texlive[47489] Master/texmf-dist: siunitx (27apr18)
commits+karl at tug.org
commits+karl at tug.org
Mon Apr 30 00:36:26 CEST 2018
Revision: 47489
http://tug.org/svn/texlive?view=revision&revision=47489
Author: karl
Date: 2018-04-30 00:36:26 +0200 (Mon, 30 Apr 2018)
Log Message:
-----------
siunitx (27apr18)
Modified Paths:
--------------
trunk/Master/texmf-dist/doc/latex/siunitx/README.md
trunk/Master/texmf-dist/doc/latex/siunitx/siunitx.pdf
trunk/Master/texmf-dist/source/latex/siunitx/siunitx.dtx
trunk/Master/texmf-dist/tex/latex/siunitx/siunitx.sty
Added Paths:
-----------
trunk/Master/texmf-dist/doc/latex/siunitx/siunitx-code.pdf
trunk/Master/texmf-dist/source/latex/siunitx/siunitx-angle.dtx
trunk/Master/texmf-dist/source/latex/siunitx/siunitx-number.dtx
trunk/Master/texmf-dist/source/latex/siunitx/siunitx-print.dtx
trunk/Master/texmf-dist/source/latex/siunitx/siunitx-table.dtx
trunk/Master/texmf-dist/source/latex/siunitx/siunitx-unit.dtx
trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v1.dtx
trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v2.dtx
trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v3.dtx
trunk/Master/texmf-dist/source/latex/siunitx/siunitx.ins
Modified: trunk/Master/texmf-dist/doc/latex/siunitx/README.md
===================================================================
--- trunk/Master/texmf-dist/doc/latex/siunitx/README.md 2018-04-29 22:36:02 UTC (rev 47488)
+++ trunk/Master/texmf-dist/doc/latex/siunitx/README.md 2018-04-29 22:36:26 UTC (rev 47489)
@@ -1,5 +1,5 @@
siunitx - A comprehensive (SI) units package
-==============================================
+============================================
Physical quantities have both numbers and units, and each
physical quantity should be expressed as the product of a number
@@ -15,41 +15,3 @@
varying typographic conventions with the same input syntax. The
package includes automated processing of numbers and units, and
the ability to control tabular alignment of numbers.
-
-Installation
-------------
-
-The package is supplied in `dtx` format and as a pre-extracted
-zip file, `siunitx.tds.zip`. The later is most convenient for
-most users: simply unzip this in your local texmf directory . If
-you want to unpack the `.dtx` yourself, running `tex
-siunitx.dtx` will extract the package whereas `latex
-siunitx.dtx` will extract it and also typeset the documentation.
-
-The package requires LaTeX3 support as provided in the
-`l3kernel` and `l3packages` bundles. Both of these are available
-on [CTAN](http://www.ctan.org/) as ready-to-install zip files.
-Suitable versions are available in MiKTeX 2.9 and TeX Live 2015
-(updating the relevant packages online may be necessary).
-LaTeX3, and so `siunitx`, requires the e-TeX extensions: these
-are available on all modern TeX systems.
-
-Typesetting the documentation requires a number of packages in
-addition to those needed to use the package. This is mainly
-because of the number of demonstration items included in the
-text. To compile the documentation without error, you will
-need the packages:
- - `amsmath`
- - `booktabs`
- - `cancel`
- - `caption`
- - `cleveref`
- - `colortbl`
- - `csquotes`
- - `datatool`
- - `helvet`
- - `mathpazo`
- - `multirow`
- - `listings`
- - `pgfplots`
- - `xcolor`
\ No newline at end of file
Added: trunk/Master/texmf-dist/doc/latex/siunitx/siunitx-code.pdf
===================================================================
(Binary files differ)
Index: trunk/Master/texmf-dist/doc/latex/siunitx/siunitx-code.pdf
===================================================================
--- trunk/Master/texmf-dist/doc/latex/siunitx/siunitx-code.pdf 2018-04-29 22:36:02 UTC (rev 47488)
+++ trunk/Master/texmf-dist/doc/latex/siunitx/siunitx-code.pdf 2018-04-29 22:36:26 UTC (rev 47489)
Property changes on: trunk/Master/texmf-dist/doc/latex/siunitx/siunitx-code.pdf
___________________________________________________________________
Added: svn:mime-type
## -0,0 +1 ##
+application/pdf
\ No newline at end of property
Modified: trunk/Master/texmf-dist/doc/latex/siunitx/siunitx.pdf
===================================================================
(Binary files differ)
Added: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-angle.dtx
===================================================================
--- trunk/Master/texmf-dist/source/latex/siunitx/siunitx-angle.dtx (rev 0)
+++ trunk/Master/texmf-dist/source/latex/siunitx/siunitx-angle.dtx 2018-04-29 22:36:26 UTC (rev 47489)
@@ -0,0 +1,84 @@
+% \iffalse meta-comment
+%
+% File: siunitx-angle.dtx Copyright (C) 2016-2017 Joseph Wright
+%
+% It may be distributed and/or modified under the conditions of the
+% LaTeX Project Public License (LPPL), either version 1.3c of this
+% license or (at your option) any later version. The latest version
+% of this license is in the file
+%
+% https://www.latex-project.org/lppl.txt
+%
+% This file is part of the "siunitx bundle" (The Work in LPPL)
+% and all files in that bundle must be distributed together.
+%
+% The released version of this bundle is available from CTAN.
+%
+% -----------------------------------------------------------------------
+%
+% The development version of the bundle can be found at
+%
+% https://github.com/josephwright/siunitx
+%
+% for those people who are interested.
+%
+% -----------------------------------------------------------------------
+%
+%<*driver>
+\documentclass{l3doc}
+% The next line is needed so that \GetFileInfo will be able to pick up
+% version data
+\usepackage{siunitx}
+\begin{document}
+ \DocInput{\jobname.dtx}
+\end{document}
+%</driver>
+% \fi
+%
+% \GetFileInfo{siunitx.sty}
+%
+% \title{^^A
+% \pkg{siunitx-angle} -- Formatting angles^^A
+% \thanks{This file describes \fileversion,
+% last revised \filedate.}^^A
+% }
+%
+% \author{^^A
+% Joseph Wright^^A
+% \thanks{^^A
+% E-mail:
+% \href{mailto:joseph.wright at morningstar2.co.uk}
+% {joseph.wright at morningstar2.co.uk}^^A
+% }^^A
+% }
+%
+% \date{Released \filedate}
+%
+% \maketitle
+%
+% \begin{documentation}
+%
+% \end{documentation}
+%
+% \begin{implementation}
+%
+% \section{\pkg{siunitx-angle} implementation}
+%
+% Start the \pkg{DocStrip} guards.
+% \begin{macrocode}
+%<*package>
+% \end{macrocode}
+%
+% Identify the internal prefix (\LaTeX3 \pkg{DocStrip} convention): only
+% internal material in this \emph{submodule} should be used directly.
+% \begin{macrocode}
+%<@@=siunitx_table>
+% \end{macrocode}
+%
+% \begin{macrocode}
+%</package>
+% \end{macrocode}
+%
+% \end{implementation}
+%
+% \PrintIndex
\ No newline at end of file
Property changes on: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-angle.dtx
___________________________________________________________________
Added: svn:eol-style
## -0,0 +1 ##
+native
\ No newline at end of property
Added: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-number.dtx
===================================================================
--- trunk/Master/texmf-dist/source/latex/siunitx/siunitx-number.dtx (rev 0)
+++ trunk/Master/texmf-dist/source/latex/siunitx/siunitx-number.dtx 2018-04-29 22:36:26 UTC (rev 47489)
@@ -0,0 +1,1862 @@
+% \iffalse meta-comment
+%
+% File: siunitx-number.dtx Copyright (C) 2014-2017 Joseph Wright
+%
+% It may be distributed and/or modified under the conditions of the
+% LaTeX Project Public License (LPPL), either version 1.3c of this
+% license or (at your option) any later version. The latest version
+% of this license is in the file
+%
+% https://www.latex-project.org/lppl.txt
+%
+% This file is part of the "siunitx bundle" (The Work in LPPL)
+% and all files in that bundle must be distributed together.
+%
+% The released version of this bundle is available from CTAN.
+%
+% -----------------------------------------------------------------------
+%
+% The development version of the bundle can be found at
+%
+% https://github.com/josephwright/siunitx
+%
+% for those people who are interested.
+%
+% -----------------------------------------------------------------------
+%
+%<*driver>
+\documentclass{l3doc}
+% The next line is needed so that \GetFileInfo will be able to pick up
+% version data
+\usepackage{siunitx}
+\begin{document}
+ \DocInput{\jobname.dtx}
+\end{document}
+%</driver>
+% \fi
+%
+% \GetFileInfo{siunitx.sty}
+%
+% \title{^^A
+% \pkg{siunitx-number} -- Parsing and formatting numbers^^A
+% \thanks{This file describes \fileversion,
+% last revised \filedate.}^^A
+% }
+%
+% \author{^^A
+% Joseph Wright^^A
+% \thanks{^^A
+% E-mail:
+% \href{mailto:joseph.wright at morningstar2.co.uk}
+% {joseph.wright at morningstar2.co.uk}^^A
+% }^^A
+% }
+%
+% \date{Released \filedate}
+%
+% \maketitle
+%
+% \begin{documentation}
+%
+% \begin{function}{\siunitx_number_format:nN, \siunitx_number_format:VN}
+% \begin{syntax}
+% \cs{siunitx_number_format:nN} \Arg{number} \meta{tl~var}
+% \end{syntax}
+% \end{function}
+%
+% \begin{function}{\siunitx_number_format:nNN}
+% \begin{syntax}
+% \cs{siunitx_number_format:nNN} \Arg{number} \meta{tl~var} \meta{marker}
+% \end{syntax}
+% \end{function}
+%
+% \begin{function}[TF]{\siunitx_if_number:n}
+% \begin{syntax}
+% \cs{siunitx_if_number_token:NTF} \Arg{tokens}
+% \Arg{true code} \Arg{false code}
+% \end{syntax}
+% Determines if the \meta{tokens} form a valid number which can be fully
+% parsed by \pkg{siunitx}.
+% \end{function}
+%
+% \begin{function}[TF]{\siunitx_if_number_token:N}
+% \begin{syntax}
+% \cs{siunitx_if_number_token:NTF} \Arg{token}
+% \Arg{true code} \Arg{false code}
+% \end{syntax}
+% Determines if the \meta{token} is valid in a number based on those
+% tokens currently set up for detection in a number.
+% \end{function}
+%
+% \end{documentation}
+%
+% \begin{implementation}
+%
+% \section{\pkg{siunitx-number} implementation}
+%
+% Start the \pkg{DocStrip} guards.
+% \begin{macrocode}
+%<*package>
+% \end{macrocode}
+%
+% Identify the internal prefix (\LaTeX3 \pkg{DocStrip} convention): only
+% internal material in this \emph{submodule} should be used directly.
+% \begin{macrocode}
+%<@@=siunitx_number>
+% \end{macrocode}
+%
+% \subsection{Initial set-up}
+%
+% Variants not provided by \pkg{expl3}.
+% \begin{macrocode}
+\cs_generate_variant:Nn \tl_if_blank:nTF { f }
+\cs_generate_variant:Nn \tl_if_blank_p:n { f }
+\cs_generate_variant:Nn \tl_if_in:NnTF { NV }
+% \end{macrocode}
+%
+% \begin{variable}{\l_@@_tmp_tl}
+% Scratch space.
+% \begin{macrocode}
+\tl_new:N \l_@@_tmp_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \subsection{Main formatting routine}
+%
+% \begin{variable}{\l_@@_formatted_tl}
+% A token list for the final formatted result: may or may not be generated
+% by the parser, depending on settings which are active.
+% \begin{macrocode}
+\tl_new:N \l_@@_formatted_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_tab_tl}
+% A token list for marking the position of tabular alignments in formatted
+% output.
+% \begin{macrocode}
+\tl_new:N \l_@@_tab_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\siunitx_number_format:nN, \siunitx_number_format:VN}
+% \begin{macro}{\siunitx_number_format:nNN}
+% \begin{macro}{\@@_format:nN}
+% \begin{macrocode}
+\cs_new_protected:Npn \siunitx_number_format:nN #1#2
+ {
+ \tl_clear:N \l_@@_tab_tl
+ \@@_format:nN {#1} #2
+ }
+\cs_generate_variant:Nn \siunitx_number_format:nN { V }
+\cs_new_protected:Npn \siunitx_number_format:nNN #1#2#3
+ {
+ \tl_set:Nn \l_@@_tab_tl {#3}
+ \@@_format:nN {#1} #2
+ }
+\cs_new_protected:Npn \@@_format:nN #1#2
+ {
+ \group_begin:
+ \@@_parse:n {#1}
+ \@@_format:
+ \exp_args:NNNV \group_end:
+ \tl_set:Nn #2 \l_@@_formatted_tl
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Parsing numbers}
+%
+% Before numbers can be manipulated or formatted they need to be parsed into
+% an internal form. In particular, if multiple code paths are to be avoided,
+% it is necessary to do such parsing even for relatively simple cases such
+% as converting |1e10| to |1 \times 10^{10}|.
+%
+% Storing the result of such parsing can be done in a number of ways. In the
+% first version of \pkg{siunitx} a series of separate data stores were used.
+% This is potentially quite fast (as recovery of items relies only on \TeX{}'s
+% hash table) but makes managing the various data entries somewhat tedious and
+% error-prone. For version two of the package, a single data structure
+% (property list) was used for each part of the parsed number. Whilst this is
+% easy to manage and extend, it is somewhat slower as at a \TeX{} level there
+% are repeated pack--unpack steps. In particular, the fact that there are a
+% limited number of items to track for a \enquote{number} means that a more
+% efficient approach is desirable (contrast parsing units, which is open-ended
+% and therefore fits well with using a property list).
+%
+% To allow for complex numbers, two parallel data structures are used, one for
+% the real part and one for the imaginary part. If the part is entirely absent
+% then the data structures are left empty. Within each part, the structure
+% is
+% \begin{quote}
+% \marg{comparator}\meta{sign}\marg{integer}\marg{decimal}
+% \marg{uncertainty}\\
+% \meta{exponent sign}\marg{exponent}
+% \end{quote}
+% where the two sign parts must be single tokens and all other components
+% must be given in braces. \emph{All} of the components must be present in
+% a stored number (\emph{i.e.}~at the end of parsing). The number must have
+% at least one digit for both the \meta{integer} and \meta{exponent} parts.
+%
+% \begin{variable}
+% {
+% \l_@@_expression_bool ,
+% \l_@@_input_uncert_close_tl ,
+% \l_@@_input_complex_tl ,
+% \l_@@_input_comparator_tl ,
+% \l_@@_input_decimal_tl ,
+% \l_@@_input_digit_tl ,
+% \l_@@_input_exponent_tl ,
+% \l_@@_input_ignore_tl ,
+% \l_@@_input_uncert_open_tl ,
+% \l_@@_input_sign_tl ,
+% \l_@@_input_uncert_sign_tl
+% }
+% \begin{macro}[EXP]{\@@_expression:n}
+% Options which determine the various valid parts of a parsed number.
+% \begin{macrocode}
+\keys_define:nn { siunitx }
+ {
+ evaluate-expression .bool_set:N =
+ \l_@@_expression_bool ,
+ expression .code:n =
+ \cs_set:Npn \@@_expression:n ##1 {#1} ,
+ input-close-uncertainty .tl_set:N =
+ \l_@@_input_uncert_close_tl ,
+ input-complex-roots .tl_set:N =
+ \l_@@_input_complex_tl ,
+ input-comparators .tl_set:N =
+ \l_@@_input_comparator_tl ,
+ input-decimal-markers .tl_set:N =
+ \l_@@_input_decimal_tl ,
+ input-digits .tl_set:N =
+ \l_@@_input_digit_tl ,
+ input-exponent-markers .tl_set:N =
+ \l_@@_input_exponent_tl ,
+ input-ignore .tl_set:N =
+ \l_@@_input_ignore_tl ,
+ input-open-uncertainty .tl_set:N =
+ \l_@@_input_uncert_open_tl ,
+ input-signs .tl_set:N =
+ \l_@@_input_sign_tl ,
+ input-uncertainty-signs .code:n =
+ {
+ \tl_set:Nn \l_@@_input_uncert_sign_tl {#1}
+ \tl_map_inline:nn {#1}
+ {
+ \tl_if_in:NnF \l_@@_input_sign_tl {##1}
+ { \tl_put_right:Nn \l_@@_input_sign_tl {##1} }
+ }
+ }
+ }
+\cs_new:Npn \@@_expression:n #1 { }
+% \end{macrocode}
+% \end{macro}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_arg_tl}
+% The input argument or a part thereof, depending on the position in
+% the parsing routine.
+% \begin{macrocode}
+\tl_new:N \l_@@_arg_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_comparator_tl}
+% A comparator, if found, is held here.
+% \begin{macrocode}
+\tl_new:N \l_@@_comparator_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_exponent_tl}
+% The exponent part of a parsed number. It is easiest to find this
+% relatively early in the parsing process, but as it needs to go at
+% the end of the internal format is held separately until required.
+% \begin{macrocode}
+\tl_new:N \l_@@_exponent_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_flex_tl}
+% When parsing for a separate uncertainty or complex number, the nature
+% of the grabbed part cannot be determined until the end of the number.
+% To avoid abusing the storage areas, this dedicated one is used for
+% \enquote{flexible} cases.
+% \begin{macrocode}
+\tl_new:N \l_@@_flex_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_imaginary_tl, \l_@@_real_tl}
+% Used to hold the real and imaginary parts of a number in the standardised
+% format.
+% \begin{macrocode}
+\tl_new:N \l_@@_imaginary_tl
+\tl_new:N \l_@@_real_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_input_tl}
+% The numerical input exactly as given by the user.
+% \begin{macrocode}
+\tl_new:N \l_@@_input_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_partial_tl}
+% To avoid needing to worry about the fact that the final data stores are
+% somewhat tricky to add to token-by-token, a simple store is used to build
+% up the parsed part of a number before transferring in one go.
+% \begin{macrocode}
+\tl_new:N \l_@@_partial_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_validate_bool}
+% Used to set up for validation with no error production.
+% \begin{macrocode}
+\bool_new:N \l_@@_validate_bool
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_parse:n}
+% After some initial set up, the parser expands the input and then replaces
+% as far as possible tricky tokens with ones that can be handled using
+% delimited arguments. The parser begins with the assumption that the input
+% is a real number. To avoid multiple conditionals here, the parser is
+% set up as a chain of commands initially, with a loop only later. This
+% avoids more conditionals than are necessary.
+% \begin{macrocode}q
+\cs_new_protected:Npn \@@_parse:n #1
+ {
+ \tl_clear:N \l_@@_imaginary_tl
+ \tl_clear:N \l_@@_real_tl
+ \protected at edef \l_@@_arg_tl
+ {
+ \bool_if:NTF \l_@@_expression_bool
+ { \fp_eval:n { \@@_expression:n {#1} } }
+ {#1}
+ }
+ \tl_set_eq:NN \l_@@_input_tl \l_@@_arg_tl
+ \@@_parse_replace:
+ \tl_if_empty:NF \l_@@_arg_tl
+ { \@@_parse_comparator: }
+ \@@_parse_check:
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_check:}
+% After the loop there is one case that might need tidying up. If a
+% separated uncertainty was found it will be currently in \cs{l_@@_flex_tl}
+% and needs moving. A series of tests pick up that case, then the check is
+% made that some content was found for at least one of the real or imaginary
+% parts of the number.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_check:
+ {
+ \tl_if_empty:NF \l_@@_flex_tl
+ {
+ \bool_lazy_and:nnTF
+ {
+ \tl_if_blank_p:f
+ { \exp_after:wN \use_iv:nnnn \l_@@_real_tl }
+ }
+ {
+ \tl_if_blank_p:f
+ { \exp_after:wN \use_iv:nnnn \l_@@_flex_tl }
+ }
+ {
+ \tl_set:Nx \l_@@_tmp_tl
+ { \exp_after:wN \use_i:nnnn \l_@@_flex_tl }
+ \tl_if_in:NVTF \l_@@_input_uncert_sign_tl
+ \l_@@_tmp_tl
+ { \@@_parse_combine_uncert: }
+ { \tl_clear:N \l_@@_real_tl }
+ }
+ { \tl_clear:N \l_@@_real_tl }
+ }
+ \bool_lazy_and:nnTF
+ { \tl_if_empty_p:N \l_@@_real_tl }
+ { \tl_if_empty_p:N \l_@@_imaginary_tl }
+ {
+ \bool_if:NF \l_@@_validate_bool
+ {
+ \msg_error:nnx { siunitx } { number / invalid-input }
+ { \exp_not:V \l_@@_input_tl }
+ }
+ }
+ { \@@_parse_finalise: }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_combine_uncert:}
+% \begin{macro}{\@@_parse_combine_uncert_auxi:NnnnNnnn}
+% \begin{macro}
+% {
+% \@@_parse_combine_uncert_auxii:nnnnn,
+% \@@_parse_combine_uncert_auxii:fnnnn
+% }
+% \begin{macro}
+% {
+% \@@_parse_combine_uncert_auxiii:nnnnnn,
+% \@@_parse_combine_uncert_auxiii:fnnnnn
+% }
+% \begin{macro}{\@@_parse_combine_uncert_auxiv:nnnn}
+% \begin{macro}[EXP]{\@@_parse_combine_uncert_auxv:w}
+% \begin{macro}[EXP]{\@@_parse_combine_uncert_auxvi:w}
+% Conversion of a second numerical part to an uncertainty needs a bit of
+% work. The first step is to extract the useful information from the two
+% stores: the sign, integer and decimal parts from the real number and the
+% integer and decimal parts from the second number. That is done using the
+% input stack to avoid lots of assignments.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_combine_uncert:
+ {
+ \exp_after:wN \exp_after:wN \exp_after:wN
+ \@@_parse_combine_uncert_auxi:NnnnNnnn
+ \exp_after:wN \l_@@_real_tl \l_@@_flex_tl
+ }
+% \end{macrocode}
+% Here, |#4|, |#5| and |#8| are all junk arguments simply there to mop up
+% tokens, while |#1| will be recovered later from \cs{l_@@_real_tl} so does
+% not need to be passed about. The difference in places between the two
+% decimal parts is now found: this is done just once to avoid having to
+% parse token lists twice. The value is then used to generate a number of
+% filler |0| tokens, and these are added to the appropriate part of the
+% number. Finally, everything is recombined: the integer part only needs
+% a test to avoid an empty main number.
+% \begin{macrocode}
+\cs_new_protected:Npn
+ \@@_parse_combine_uncert_auxi:NnnnNnnn #1#2#3#4#5#6#7#8
+ {
+ \int_compare:nNnTF { \tl_count:n {#6} } > { \tl_count:n {#2} }
+ {
+ \tl_clear:N \l_@@_real_tl
+ \tl_clear:N \l_@@_flex_tl
+ }
+ {
+ \@@_parse_combine_uncert_auxii:fnnnn
+ { \int_eval:n { \tl_count:n {#3} - \tl_count:n {#7} } }
+ {#2} {#3} {#6} {#7}
+ }
+ }
+\cs_new_protected:Npn
+ \@@_parse_combine_uncert_auxii:nnnnn #1
+ {
+ \@@_parse_combine_uncert_auxiii:fnnnnn
+ { \prg_replicate:nn { \int_abs:n {#1} } { 0 } }
+ {#1}
+ }
+\cs_generate_variant:Nn \@@_parse_combine_uncert_auxii:nnnnn { f }
+\cs_new_protected:Npn
+ \@@_parse_combine_uncert_auxiii:nnnnnn #1#2#3#4#5#6
+ {
+ \int_compare:nNnTF {#2} > 0
+ {
+ \@@_parse_combine_uncert_auxiv:nnnn
+ {#3} {#4} {#5} { #6 #1 }
+ }
+ {
+ \@@_parse_combine_uncert_auxiv:nnnn
+ {#3} { #4 #1 } {#5} {#6}
+ }
+ }
+\cs_generate_variant:Nn
+ \@@_parse_combine_uncert_auxiii:nnnnnn { f }
+\cs_new_protected:Npn
+ \@@_parse_combine_uncert_auxiv:nnnn #1#2#3#4
+ {
+ \tl_set:Nx \l_@@_real_tl
+ {
+ \tl_head:V \l_@@_real_tl
+ { \exp_not:n {#1} }
+ {
+ \bool_lazy_and:nnTF
+ { \tl_if_blank_p:n {#2} }
+ { ! \tl_if_blank_p:n {#4} }
+ { 0 }
+ { \exp_not:n {#2} }
+ }
+ {
+ \@@_parse_combine_uncert_auxv:w #3#4
+ \q_recursion_tail \q_recursion_stop
+ }
+ }
+ }
+% \end{macrocode}
+% A short routine to remove any leading zeros in the uncertainty part,
+% which are not needed for the compact representation used by the module.
+% \begin{macrocode}
+\cs_new:Npn \@@_parse_combine_uncert_auxv:w #1
+ {
+ \quark_if_recursion_tail_stop:N #1
+ \str_if_eq:nnTF {#1} { 0 }
+ { \@@_parse_combine_uncert_auxv:w }
+ { \@@_parse_combine_uncert_auxvi:w #1 }
+ }
+\cs_new:Npn \@@_parse_combine_uncert_auxvi:w
+ #1 \q_recursion_tail \q_recursion_stop
+ { \exp_not:n {#1} }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_comparator:}
+% \begin{macro}{\@@_parse_comparator_aux:Nw}
+% A comparator has to be the very first token in the input. A such, the
+% test for this can be very fast: grab the first token, do a check and
+% if appropriate store the result.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_comparator:
+ {
+ \exp_after:wN \@@_parse_comparator_aux:Nw
+ \l_@@_arg_tl \q_stop
+ }
+\cs_new_protected:Npn \@@_parse_comparator_aux:Nw #1#2 \q_stop
+ {
+ \tl_if_in:NnTF \l_@@_input_comparator_tl {#1}
+ {
+ \tl_set:Nn \l_@@_comparator_tl {#1}
+ \tl_set:Nn \l_@@_arg_tl {#2}
+ }
+ { \tl_clear:N \l_@@_comparator_tl }
+ \tl_if_empty:NF \l_@@_arg_tl
+ { \@@_parse_sign: }
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_exponent:}
+% \begin{macro}{\@@_parse_exponent_aux:w}
+% \begin{macro}{\@@_parse_exponent_aux:nn}
+% \begin{macro}{\@@_parse_exponent_aux:Nw}
+% \begin{macro}{\@@_parse_exponent_aux:Nn}
+% \begin{macro}
+% {\@@_parse_exponent_zero_test:N, \@@_parse_exponent_check:N}
+% \begin{macro}{\@@_parse_exponent_cleanup:N}
+% An exponent part of a number has to come at the end and can only occur
+% once. Thus it is relatively easy to parse. First, there is a check that
+% an exponent part is allowed, and if so a split is made (the previous
+% part of the chain checks that there is some content in \cs{l_@@_arg_tl}
+% before calling this function). After splitting, if there is no exponent
+% then simply save a default. Otherwise, check for a sign and then store
+% either this or an assumed |+| and the digits after a check that nothing
+% else is present after the~|e|. The only slight complication to all of
+% this is allowing an arbitrary token in the input to represent the exponent:
+% this is done by setting any exponent tokens to the first of the allowed
+% list, then using that in a delimited argument set up. Once an exponent
+% part is found, there is a loop to check that each of the tokens is a digit
+% then a tidy up step to remove any leading zeros.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_exponent:
+ {
+ \tl_if_empty:NTF \l_@@_input_exponent_tl
+ { \tl_set:Nn \l_@@_exponent_tl { +0 } }
+ {
+ \tl_set:Nx \l_@@_tmp_tl
+ { \tl_head:V \l_@@_input_exponent_tl }
+ \tl_map_inline:Nn \l_@@_input_exponent_tl
+ {
+ \tl_replace_all:NnV \l_@@_arg_tl
+ {##1} \l_@@_tmp_tl
+ }
+ \use:x
+ {
+ \cs_set_protected:Npn
+ \exp_not:N \@@_parse_exponent_aux:w
+ ####1 \exp_not:V \l_@@_tmp_tl
+ ####2 \exp_not:V \l_@@_tmp_tl
+ ####3 \exp_not:N \q_stop
+ }
+ { \@@_parse_exponent_aux:nn {##1} {##2} }
+ \use:x
+ {
+ \@@_parse_exponent_aux:w
+ \exp_not:V \l_@@_arg_tl
+ \exp_not:V \l_@@_tmp_tl \exp_not:N \q_nil
+ \exp_not:V \l_@@_tmp_tl \exp_not:N \q_stop
+ }
+ }
+ }
+\cs_new_protected:Npn \@@_parse_exponent_aux:w { }
+\cs_new_protected:Npn \@@_parse_exponent_aux:nn #1#2
+ {
+ \quark_if_nil:nTF {#2}
+ { \tl_set:Nn \l_@@_exponent_tl { +0 } }
+ {
+ \tl_set:Nn \l_@@_arg_tl {#1}
+ \tl_if_blank:nTF {#2}
+ { \tl_clear:N \l_@@_real_tl }
+ { \@@_parse_exponent_aux:Nw #2 \q_stop }
+ }
+ \tl_if_empty:NF \l_@@_real_tl
+ { \@@_parse_loop: }
+ }
+\cs_new_protected:Npn \@@_parse_exponent_aux:Nw #1#2 \q_stop
+ {
+ \tl_if_in:NnTF \l_@@_input_sign_tl {#1}
+ { \@@_parse_exponent_aux:Nn #1 {#2} }
+ { \@@_parse_exponent_aux:Nn + {#1#2} }
+ \tl_if_empty:NT \l_@@_exponent_tl
+ { \tl_clear:N \l_@@_real_tl }
+ }
+\cs_new_protected:Npn \@@_parse_exponent_aux:Nn #1#2
+ {
+ \tl_set:Nn \l_@@_exponent_tl { #1 }
+ \tl_if_blank:nTF {#2}
+ { \tl_clear:N \l_@@_real_tl }
+ {
+ \@@_parse_exponent_zero_test:N #2
+ \q_recursion_tail \q_recursion_stop
+ }
+ }
+\cs_new_protected:Npn \@@_parse_exponent_zero_test:N #1
+ {
+ \quark_if_recursion_tail_stop_do:Nn #1
+ { \tl_set:Nn \l_@@_exponent_tl { +0 } }
+ \str_if_eq:nnTF {#1} { 0 }
+ { \@@_parse_exponent_zero_test:N }
+ { \@@_parse_exponent_check:N #1 }
+ }
+\cs_new_protected:Npn \@@_parse_exponent_check:N #1
+ {
+ \quark_if_recursion_tail_stop:N #1
+ \tl_if_in:NnTF \l_@@_input_digit_tl {#1}
+ {
+ \tl_put_right:Nn \l_@@_exponent_tl {#1}
+ \@@_parse_exponent_check:N
+ }
+ { \@@_parse_exponent_cleanup:wN }
+ }
+\cs_new_protected:Npn \@@_parse_exponent_cleanup:wN
+ #1 \q_recursion_stop
+ { \tl_clear:N \l_@@_real_tl }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_replace:}
+% \begin{macro}{\@@_parse_replace_aux:nN}
+% \begin{macro}{\@@_parse_replace_sign:}
+% \begin{variable}{\c_@@_parse_sign_replacement_tl}
+% There are two parts to the replacement code. First, any active
+% hyphens signs are normalised: these can come up with some packages and
+% cause issues. Multi-token signs then are converted to the single token
+% equivalents so that everything else can work on a one token basis.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_replace:
+ {
+ \@@_parse_replace_minus:
+ \exp_last_unbraced:NV \@@_parse_replace_aux:nN
+ \c_@@_parse_sign_replacement_tl
+ { ? } \q_recursion_tail
+ \q_recursion_stop
+ }
+\cs_set_protected:Npn \@@_parse_replace_aux:nN #1#2
+ {
+ \quark_if_recursion_tail_stop:N #2
+ \tl_replace_all:Nnn \l_@@_arg_tl {#1} {#2}
+ \@@_parse_replace_aux:nN
+ }
+\tl_const:Nn \c_@@_parse_sign_replacement_tl
+ {
+ { -+ } \mp
+ { +- } \pm
+ { << } \ll
+ { <= } \le
+ { >> } \gg
+ { >= } \ge
+ }
+\group_begin:
+ \char_set_catcode_active:N \-
+ \cs_new_protected:Npx \@@_parse_replace_minus:
+ {
+ \tl_replace_all:Nnn \exp_not:N \l_@@_arg_tl
+ { \exp_not:N - } { \token_to_str:N - }
+ }
+\group_end:
+% \end{macrocode}
+% \end{variable}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_finalise:}
+% \begin{macro}{\@@_parse_finalise_aux:N}
+% \begin{macro}{\@@_parse_finalise_aux:Nw}
+% Combine all of the bits of a number together: both the real and
+% imaginary parts contain all of the data.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_finalise:
+ {
+ \@@_parse_finalise_aux:N \l_@@_real_tl
+ \@@_parse_finalise_aux:N \l_@@_imaginary_tl
+ }
+\cs_new_protected:Npn \@@_parse_finalise_aux:N #1
+ {
+ \tl_if_empty:NF #1
+ {
+ \tl_set:Nx #1
+ {
+ { \exp_not:V \l_@@_comparator_tl }
+ \exp_not:V #1
+ \exp_after:wN \@@_parse_finalise_aux:Nw
+ \l_@@_exponent_tl \q_stop
+ }
+ }
+ }
+\cs_new:Npn \@@_parse_finalise_aux:Nw #1#2 \q_stop
+ {
+ \exp_not:N #1
+ { \exp_not:n {#2} }
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_loop:}
+% \begin{macro}{\@@_parse_loop_first:N}
+% \begin{macro}{\@@_parse_loop_main:NNNNN}
+% \begin{macro}{\@@_parse_loop_main_end:NN}
+% \begin{macro}{\@@_parse_loop_main_digit:NNNNN}
+% \begin{macro}{\@@_parse_loop_main_decimal:NN}
+% \begin{macro}{\@@_parse_loop_main_uncert:NNN}
+% \begin{macro}{\@@_parse_loop_main_complex:N}
+% \begin{macro}{\@@_parse_loop_main_sign:NNN}
+% \begin{macro}{\@@_parse_loop_main_store:NNN}
+% \begin{macro}{\@@_parse_loop_after_decimal:NNN}
+% \begin{macro}{\@@_parse_loop_uncert:NNNNN}
+% \begin{macro}{\@@_parse_loop_after_uncert:NNN}
+% \begin{macro}{\@@_parse_loop_root_swap:NNwNN}
+% \begin{macro}{\@@_parse_loop_complex_cleanup:wN}
+% \begin{macro}{\@@_parse_loop_break:wN}
+% At this stage, the partial input \cs{l_@@_arg_tl} will contain any
+% mantissa, which may contain an uncertainty or complex part. Parsing this
+% and allowing for all of the different formats possible is best done using
+% a token-by-token approach. However, as at each stage only a subset of
+% tokens are valid, the approach take is to use a set of semi-dedicated
+% functions to parse different components along with switches to allow a
+% sensible amount of code sharing.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop:
+ {
+ \tl_clear:N \l_@@_partial_tl
+ \exp_after:wN \@@_parse_loop_first:NNN
+ \exp_after:wN \l_@@_real_tl \exp_after:wN \c_true_bool
+ \l_@@_arg_tl
+ \q_recursion_tail \q_recursion_stop
+ }
+% \end{macrocode}
+% The very first token of the input is handled with a dedicated function.
+% Valid cases here are
+% \begin{itemize}
+% \item Entirely blank if the original input was for example |+e10|:
+% simply clean up if in the integer part of issue an error if in
+% a second part (complex number, \emph{etc.}).
+% \item An integer part digit: pass through to the main collection
+% routine.
+% \item A decimal marker: store an empty integer part and move to
+% the main collection routine for a decimal part.
+% \item A complex root token: shuffle to the end of the input.
+% \end{itemize}
+% Anything else is invalid and sends the code to the abort function.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_first:NNN #1#2#3
+ {
+ \quark_if_recursion_tail_stop_do:Nn #3
+ {
+ \bool_if:NTF #2
+ { \tl_put_right:Nn #1 { { 1 } { } { } } }
+ { \@@_parse_loop_break:wN \q_recursion_stop }
+ }
+ \tl_if_in:NnTF \l_@@_input_digit_tl {#3}
+ {
+ \@@_parse_loop_main:NNNNN
+ #1 \c_true_bool \c_false_bool #2 #3
+ }
+ {
+ \tl_if_in:NnTF \l_@@_input_decimal_tl {#3}
+ {
+ \tl_put_right:Nn #1 { { 0 } }
+ \@@_parse_loop_after_decimal:NNN #1 #2
+ }
+ {
+ \tl_if_in:NnTF \l_@@_input_complex_tl {#3}
+ { \@@_parse_loop_root_swap:NNwNN #1 #3 }
+ { \@@_parse_loop_break:wN }
+ }
+ }
+ }
+% \end{macrocode}
+% A single function is used to cover the \enquote{main} part of numbers:
+% finding real, complex or separated uncertainty parts and covering both
+% the integer and decimal components. This works because these elements
+% share a lot of concepts: a small number of switches can be used to
+% differentiate between them. To keep the code at least somewhat readable,
+% this main function deals with the validity testing but hands off other
+% tasks to dedicated auxiliaries for each case.
+%
+% The possibilities are
+% \begin{itemize}
+% \item The number terminates, meaning that some digits were collected
+% and everything is simply tidied up (as far as the loop is concerned).
+% \item A digit is found: this is the common case and leads to a storage
+% auxiliary (which handles non-significant zeros).
+% \item A decimal marker is found: only valid in the integer part and
+% there leading to a store-and-switch situation.
+% \item An open-uncertainty token: switch to the dedicated collector
+% for uncertainties.
+% \item A complex root token: store the current number as an imaginary
+% part and terminate the loop.
+% \item A sign token (if allowed): stop collecting this number and
+% restart collection for the second part.
+% \end{itemize}
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_main:NNNNN #1#2#3#4#5
+ {
+ \quark_if_recursion_tail_stop_do:Nn #5
+ { \@@_parse_loop_main_end:NN #1#2 }
+ \tl_if_in:NnTF \l_@@_input_digit_tl {#5}
+ { \@@_parse_loop_main_digit:NNNNN #1#2#3#4#5 }
+ {
+ \tl_if_in:NnTF \l_@@_input_decimal_tl {#5}
+ {
+ \bool_if:NTF #2
+ { \@@_parse_loop_main_decimal:NN #1 #4 }
+ { \@@_parse_loop_break:wN }
+ }
+ {
+ \tl_if_in:NnTF \l_@@_input_uncert_open_tl {#5}
+ { \@@_parse_loop_main_uncert:NNN #1#2 #4 }
+ {
+ \tl_if_in:NnTF \l_@@_input_complex_tl {#5}
+ {
+ \@@_parse_loop_main_store:NNN
+ #1 #2 \c_true_bool
+ \@@_parse_loop_main_complex:N #1
+ }
+ {
+ \bool_if:NTF #4
+ {
+ \tl_if_in:NnTF \l_@@_input_sign_tl {#5}
+ {
+ \@@_parse_loop_main_sign:NNN
+ #1#2 #5
+ }
+ { \@@_parse_loop_break:wN }
+ }
+ { \@@_parse_loop_break:wN }
+ }
+ }
+ }
+ }
+ }
+% \end{macrocode}
+% If the main loop finds the end marker then there is a tidy up phase.
+% The current partial number is stored either as the integer or decimal,
+% depending on the setting for the indicator switch. For the integer
+% part, if no number has been collected then one or more non-significant
+% zeros have been dropped. Exactly one zero is therefore needed to make
+% sure the parsed result is correct.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_main_end:NN #1#2
+ {
+ \bool_lazy_and:nnT
+ {#2} { \tl_if_empty_p:N \l_@@_partial_tl }
+ { \tl_set:Nn \l_@@_partial_tl { 0 } }
+ \tl_put_right:Nx #1
+ {
+ { \exp_not:V \l_@@_partial_tl }
+ \bool_if:NT #2 { { } }
+ { }
+ }
+ }
+% \end{macrocode}
+% The most common case for the main loop collector is to find a digit.
+% Here, in the integer part it is possible that zeros are non-significant:
+% that is handled using a combination of a switch and a string test. Other
+% than that, the situation here is simple: store the input and loop.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_main_digit:NNNNN #1#2#3#4#5
+ {
+ \bool_lazy_or:nnTF
+ {#3} { ! \str_if_eq_p:nn {#5} { 0 } }
+ {
+ \tl_put_right:Nn \l_@@_partial_tl {#5}
+ \@@_parse_loop_main:NNNNN #1 #2 \c_true_bool #4
+ }
+ { \@@_parse_loop_main:NNNNN #1 #2 \c_false_bool #4 }
+ }
+% \end{macrocode}
+% When a decimal marker was found, move the integer part to the
+% store and then go back to the loop with the flags set correctly.
+% There is the case of non-significant zeros to cover before that, of course.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_main_decimal:NN #1#2
+ {
+ \@@_parse_loop_main_store:NNN #1 \c_false_bool \c_false_bool
+ \@@_parse_loop_after_decimal:NNN #1 #2
+ }
+% \end{macrocode}
+% Starting an uncertainty part means storing the number to date as in other
+% cases, with the possibility of a blank decimal part allowed for. The
+% uncertainty itself is collected by a dedicated function as it is extremely
+% restricted.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_main_uncert:NNN #1#2#3
+ {
+ \@@_parse_loop_main_store:NNN #1 #2 \c_false_bool
+ \@@_parse_loop_uncert:NNNNN
+ #1 \c_true_bool \c_false_bool #3
+ }
+% \end{macrocode}
+% A complex root token has to be at the end of the input (leading ones
+% are dealt with specially). Thus after moving the data to the correct
+% place there is a hand-off to a cleanup function. The case where only the
+% complex root token was given is covered by
+% \cs{@@_parse_loop_root_swap:NNwNN}.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_main_complex:N #1
+ {
+ \tl_set_eq:NN \l_@@_imaginary_tl #1
+ \tl_clear:N #1
+ \@@_parse_loop_complex_cleanup:wN
+ }
+% \end{macrocode}
+% If a sign is found, terminate the current number, store the sign as the
+% first token of the second part and go back to do the dedicated first-token
+% function.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_main_sign:NNN #1#2#3
+ {
+ \@@_parse_loop_main_store:NNN #1 #2 \c_true_bool
+ \tl_set:Nn \l_@@_flex_tl {#3}
+ \@@_parse_loop_first:NNN
+ \l_@@_flex_tl \c_false_bool
+ }
+% \end{macrocode}
+% A common auxiliary for the various non-digit token functions: tidy up the
+% integer and decimal parts of a number. Here, the two flags are used to
+% indicate if empty decimal and uncertainty parts should be included in
+% the storage cycle.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_main_store:NNN #1#2#3
+ {
+ \tl_if_empty:NT \l_@@_partial_tl
+ { \tl_set:Nn \l_@@_partial_tl { 0 } }
+ \tl_put_right:Nx #1
+ {
+ { \exp_not:V \l_@@_partial_tl }
+ \bool_if:NT #2 { { } }
+ \bool_if:NT #3 { { } }
+ }
+ \tl_clear:N \l_@@_partial_tl
+ }
+% \end{macrocode}
+% After a decimal marker there has to be a digit if there wasn't one before
+% it. That is handled by using a dedicated function, which checks for
+% an empty integer part first then either simply hands off or looks for
+% a digit.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_after_decimal:NNN #1#2#3
+ {
+ \tl_if_blank:fTF { \exp_after:wN \use_none:n #1 }
+ {
+ \quark_if_recursion_tail_stop_do:Nn #3
+ { \@@_parse_loop_break:wN \q_recursion_stop }
+ \tl_if_in:NnTF \l_@@_input_digit_tl {#1}
+ {
+ \tl_put_right:Nn \l_@@_partial_tl {#3}
+ \@@_parse_loop_main:NNNNN
+ #1 \c_false_bool \c_true_bool #2
+ }
+ { \@@_parse_loop_break:wN }
+ }
+ {
+ \@@_parse_loop_main:NNNNN
+ #1 \c_false_bool \c_true_bool #2 #3
+ }
+ }
+% \end{macrocode}
+% Inside the brackets for an uncertainty the range of valid choices is
+% very limited. Either the token is a digit, in which case there is a
+% test to look for non-significant zeros, or it is a closing bracket. The
+% latter is not valid for the very first token, which is handled using a
+% switch (it's a simple enough difference).
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_uncert:NNNNN #1#2#3#4#5
+ {
+ \quark_if_recursion_tail_stop_do:Nn #5
+ { \@@_parse_loop_break:wN \q_recursion_stop }
+ \tl_if_in:NnTF \l_@@_input_digit_tl {#5}
+ {
+ \bool_lazy_or:nnTF
+ {#3} { ! \str_if_eq_p:nn {#5} { 0 } }
+ {
+ \tl_put_right:Nn \l_@@_partial_tl {#5}
+ \@@_parse_loop_uncert:NNNNN
+ #1 \c_false_bool \c_true_bool #4
+ }
+ {
+ \@@_parse_loop_uncert:NNNNN
+ #1 \c_false_bool \c_false_bool #4
+ }
+ }
+ {
+ \tl_if_in:NnTF \l_@@_input_uncert_close_tl {#5}
+ {
+ \bool_if:NTF #2
+ { \@@_parse_loop_break:wN }
+ {
+ \@@_parse_loop_main_store:NNN #1
+ \c_false_bool \c_false_bool
+ \@@_parse_loop_after_uncert:NNN #1 #3
+ }
+ }
+ { \@@_parse_loop_break:wN }
+ }
+ }
+% \end{macrocode}
+% After a bracketed uncertainty there are only a very small number of
+% valid choices. The number can end, there can be a complex root token
+% or there can be a sign. The latter is only allowed if the part being
+% parsed at the moment was the first part of the number. The case where
+% there is no root symbol but there should have been is cleared up after
+% the loop code, so at this stage there is no check.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_after_uncert:NNN #1#2#3
+ {
+ \quark_if_recursion_tail_stop:N #3
+ \tl_if_in:NnTF \l_@@_input_complex_tl {#3}
+ { \@@_parse_loop_main_complex:N #1 }
+ {
+ \bool_if:NTF #2
+ {
+ \tl_if_in:NnTF \l_@@_input_sign_tl {#3}
+ {
+ \tl_set:Nn \l_@@_flex_tl {#3}
+ \@@_parse_loop_first:NNN
+ \l_@@_flex_tl \c_false_bool
+ }
+ { \@@_parse_loop_break:wN }
+ }
+ { \@@_parse_loop_break:wN }
+ }
+ }
+% \end{macrocode}
+% When the complex root symbol comes at the start of the number rather than
+% at the end, the easiest approach is to shuffle it to the \enquote{normal}
+% position. As the exponent has already been removed, this must be the last
+% token of the input and any duplication will be picked up. The case where
+% just a complex root token has to be covered: in that situation, there is
+% an implicit |1| to store after which the loop stops.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_root_swap:NNwNN #1#2#3
+ \q_recursion_tail \q_recursion_stop
+ {
+ \tl_if_blank:nTF {#3}
+ {
+ \tl_set:Nx \l_@@_imaginary_tl
+ {
+ \exp_not:V #1
+ { 1 } { } { }
+ }
+ \tl_clear:N #1
+ }
+ {
+ \use:x
+ {
+ \tl_clear:N \exp_not:N #1
+ \tl_set:Nn \exp_not:N \l_@@_flex_tl { \exp_not:V #1 }
+ }
+ \@@_parse_loop_first:NNN
+ \l_@@_flex_tl \c_false_bool
+ #3 #2 \q_recursion_tail \q_recursion_stop
+ }
+ }
+% \end{macrocode}
+% Nothing is allowed after a complex root token: check and if there is
+% kill the parsing.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_complex_cleanup:wN
+ #1 \q_recursion_tail \q_recursion_stop
+ {
+ \tl_if_blank:nF {#1}
+ { \@@_parse_loop_break:wN \q_recursion_stop }
+ }
+% \end{macrocode}
+% Something is not right: remove all of the remaining tokens from the
+% number and clear the storage areas as a signal for the next part of the
+% code.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_loop_break:wN
+ #1 \q_recursion_stop
+ {
+ \tl_clear:N \l_@@_imaginary_tl
+ \tl_clear:N \l_@@_flex_tl
+ \tl_clear:N \l_@@_real_tl
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_sign:}
+% \begin{macro}{\@@_parse_sign_aux:Nw}
+% The first token of a number after a comparator could be a sign. A quick
+% check is made and if found stored; if there is no sign then the internal
+% format requires that |+| is used. For the number to be valid it has to be
+% more than just a sign, so the next part of the chain is only called if that
+% is the case.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_sign:
+ {
+ \exp_after:wN \@@_parse_sign_aux:Nw
+ \l_@@_arg_tl \q_stop
+ }
+\cs_new_protected:Npn \@@_parse_sign_aux:Nw #1#2 \q_stop
+ {
+ \tl_if_in:NnTF \l_@@_input_sign_tl {#1}
+ {
+ \tl_set:Nn \l_@@_arg_tl {#2}
+ \tl_set:Nn \l_@@_real_tl {#1}
+ }
+ { \tl_set:Nn \l_@@_real_tl { + } }
+ \tl_if_empty:NTF \l_@@_arg_tl
+ { \tl_clear:N \l_@@_real_tl }
+ { \@@_parse_exponent: }
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Processing numbers}
+%
+% \begin{variable}
+% {
+% \l_@@_round_half_up_bool ,
+% \l_@@_round_min_tl ,
+% \l_@@_round_mode_tl ,
+% \l_@@_round_precision_int
+% }
+% \begin{macrocode}
+\keys_define:nn { siunitx }
+ {
+ round-half .choice: ,
+ round-half / even .code:n =
+ { \bool_set_false:N \l_@@_round_half_up_bool } ,
+ round-half / up .code:n =
+ { \bool_set_true:N \l_@@_round_half_up_bool } ,
+ round-minimum .tl_set:N =
+ \l_@@_round_min_tl ,
+ round-mode .choice: ,
+ round-mode / figures .code:n =
+ { \tl_set:Nn \_@@_round_mode_tl { figures } } ,
+ round-mode / none .code:n =
+ { \tl_set:Nn \_@@_round_mode_tl { none } } ,
+ round-mode / places .code:n =
+ { \tl_set:Nn \_@@_round_mode_tl { places } } ,
+ round-mode / uncertainty .code:n =
+ { \tl_set:Nn \_@@_round_mode_tl { uncertainty } } ,
+ round-precision .int_set:N =
+ \l_@@_round_precision_int ,
+ }
+\bool_new:N \l_@@_round_half_up_bool
+\tl_new:N \_@@_round_mode_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_round:}
+% \begin{macro}{\@@_round:N}
+% \begin{macro}{\@@_round_none:nNnnnNn}
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_round:
+ {
+ \@@_round:N \l_@@_real_tl
+ \@@_round:N \l_@@_imaginary_tl
+ }
+\cs_new_protected:Npn \@@_round:N #1
+ {
+ \tl_if_empty:NF #1
+ {
+ \tl_set:Nx #1
+ {
+ \cs:w
+ @@_round_ \_@@_round_mode_tl :nNnnnNn
+ \exp_after:wN
+ \cs_end: #1
+ }
+ }
+ }
+\cs_new:Npn \@@_round_none:nNnnnNn #1#2#3#4#5#6#7
+ { \exp_not:n { {#1} #2 {#3} {#4} {#5} #6 {#7} } }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}[EXP]{\@@_round_figures:nNnnnNn}
+% \begin{macrocode}
+\cs_new:Npn \@@_round_figures:nNnnnNn #1#2#3#4#5#6#7
+ {
+ \tl_if_blank:nTF {#5}
+ {
+ \int_compare:nNnTF \l_@@_round_precision_int > 0
+ { }
+ { { } + { 0 } { } { } + { 0 } }
+ }
+ { \exp_not:n { {#1} #2 {#3} {#4} {#5} #6 {#7} } }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}[EXP]{\@@_round_places:nNnnnNn}
+% \begin{macro}[EXP]
+% {\@@_round_places_decimal:nNnnNn, \@@_round_places_integer:nNnnNn}
+% The first step when rounding to a fixed number of places is to establish
+% if this is in the decimal or integer parts. The two require different
+% calculations for how many digits to drop from the input.
+% \begin{macrocode}
+\cs_new:Npn \@@_round_places:nNnnnNn #1#2#3#4#5#6#7
+ {
+ \tl_if_blank:nTF {#5}
+ {
+ \int_compare:nNnTF \l_@@_round_precision_int > 0
+ { \@@_round_places_decimal:nNnnNn }
+ { \@@_round_places_integer:nNnnNn }
+ {#1} #2 {#3} {#4} #6 {#7}
+ }
+ { \exp_not:n { {#1} #2 {#3} {#4} {#5} #6 {#7} } }
+ }
+\cs_new:Npn \@@_round_places_decimal:nNnnNn #1#2#3#4#5#6
+ { }
+\cs_new:Npn \@@_round_places_integer:nNnnNn #1#2#3#4#5#6
+ { }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}[EXP]{\@@_round_uncertainty:nNnnnNn}
+% \begin{macrocode}
+\cs_new:Npn \@@_round_uncertainty:nNnnnNn #1#2#3#4#5#6#7
+ {
+ \tl_if_blank:nTF {#5}
+ { \exp_not:n { {#1} #2 {#3} {#4} { } #6 {#7} } }
+ { }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{Formatting parsed numbers}
+%
+% \begin{variable}
+% {
+% \l_@@_bracket_negative_bool ,
+% \l_@@_bracket_close_tl ,
+% \l_@@_explicit_plus_bool ,
+% \l_@@_exponent_base_tl ,
+% \l_@@_exponent_product_tl ,
+% \l_@@_group_decimal_bool ,
+% \l_@@_group_integer_bool ,
+% \l_@@_group_minimum_int ,
+% \l_@@_group_separator_tl ,
+% \l_@@_negative_color_tl ,
+% \l_@@_bracket_open_tl ,
+% \l_@@_output_uncert_close_tl ,
+% \l_@@_output_complex_tl ,
+% \l_@@_output_decimal_tl ,
+% \l_@@_output_uncert_open_tl ,
+% \l_@@_uncert_separate_bool ,
+% \l_@@_tight_bool ,
+% \l_@@_unity_mantissa_bool ,
+% \l_@@_zero_exponent_bool
+% }
+% Keys producing tokens in the output.
+% \begin{macrocode}
+\keys_define:nn { siunitx }
+ {
+ bracket-negative .bool_set:N =
+ \l_@@_bracket_negative_bool ,
+ explicit-plus .bool_set:N =
+ \l_@@_explicit_plus_bool ,
+ exponent-base .tl_set:N =
+ \l_@@_exponent_base_tl ,
+ exponent-product .tl_set:N =
+ \l_@@_exponent_product_tl ,
+ group-digits .choice: ,
+ group-digits / all .code:n =
+ {
+ \bool_set_true:N \l_@@_group_decimal_bool
+ \bool_set_true:N \l_@@_group_integer_bool
+ } ,
+ group-digits / decimal .code:n =
+ {
+ \bool_set_true:N \l_@@_group_decimal_bool
+ \bool_set_false:N \l_@@_group_integer_bool
+ } ,
+ group-digits / integer .code:n =
+ {
+ \bool_set_false:N \l_@@_group_decimal_bool
+ \bool_set_true:N \l_@@_group_integer_bool
+ } ,
+ group-digits / none .code:n =
+ {
+ \bool_set_false:N \l_@@_group_decimal_bool
+ \bool_set_false:N \l_@@_group_integer_bool
+ } ,
+ group-digits .default:n = all ,
+ group-minimum-digits .int_set:N =
+ \l_@@_group_minimum_int ,
+ group-separator .tl_set:N =
+ \l_@@_group_separator_tl ,
+ negative-color .tl_set:N =
+ \l_@@_negative_color_tl ,
+ number-close-bracket .tl_set:N =
+ \l_@@_bracket_close_tl ,
+ number-open-bracket .tl_set:N =
+ \l_@@_bracket_open_tl ,
+ output-close-uncertainty .tl_set:N =
+ \l_@@_output_uncert_close_tl ,
+ output-complex-root .tl_set:N =
+ \l_@@_output_complex_tl ,
+ output-decimal-marker .tl_set:N =
+ \l_@@_output_decimal_tl ,
+ output-open-uncertainty .tl_set:N =
+ \l_@@_output_uncert_open_tl ,
+ separate-uncertainty .bool_set:N =
+ \l_@@_uncert_separate_bool ,
+ tight-spacing .bool_set:N =
+ \l_@@_tight_bool ,
+ unity-mantissa .bool_set:N =
+ \l_@@_unity_mantissa_bool ,
+ zero-exponent .bool_set:N =
+ \l_@@_zero_exponent_bool ,
+ }
+\bool_new:N \l_@@_group_decimal_bool
+\bool_new:N \l_@@_group_integer_bool
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_format:}
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format:
+ {
+ \tl_set:Nx \l_@@_formatted_tl
+ {
+ \tl_if_empty:NTF \l_@@_real_tl
+ {
+ \tl_if_empty:NF \l_@@_imaginary_tl
+ { \@@_format:N \l_@@_imaginary_tl }
+ }
+ {
+ \tl_if_empty:NTF \l_@@_imaginary_tl
+ { \@@_format:N \l_@@_real_tl }
+ { ??? }
+ }
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}[EXP]{\@@_format:N}
+% \begin{macro}[EXP]{\@@_format:nNnnnNn}
+% \begin{macro}[EXP]{\@@_format_comparator:n}
+% \begin{macro}[EXP]{\@@_format_sign:N, \@@_format_sign_aux:N}
+% \begin{macro}[EXP]
+% {\@@_format_sign_color:w, \@@_format_sign_brackets:w}
+% \begin{macro}[EXP]{\@@_format_integer:nnn}
+% \begin{macro}[EXP]{\@@_format_decimal:n, \@@_format_decimal:f}
+% \begin{macro}[EXP]{\@@_format_digits:nn}
+% \begin{macro}[EXP]{\@@_format_integer_aux:n}
+% \begin{macro}[EXP]
+% {
+% \@@_format_integer_aux_0:n,
+% \@@_format_integer_aux_1:n,
+% \@@_format_integer_aux_2:n
+% }
+% \begin{macro}[EXP]{\@@_format_decimal_aux:n}
+% \begin{macro}[EXP]{\@@_format_decimal_loop:NNNN}
+% \begin{macro}[EXP]{\@@_format_integer_first:nnNN}
+% \begin{macro}[EXP]{\@@_format_integer_loop:NNNN}
+% \begin{macro}[EXP]{\@@_format_uncertainty:nn}
+% \begin{macro}[EXP]{\@@_format_uncertainty_unaligned:}
+% \begin{macro}[EXP]
+% {\@@_format_uncertainty_aux:nn, \@@_format_uncertainty_aux:fn}
+% \begin{macro}[EXP]
+% {\@@_format_uncertainty:nnw, \@@_format_uncertainty:fnw}
+% \begin{macro}[EXP]{\@@_format_uncertainty:nw}
+% \begin{macro}[EXP]{\@@_format_exponent:Nnn}
+% \begin{macro}[EXP]{\@@_format_end:}
+% The approach to formatting a single number is to split into
+% the constituent parts. All of the parts are assembled including
+% inserting tabular alignment markers (which may be empty) for each
+% separate unit.
+% \begin{macrocode}
+\cs_new:Npn \@@_format:N #1
+ { \exp_after:wN \@@_format:nNnnnNn #1 }
+\cs_new:Npn \@@_format:nNnnnNn #1#2#3#4#5#6#7
+ {
+ \@@_format_comparator:n {#1}
+ \@@_format_sign:N #2
+ \@@_format_integer:nnn {#3} {#4} {#7}
+ \@@_format_decimal:n {#4}
+ \@@_format_uncertainty:nn {#5} {#4}
+ \@@_format_exponent:Nnn #6 {#7} { #3 . #4 }
+ \@@_format_end:
+ }
+% \end{macrocode}
+% To get the spacing correct this needs to be an ordinary math character.
+% \begin{macrocode}
+\cs_new:Npn \@@_format_comparator:n #1
+ {
+ \tl_if_blank:nF {#1}
+ { \exp_not:n { \mathord {#1} } }
+ \exp_not:V \l_@@_tab_tl
+ }
+% \end{macrocode}
+% Formatting signs has to deal with some additional formatting requirements
+% for negative numbers. Both making such numbers a fixed color and bracketing
+% them needs some rearrangement of the order of tokens, which is set up in
+% the main formatting macro by the dedicated do-nothing end function.
+% \begin{macrocode}
+\cs_new:Npn \@@_format_sign:N #1
+ {
+ \str_if_eq:nnTF {#1} { + }
+ {
+ \bool_if:NT \l_@@_explicit_plus_bool
+ { \@@_format_sign_aux:N #1 }
+ }
+ {
+ \str_if_eq:nnTF {#1} { - }
+ {
+ \tl_if_empty:NF \l_@@_negative_color_tl
+ { \@@_format_sign_color:w }
+ \bool_if:NTF \l_@@_bracket_negative_bool
+ { \@@_format_sign_brackets:w }
+ { \@@_format_sign_aux:N #1 }
+ }
+ { \@@_format_sign_aux:N #1 }
+ }
+ }
+\cs_new:Npn \@@_format_sign_aux:N #1
+ {
+ \bool_if:NTF \l_@@_tight_bool
+ { \exp_not:n { \mathord {#1} } }
+ { \exp_not:n {#1} }
+ }
+\cs_new:Npn
+ \@@_format_sign_color:w #1 \@@_format_end:
+ {
+ \exp_not:N \textcolor { \exp_not:V \l_@@_negative_color_tl }
+ {
+ #1
+ \@@_format_end:
+ }
+ }
+\cs_new:Npn
+ \@@_format_sign_brackets:w #1 \@@_format_end:
+ {
+ \exp_not:V \l_@@_bracket_open_tl
+ #1
+ \exp_not:V \l_@@_bracket_close_tl
+ \@@_format_end:
+ }
+% \end{macrocode}
+% Digit formatting leads off with separate functions to allow for a few
+% \enquote{up front} items before using a common set of tests for some common
+% cases. The code then splits again as the two types of grouping need
+% different strategies.
+% \begin{macrocode}
+\cs_new:Npn \@@_format_integer:nnn #1#2#3
+ {
+ \bool_lazy_all:nF
+ {
+ { \str_if_eq_p:nn {#1} { 1 } }
+ { \tl_if_blank_p:n {#2} }
+ { ! \str_if_eq_p:nn {#3} { 0 } }
+ { ! \l_@@_unity_mantissa_bool }
+ }
+ { \@@_format_digits:nn { integer } {#1} }
+ }
+\cs_new:Npn \@@_format_decimal:n #1
+ {
+ \exp_not:V \l_@@_tab_tl
+ \tl_if_blank:nF {#1}
+ { \exp_not:V \l_@@_output_decimal_tl }
+ \exp_not:V \l_@@_tab_tl
+ \@@_format_digits:nn { decimal } {#1}
+ }
+\cs_generate_variant:Nn \@@_format_decimal:n { f }
+\cs_new:Npn \@@_format_digits:nn #1#2
+ {
+ \bool_if:cTF { l_@@_group_ #1 _ bool }
+ {
+ \int_compare:nNnTF
+ { \tl_count:n {#2} } < \l_@@_group_minimum_int
+ { \exp_not:n {#2} }
+ { \use:c { @@_format_ #1 _aux:n } {#2} }
+ }
+ { \exp_not:n {#2} }
+ }
+% \end{macrocode}
+% For integers, we need to know how many digits there are to allow for the
+% correct insertion of separators. That is done using a two-part set up such
+% that there is no separator on the first pass.
+% \begin{macrocode}
+\cs_new:Npn \@@_format_integer_aux:n #1
+ {
+ \use:c
+ {
+ @@_format_integer_aux_
+ \int_eval:n { \int_mod:nn { \tl_count:n {#1} } { 3 } }
+ :n
+ } {#1}
+ }
+\cs_new:cpn { @@_format_integer_aux_0:n } #1
+ { \@@_format_integer_first:nnNN #1 \q_nil }
+\cs_new:cpn { @@_format_integer_aux_1:n } #1
+ { \@@_format_integer_first:nnNN { } { } #1 \q_nil }
+\cs_new:cpn { @@_format_integer_aux_2:n } #1
+ { \@@_format_integer_first:nnNN { } #1 \q_nil }
+\cs_new:Npn \@@_format_integer_first:nnNN #1#2#3#4
+ {
+ \exp_not:n {#1#2#3}
+ \quark_if_nil:NF #4
+ { \@@_format_integer_loop:NNNN #4 }
+ }
+\cs_new:Npn \@@_format_integer_loop:NNNN #1#2#3#4
+ {
+ \exp_not:V \l_@@_group_separator_tl
+ \exp_not:n {#1#2#3}
+ \quark_if_nil:NF #4
+ { \@@_format_integer_loop:NNNN #4 }
+ }
+% \end{macrocode}
+% For decimals, no need to do any counting, just loop using enough markers to
+% find the end of the list. By passing the decimal marker, it is possible not
+% to have to use a check on the content of the rest of the number. The
+% |\use_none:n(n)| mop up the remaining |\q_nil| tokens.
+% \begin{macrocode}
+\cs_new:Npn \@@_format_decimal_aux:n #1
+ {
+ \@@_format_decimal_loop:NNNN \c_empty_tl
+ #1 \q_nil \q_nil \q_nil
+ }
+\cs_new:Npn \@@_format_decimal_loop:NNNN #1#2#3#4
+ {
+ \quark_if_nil:NF #2
+ {
+ \exp_not:V #1
+ \exp_not:n {#2}
+ \quark_if_nil:NTF #3
+ { \use_none:n }
+ {
+ \exp_not:n {#3}
+ \quark_if_nil:NTF #4
+ { \use_none:nn }
+ {
+ \exp_not:n {#4}
+ \@@_format_decimal_loop:NNNN
+ \l_@@_group_separator_tl
+ }
+ }
+ }
+ }
+% \end{macrocode}
+% Uncertainties which are directly attached are easy to deal with. For those
+% that are separated, the first step is to find if they are entirely
+% contained within the decimal part, and to pad if they are. For the case
+% where the boundary is crossed to the integer part, the correct number of
+% digit tokens need to be removed from the start of the uncertainty and
+% the split result sent to the appropriate auxiliaries.
+% \begin{macrocode}
+\cs_new:Npn \@@_format_uncertainty:nn #1#2
+ {
+ \tl_if_blank:nTF {#1}
+ { \@@_format_uncertainty_unaligned: }
+ {
+ \bool_if:NTF \l_@@_uncert_separate_bool
+ {
+ \exp_not:V \l_@@_tab_tl
+ \@@_format_sign_aux:N \pm
+ \exp_not:V \l_@@_tab_tl
+ \@@_format_uncertainty_aux:fn
+ { \int_eval:n { \tl_count:n {#1} - \tl_count:n {#2} } }
+ {#1}
+ }
+ {
+ \exp_not:V \l_@@_output_uncert_open_tl
+ \exp_not:n {#1}
+ \exp_not:V \l_@@_output_uncert_close_tl
+ \@@_format_uncertainty_unaligned:
+ }
+ }
+ }
+\cs_new:Npn \@@_format_uncertainty_unaligned:
+ {
+ \exp_not:V \l_@@_tab_tl
+ \exp_not:V \l_@@_tab_tl
+ \exp_not:V \l_@@_tab_tl
+ \exp_not:V \l_@@_tab_tl
+ }
+\cs_new:Npn \@@_format_uncertainty_aux:nn #1#2
+ {
+ \int_compare:nNnTF {#1} > 0
+ {
+ \@@_format_uncertainty_aux:fnw
+ { \int_eval:n { #1 - 1 } }
+ { }
+ #2 \q_nil
+ }
+ {
+ 0
+ \@@_format_decimal:f
+ {
+ \prg_replicate:nn { \int_abs:n {#1} } { 0 }
+ #2
+ }
+ }
+ }
+\cs_generate_variant:Nn \@@_format_uncertainty_aux:nn { f }
+\cs_new:Npn \@@_format_uncertainty_aux:nnw #1#2#3
+ {
+ \quark_if_nil:NF #3
+ {
+ \int_compare:nNnTF {#1} = 0
+ { \@@_format_uncertainty_aux:nw {#2#3} }
+ {
+ \@@_format_uncertainty_aux:fnw
+ { \int_eval:n { #1 - 1 } }
+ {#2#3}
+ }
+ }
+ }
+\cs_generate_variant:Nn \@@_format_uncertainty_aux:nnw { f }
+\cs_new:Npn \@@_format_uncertainty_aux:nw #1#2 \q_nil
+ {
+ \@@_format_digits:nn { integer } {#1}
+ \@@_format_decimal:n {#2}
+ }
+% \end{macrocode}
+% Setting the exponent part requires some information about the mantissa:
+% was it there or not. This means that whilst only the sign and value for
+% the exponent are typeset here, there is a need to also have access to the
+% combined mantissa part (with a decimal marker). The rest of the work is
+% about picking up the various options and getting the combinations right.
+% For signs, the auxiliary from the main sign routine can be used, but not
+% the main function: negative exponents don't have special handling.
+% \begin{macrocode}
+\cs_new:Npn \@@_format_exponent:Nnn #1#2#3
+ {
+ \exp_not:V \l_@@_tab_tl
+ \bool_lazy_or:nnTF
+ { \l_@@_zero_exponent_bool }
+ { ! \str_if_eq_p:nn {#2} { 0 } }
+ {
+ \bool_lazy_and:nnTF
+ { \str_if_eq_p:nn {#3} { 1. } }
+ { ! \l_@@_unity_mantissa_bool }
+ { \exp_not:V \l_@@_tab_tl }
+ {
+ \bool_if:NTF \l_@@_tight_bool
+ {
+ \exp_not:N \mathord
+ { \exp_not:V \l_@@_exponent_product_tl }
+ }
+ { \exp_not:V \l_@@_exponent_product_tl }
+ \exp_not:V \l_@@_tab_tl
+ }
+ \exp_not:V \l_@@_exponent_base_tl
+ ^
+ {
+ \bool_lazy_or:nnT
+ { \l_@@_explicit_plus_bool }
+ { ! \str_if_eq_p:nn {#1} { + } }
+ { \@@_format_sign_aux:N #1 }
+ \@@_format_digits:nn { integer } {#2}
+ }
+ }
+ { \exp_not:V \l_@@_tab_tl }
+ }
+% \end{macrocode}
+% A do-nothing marker used to allow shuffling of the output and so expandable
+% operations for formatting.
+% \begin{macrocode}
+\cs_new:Npn \@@_format_end: { }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Miscellaneous tools}
+%
+% \begin{variable}{\l_@@_valid_tl}
+% The list of valid tokens.
+% \begin{macrocode}
+\tl_new:N \l_@@_valid_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}[TF]{\siunitx_if_number:n}
+% Test if an entire number is valid: this means parsing the number but not
+% returning anything.
+% \begin{macrocode}
+\prg_new_protected_conditional:Npnn \siunitx_if_number:n #1
+ { T , F , TF }
+ {
+ \group_begin:
+ \bool_set_true:N \l_@@_validate_bool
+ \@@_parse:n {#1}
+ \bool_lazy_and:nnTF
+ { \tl_if_empty_p:N \l_@@_real_tl }
+ { \tl_if_empty_p:N \l_@@_imaginary_tl }
+ {
+ \group_end:
+ \prg_return_false:
+ }
+ {
+ \group_end:
+ \prg_return_true:
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}[TF]{\siunitx_if_number_token:N}
+% A simple conditional to answer the question of whether a specific token is
+% possibly valid in a number.
+% \begin{macrocode}
+\prg_new_protected_conditional:Npnn \siunitx_if_number_token:N #1
+ { T , F , TF }
+ {
+ \tl_set:Nx \l_@@_valid_tl
+ {
+ \exp_not:V \l_@@_input_uncert_close_tl
+ \exp_not:V \l_@@_input_complex_tl
+ \exp_not:V \l_@@_input_comparator_tl
+ \exp_not:V \l_@@_input_decimal_tl
+ \exp_not:V \l_@@_input_digit_tl
+ \exp_not:V \l_@@_input_exponent_tl
+ \exp_not:V \l_@@_input_ignore_tl
+ \exp_not:V \l_@@_input_uncert_open_tl
+ \exp_not:V \l_@@_input_sign_tl
+ \exp_not:V \l_@@_input_uncert_sign_tl
+ }
+ \tl_if_in:VnTF \l_@@_valid_tl {#1}
+ { \prg_return_true: }
+ { \prg_return_false: }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{Messages}
+%
+% \begin{macrocode}
+\msg_new:nnnn { siunitx } { number / invalid-input }
+ { Invalid~number~'#1'. }
+ {
+ The~input~'#1'~could~not~be~parsed~as~a~number~following~the~
+ format~defined~in~module~documentation.
+ }
+% \end{macrocode}
+%
+% \subsection{Standard settings for module options}
+%
+% Some of these follow naturally from the point of definition
+% (\emph{e.g.}~boolean variables are always |false| to begin with),
+% but for clarity everything is set here.
+% \begin{macrocode}
+\keys_set:nn { siunitx }
+ {
+ bracket-negative = false ,
+ evaluate-expression = false ,
+ explicit-plus = false ,
+ exponent-base = 10 ,
+ exponent-product = \times ,
+ expression = #1 ,
+ group-digits = all ,
+ group-minimum-digits = 4 ,
+ group-separator = \, , % (
+ input-close-uncertainty = ) ,
+ input-complex-roots = ij ,
+ input-comparators = { <=>\approx\ge\geq\gg\le\leq\ll\sim } ,
+ input-decimal-markers = { ., } ,
+ input-digits = 0123456789 ,
+ input-exponent-markers = dDeE ,
+ input-ignore = \, ,
+ input-open-uncertainty = ( , % )
+ input-signs = +-\mp\pm ,
+ input-uncertainty-signs = \pm ,
+ negative-color = , % (
+ number-close-bracket = ) ,
+ number-open-bracket = ( , % )
+ output-close-uncertainty = ) ,
+ output-complex-root = \mathrm { i } ,
+ output-decimal-marker = . ,
+ output-open-uncertainty = ( , % )
+ round-half = up ,
+ round-minimum = 0 ,
+ round-mode = none ,
+ round-precision = 2 ,
+ separate-uncertainty = false ,
+ tight-spacing = false ,
+ unity-mantissa = false ,
+ zero-exponent = false
+ }
+% \end{macrocode}
+%
+% \begin{macrocode}
+%</package>
+% \end{macrocode}
+%
+% \end{implementation}
+%
+% \PrintIndex
Property changes on: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-number.dtx
___________________________________________________________________
Added: svn:eol-style
## -0,0 +1 ##
+native
\ No newline at end of property
Added: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-print.dtx
===================================================================
--- trunk/Master/texmf-dist/source/latex/siunitx/siunitx-print.dtx (rev 0)
+++ trunk/Master/texmf-dist/source/latex/siunitx/siunitx-print.dtx 2018-04-29 22:36:26 UTC (rev 47489)
@@ -0,0 +1,767 @@
+% \iffalse meta-comment
+%
+% File: siunitx-print.dtx Copyright (C) 2016-2017 Joseph Wright
+%
+% It may be distributed and/or modified under the conditions of the
+% LaTeX Project Public License (LPPL), either version 1.3c of this
+% license or (at your option) any later version. The latest version
+% of this license is in the file
+%
+% https://www.latex-project.org/lppl.txt
+%
+% This file is part of the "siunitx bundle" (The Work in LPPL)
+% and all files in that bundle must be distributed together.
+%
+% The released version of this bundle is available from CTAN.
+%
+% -----------------------------------------------------------------------
+%
+% The development version of the bundle can be found at
+%
+% https://github.com/josephwright/siunitx
+%
+% for those people who are interested.
+%
+% -----------------------------------------------------------------------
+%
+%<*driver>
+\documentclass{l3doc}
+% The next line is needed so that \GetFileInfo will be able to pick up
+% version data
+\usepackage{siunitx}
+\begin{document}
+ \DocInput{\jobname.dtx}
+\end{document}
+%</driver>
+% \fi
+%
+% \GetFileInfo{siunitx.sty}
+%
+% \title{^^A
+% \pkg{siunitx-print} -- Printing material with font control^^A
+% \thanks{This file describes \fileversion,
+% last revised \filedate.}^^A
+% }
+%
+% \author{^^A
+% Joseph Wright^^A
+% \thanks{^^A
+% E-mail:
+% \href{mailto:joseph.wright at morningstar2.co.uk}
+% {joseph.wright at morningstar2.co.uk}^^A
+% }^^A
+% }
+%
+% \date{Released \filedate}
+%
+% \maketitle
+%
+% \begin{documentation}
+%
+% This submodule is focussed on providing controlled printing for numbers and
+% units. Key to this is control of font: conventions for printing quantities
+% mean that the exact nature of the output is important. At the same time, this
+% module provides flexibility for the user in terms of which aspects of the font
+% are responsive to the surrounding general text. Printing material may also
+% take place in text or math mode.
+%
+% The printing routines assume that normal \LaTeXe{} font selection commands
+% are available, in particular \cs{bfsries}, \cs{mathrm}, \cs{mathversion},
+% \cs{mdseries}, \cs{rmfamily} and \cs{upshape}. It also requires the standard
+% \LaTeXe{} kernel commands \cs{ensuremath}, \cs{textsubscript} and
+% \cs{textsuperscript} for printing in text mode. The following packages are
+% also required to provide the functionality detailed.
+% \begin{itemize}
+% \item \pkg{color}: support for color using \cs{textcolor}
+% \item \pkg{textcomp}: \cs{textminus} and \cs{textpm} for printing
+% in text mode
+% \item \pkg{amstext}: the \cs{text} command for printing in text mode
+% \end{itemize}
+%
+% \begin{function}{\siunitx_print:nn, \siunitx_print:nV}
+% \begin{syntax}
+% \cs{siunitx_print:nn} \Arg{type} \Arg{material}
+% \end{syntax}
+% Prints the \meta{material} according the the prevailing settings for the
+% submodule as applicable to the \meta{type} of content: the latter should be
+% either |number| or |unit|. The \meta{material} should comprise normal
+% \LaTeX{} mark-up for numbers or units. In particular, units will typically
+% use |\mathrm| to indicate material to be printed in the current upright
+% roman font, and |^| and |_| will typically be used to indicate super- and
+% subscripts, respectively. These elements will be correctly handled when
+% printing for example using |\mathsf| in math mode, or using only text
+% fonts.
+% \end{function}
+%
+% \begin{variable}{\l_siunitx_print_series_prop}
+% This properly list contains mappings from text mode font weights to math
+% mode font versions. The standard settings here cover
+% \begin{itemize}
+% \item[\texttt{m}] Standard (mid) weight, mapping to math version
+% |normal|.
+% \item[\texttt{bx}] Bold exanded weight, mapping to math version |bold|.
+% \item[\texttt{lt}] Light weight, mapping to math version |light|.
+% \item[\texttt{l}] Light weight, mapping to math version |light|.
+% \end{itemize}
+% \end{variable}
+%
+% \subsection{Key--value options}
+%
+% The options defined by this submodule are available within the \pkg{l3keys}
+% |siunitx| tree.
+%
+% \begin{function}{color}
+% \begin{syntax}
+% |color| = \meta{color}
+% \end{syntax}
+% Color to apply to printed output: the latter should be a named color
+% defined for use with \cs{textcolor}. The standard setting is empty (no
+% color).
+% \end{function}
+%
+% \begin{function}{mode}
+% \begin{syntax}
+% |mode| = \meta{choice}
+% \end{syntax}
+% Selects which mode (math or text) the output is printed in: a choice
+% from the options |match|, |math| or |text|. The option |match| matches
+% the mode prevailing at the point \cs{siunitx_print:nn} is called. The
+% |math| and |text| options choose the relevant \TeX{} mode for printing.
+% The standard setting is |math|.
+% \end{function}
+%
+% \begin{function}{number-color}
+% \begin{syntax}
+% |number-color| = \meta{color}
+% \end{syntax}
+% Color to apply to numbers in output: the latter should be a named color
+% defined for use with \cs{textcolor}. The standard setting is empty (no
+% color).
+% \end{function}
+%
+% \begin{function}{number-mode}
+% \begin{syntax}
+% |number-mode| = \meta{choice}
+% \end{syntax}
+% Selects which mode (math or text) the numbers are printed in: a choice
+% from the options |match|, |math| or |text|. The option |match| matches
+% the mode prevailing at the point \cs{siunitx_print:nn} is called. The
+% |math| and |text| options choose the relevant \TeX{} mode for printing.
+% The standard setting is |math|.
+% \end{function}
+%
+% \begin{function}{propagate-math-font}
+% \begin{syntax}
+% |propagate-math-font| = |true|\verb"|"|false|
+% \end{syntax}
+% Switch to determine if the currently-active math font is applied within
+% printed output. This is relevant only when \cs{siunitx_print:nn} is
+% called from within math mode: in text mode there is not active math
+% font. When not active, math mode material will be typeset using
+% standard math mode fonts without any changes being made to the
+% supplied argument. The standard setting is |false|.
+% \end{function}
+%
+% \begin{function}{reset-math-version}
+% \begin{syntax}
+% |reset-math-version| = |true|\verb"|"|false|
+% \end{syntax}
+% Switch to determine whether the active \cs{mathversion} is reset to
+% |normal| when printing in math mode. Note that math version is typically
+% used to select \cs{boldmath}, though it is also be used by
+% \foreign{e.g.}~\pkg{sansmath}. The standard setting is |true|.
+% \end{function}
+%
+% \begin{function}{reset-text-family}
+% \begin{syntax}
+% |reset-text-family| = |true|\verb"|"|false|
+% \end{syntax}
+% Switch to determine whether the active math family is reset to
+% \cs{rmfamily} when printing in text mode. The standard setting is |true|.
+% \end{function}
+%
+% \begin{function}{reset-text-series}
+% \begin{syntax}
+% |reset-text-series| = |true|\verb"|"|false|
+% \end{syntax}
+% Switch to determine whether the active math series is reset to
+% \cs{mdseries} when printing in text mode. The standard setting is |true|.
+% \end{function}
+%
+% \begin{function}{reset-text-shape}
+% \begin{syntax}
+% |reset-text-shape| = |true|\verb"|"|false|
+% \end{syntax}
+% Switch to determine whether the active math shape is reset to
+% \cs{upshape} when printing in text mode. The standard setting is |true|.
+% \end{function}
+%
+% \begin{function}{text-family-to-math}
+% \begin{syntax}
+% |text-family-to-math| = |true|\verb"|"|false|
+% \end{syntax}
+% Switch to determine if the family of the current text font should be
+% applied (where possible) to printing in math mode. The standard setting is
+% |false|.
+% \end{function}
+%
+% \begin{function}{text-weight-to-math}
+% \begin{syntax}
+% |text-weight-to-math| = |true|\verb"|"|false|
+% \end{syntax}
+% Switch to determine if the weight of the current text font should be
+% applied (where possible) to printing in math mode. This is achieved by
+% setting the \cs{mathversion}, and so will override |reset-math-version|.
+% The mappings between text and math weight are stored in
+% \cs{l_siunitx_print_series_prop}. The standard setting is |false|.
+% \end{function}
+%
+% \begin{function}{unit-color}
+% \begin{syntax}
+% |unit-color| = \meta{color}
+% \end{syntax}
+% Color to apply to units in output: the latter should be a named color
+% defined for use with \cs{textcolor}. The standard setting is empty (no
+% color).
+% \end{function}
+%
+% \begin{function}{unit-mode}
+% \begin{syntax}
+% |unit-mode| = \meta{choice}
+% \end{syntax}
+% Selects which mode (math or text) units are printed in: a choice
+% from the options |match|, |math| or |text|. The option |match| matches
+% the mode prevailing at the point \cs{siunitx_print:nn} is called. The
+% |math| and |text| options choose the relevant \TeX{} mode for printing.
+% The standard setting is |math|.
+% \end{function}
+%
+% \end{documentation}
+%
+% \begin{implementation}
+%
+% \section{\pkg{siunitx-print} implementation}
+%
+% Start the \pkg{DocStrip} guards.
+% \begin{macrocode}
+%<*package>
+% \end{macrocode}
+%
+% Identify the internal prefix (\LaTeX3 \pkg{DocStrip} convention): only
+% internal material in this \emph{submodule} should be used directly.
+% \begin{macrocode}
+%<@@=siunitx_print>
+% \end{macrocode}
+%
+% \subsection{Initial set up}
+%
+% The printing routines depend on \pkg{amstext} for text mode working.
+% \begin{macrocode}
+\RequirePackage { amstext }
+% \end{macrocode}
+%
+% Color support is always required.
+% \begin{macrocode}
+\RequirePackage { color }
+% \end{macrocode}
+%
+% For a sensible \cs{textminus} we load \pkg{textcomp} if \pkg{fontspec}
+% is not in use.
+% \begin{macrocode}
+\AtBeginDocument
+ {
+ \@ifpackageloaded { fontspec }
+ { }
+ { \RequirePackage { textcomp } }
+ }
+% \end{macrocode}
+%
+% \begin{variable}{\l_@@_tmp_box}
+% Scratch space.
+% \begin{macrocode}
+\box_new:N \l_@@_tmp_box
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\document}
+% \begin{macro}{\@@_store_fam:n}
+% \begin{variable}{\c_@@_mathrm_int, \c_@@_mathsf_int, \c_@@_mathtt_int}
+% In order to test math fonts, we need information about the \cs{fam} used
+% by the various options. This is run as a hook onto \cs{document}, rather
+% than using \tn{AtBeginDocument} as it has to come after anything that
+% \pkg{fontspec} does (nasty errors arise otherwise). As this is a true
+% one-off, we avoid wasting a box.
+% \begin{macrocode}
+\tl_put_right:Nn \document
+ {
+ \@@_store_fam:n { rm }
+ \@@_store_fam:n { sf }
+ \@@_store_fam:n { tt }
+ \ignorespaces
+ }
+\cs_new_protected:Npn \@@_store_fam:n #1
+ {
+ \group_begin:
+ \hbox_set:Nn \l_@@_tmp_box
+ {
+ \ensuremath
+ {
+ \use:c { math #1 }
+ { \int_const:cn { c_@@_math #1 _int } { \fam } }
+ }
+ }
+ \group_end:
+ }
+% \end{macrocode}
+% \end{variable}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Printing routines}
+%
+% \begin{variable}
+% {
+% \l_@@_number_color_tl ,
+% \l_@@_number_mode_tl ,
+% \l_@@_unit_color_tl ,
+% \l_@@_unit_mode_tl ,
+% \l_@@_math_font_bool ,
+% \l_@@_math_version_bool ,
+% \l_@@_math_family_bool ,
+% \l_@@_math_weight_bool ,
+% \l_@@_text_family_tl ,
+% \l_@@_text_series_tl ,
+% \l_@@_text_shape_tl
+% }
+% Options which apply to the main formatting routine, and so are not tied
+% to either symbolic or literal input.
+% \begin{macrocode}
+\tl_new:N \l_@@_number_mode_tl
+\tl_new:N \l_@@_unit_mode_tl
+\keys_define:nn { siunitx }
+ {
+ color .meta:n =
+ { number-color = #1 , unit-color = #1 } ,
+ mode .meta:n =
+ { number-mode = #1 , unit-mode = #1 } ,
+ number-color .tl_set:N =
+ \l_@@_number_color_tl ,
+ number-mode .choices:nn =
+ { match , math , text }
+ {
+ \tl_set_eq:NN
+ \l_@@_number_mode_tl \l_keys_choice_tl
+ } ,
+ propagate-math-font .bool_set:N =
+ \l_@@_math_font_bool ,
+ reset-math-version .bool_set:N =
+ \l_@@_math_version_bool ,
+ reset-text-family .bool_set:N =
+ \l_@@_text_family_bool ,
+ reset-text-series .bool_set:N =
+ \l_@@_text_series_bool ,
+ reset-text-shape .bool_set:N =
+ \l_@@_text_shape_bool ,
+ text-family-to-math .bool_set:N =
+ \l_@@_math_family_bool ,
+ text-weight-to-math .bool_set:N =
+ \l_@@_math_weight_bool ,
+ unit-color .tl_set:N =
+ \l_@@_unit_color_tl ,
+ unit-mode .choices:nn =
+ { match , math , text }
+ {
+ \tl_set_eq:NN
+ \l_@@_unit_mode_tl \l_keys_choice_tl
+ }
+ }
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\siunitx_print:nn, \siunitx_print:nV}
+% The main printing function doesn't actually need to do very much: just set
+% the color and select the correct sub-function.
+% \begin{macrocode}
+\cs_new_protected:Npn \siunitx_print:nn #1#2
+ {
+ \tl_if_empty:cTF { l_@@_ #1 _color_tl }
+ { \use:n }
+ { \exp_args:Nv \textcolor { l_@@_ #1 _color_tl } }
+ {
+ \use:c
+ {
+ @@_
+ \tl_use:c { l_@@_ #1 _mode_tl } :n
+ }
+ {#2}
+ }
+ }
+\cs_generate_variant:Nn \siunitx_print:nn { nV }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_match:n}
+% When the \emph{output} mode should match the input, a simple selection of
+% route can be made.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_match:n #1
+ {
+ \mode_if_math:TF
+ { \@@_math:n {#1} }
+ { \@@_text:n {#1} }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{variable}{\l_siunitx_print_series_prop}
+% Mapping data to relate text series to math version.
+% \begin{macrocode}
+\prop_new:N \l_siunitx_print_series_prop
+\prop_put:Nnn \l_siunitx_print_series_prop { m } { normal }
+\prop_put:Nnn \l_siunitx_print_series_prop { bx } { bold }
+\prop_put:Nnn \l_siunitx_print_series_prop { l } { light }
+\prop_put:Nnn \l_siunitx_print_series_prop { lt } { light }
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_math:n}
+% \begin{macro}{\@@_math_version:nn, \@@_math_version:Vn}
+% \begin{macro}
+% {
+% \@@_math_auxi:n,
+% \@@_math_auxii:n,
+% \@@_math_auxiii:n,
+% \@@_math_auxiv:n,
+% \@@_math_auxv:n
+% }
+% \begin{macro}{\@@_math_aux:Nn, \@@_math_aux:cn}
+% \begin{macro}{\@@_math_sub:n, \@@_math_super:n, \@@_math_script:n}
+% \begin{macro}{\@@_math_text:n}
+% The first step in setting in math mode is to check on the math version.
+% The starting point is the question of whether text series needs to
+% propagate to math mode: if so, check on the mapping, otherwise check on
+% the current math version.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_math:n #1
+ {
+ \bool_if:NTF \l_@@_math_weight_bool
+ {
+ \prop_get:NVNTF \l_siunitx_print_series_prop
+ \f at series \l_@@_tmp_tl
+ { \@@_math_version:Vn \l_@@_tmp_tl {#1} }
+ { \@@_math_auxi:n {#1} }
+ }
+ { \@@_math_auxi:n {#1} }
+ }
+\cs_new_protected:Npn \@@_math_auxi:n #1
+ {
+ \bool_if:NTF \l_@@_math_version_bool
+ { \@@_math_version:nn { normal } {#1} }
+ { \@@_math_auxii:n {#1} }
+ }
+% \end{macrocode}
+% Any setting which changes the math version can only be set from text mode
+% (as it applies at the level of a formula). As such, the first test is to
+% see if that needs to be to check if the math version has to be set: if so,
+% switch to text mode, sort it out and switch back. That of course means
+% that in such cases, line breaking will not be possible.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_math_version:nn #1#2
+ {
+ \str_if_eq_x:nnTF { \math at version } { #1 }
+ { \@@_math_auxii:n {#2} }
+ {
+ \mode_if_math:TF
+ { \text }
+ { \use:n }
+ {
+ \mathversion {#1}
+ \@@_math_auxii:n {#2}
+ }
+ }
+ }
+\cs_generate_variant:Nn \@@_math_version:nn { V }
+% \end{macrocode}
+% At this point, force math mode then start dealing with setting math font
+% based on text family. If the text family is roman, life is slightly
+% different to if it is sanserif or monospaced. In all cases, the outcomes
+% can be handled using the same routines as for normal math mode treatment.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_math_auxii:n #1
+ { \ensuremath { \@@_math_auxiii:n {#1} } }
+\cs_new_protected:Npn \@@_math_auxiii:n #1
+ {
+ \bool_if:NTF \l_@@_math_family_bool
+ {
+ \str_case_x:nnF { \f at family }
+ {
+ { \rmdefault } { \@@_math_auxv:n }
+ { \sfdefault } { \@@_math_aux:Nn \mathsf }
+ { \ttdefault } { \@@_math_aux:Nn \mathtt }
+ }
+ { \@@_math_auxiv:n }
+ }
+ { \@@_math_auxiv:n }
+ {#1}
+ }
+% \end{macrocode}
+% Now we deal with the font selection in math mode. There are two possible
+% cases. First, we are retaining the current math font, and the active one is
+% \cs{mathsf} or \cs{mathtt}: that needs to be applied to the argument.
+% Alternatively, if the current font is not retained, ensure that
+% normal math mode rules are active. The parts here are split up to allow
+% reuse when picking up the text family.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_math_auxiv:n #1
+ {
+ \bool_if:NTF \l_@@_math_font_bool
+ {
+ \int_case:nnF \fam
+ {
+ \c_@@_mathsf_int { \@@_math_aux:Nn \mathsf }
+ \c_@@_mathtt_int { \@@_math_aux:Nn \mathtt }
+ }
+ { \use:n }
+ }
+ { \@@_math_auxv:n }
+ {#1}
+ }
+\cs_new_protected:Npn \@@_math_auxv:n #1
+ {
+ \bool_lazy_or:nnTF
+ { \int_compare_p:nNn \fam = { -1 } }
+ { \int_compare_p:nNn \fam = \c_@@_mathrm_int }
+ { \use:n }
+ { \mathrm }
+ {#1}
+ }
+% \end{macrocode}
+% Search-and-replace fun: deal with any \cs{mathrm} in the argument and
+% also inside sub/superscripts.
+% \begin{macrocode}
+\cs_new_protected:Npx \@@_math_aux:Nn #1#2
+ {
+ \group_begin:
+ \tl_set:Nn \exp_not:N \l_@@_tmp_tl {#2}
+ \tl_replace_all:Nnn \exp_not:N \l_@@_tmp_tl
+ { \exp_not:N \mathrm } { \exp_not:N \use:n }
+ \tl_replace_all:Nnn \exp_not:N \l_@@_tmp_tl
+ { \char_generate:nn { `\_ } { 8 } }
+ { \exp_not:N \@@_math_sub:n }
+ \tl_replace_all:Nnn \exp_not:N \l_@@_tmp_tl
+ { ^ }
+ { \exp_not:N \@@_math_super:n }
+ #1 { \exp_not:N \tl_use:N \exp_not:N \l_@@_tmp_tl }
+ \group_end:
+ }
+\cs_generate_variant:Nn \@@_math_aux:Nn { c }
+\cs_new_protected:Npx \@@_math_sub:n #1
+ {
+ \char_generate:nn { `\_ } { 8 }
+ { \exp_not:N \@@_math_script:n {#1} }
+ }
+\cs_new_protected:Npn \@@_math_super:n #1
+ { ^ { \@@_math_script:n {#1} } }
+\cs_new_protected:Npn \@@_math_script:n #1
+ {
+ \group_begin:
+ \tl_set:Nn \l_@@_tmp_tl {#1}
+ \tl_replace_all:Nnn \l_@@_tmp_tl { \mathrm } { \use:n }
+ \tl_use:N \l_@@_tmp_tl
+ \group_end:
+ }
+% \end{macrocode}
+% To match the text mode font, there is a simple look-up of the current
+% font family. Luckily, mappings to math mode equivalents are easy.
+% \begin{macrocode}
+
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_text:n}
+% \begin{macro}{\@@_text_replace:n}
+% \begin{macro}{\@@_text_replace:N}
+% \begin{macro}{\@@_text_replace:NNn}
+% \begin{macro}{\@@_text_sub:n, \@@_text_super:n}
+% \begin{macro}{\@@_text_scripts:NnN}
+% \begin{macro}{\@@_text_scripts:}
+% \begin{macro}{\@@_text_scripts_one:NnN}
+% \begin{macro}{\@@_text_scripts_two:NnNn}
+% \begin{macro}{\@@_text_scripts_two:nn}
+% \begin{macro}{\@@_text_scripts_two:n}
+% Typesetting in text mode is easy in font control terms but more tricky
+% in the manipulation of the input. The easy part comes first.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_text:n #1
+ {
+ \text
+ {
+ \bool_if:NT \l_@@_text_family_bool
+ { \rmfamily }
+ \bool_if:NT \l_@@_text_series_bool
+ { \mdseries }
+ \bool_if:NT \l_@@_text_shape_bool
+ { \upshape }
+ \@@_text_replace:n {#1}
+ }
+ }
+% \end{macrocode}
+% To get math mode material to print in text mode, various search-and-replace
+% steps are needed. The \cs{mathrm} command is also disabled as it is
+% invalid in text mode.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_text_replace:n #1
+ {
+ \group_begin:
+ \tl_set:Nn \l_@@_tmp_tl {#1}
+ \@@_text_replace:N \l_@@_tmp_tl
+ \tl_use:N \l_@@_tmp_tl
+ \group_end:
+ }
+\cs_new_protected:Npx \@@_text_replace:N #1
+ {
+ \exp_not:N \@@_text_replace:NNn #1
+ \exp_not:N \mathrm
+ { \exp_not:N \use:n }
+ \exp_not:N \pm
+ { \exp_not:N \textpm }
+ \exp_not:N \mp
+ { \exp_not:n { \ensuremath { \mp } } }
+ -
+ { \exp_not:N \textminus }
+ \char_generate:nn { `\_ } { 8 }
+ { \exp_not:N \@@_text_sub:n }
+ ^
+ { \exp_not:N \@@_text_super:n }
+ \exp_not:N \q_recursion_tail
+ { ? }
+ \exp_not:N \q_recursion_stop
+ }
+\cs_new_protected:Npn \@@_text_replace:NNn #1#2#3
+ {
+ \quark_if_recursion_tail_stop:N #2
+ \tl_replace_all:Nnn #1 {#2} {#3}
+ \@@_text_replace:NNn #1
+ }
+% \end{macrocode}
+% Sub- and superscripts can be in any order in the source. The first step
+% of handling them is therefore to do a look-ahead to work out whether
+% only one or both are present.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_text_sub:n #1
+ {
+ \@@_text_scripts:NnN
+ \textsubscript {#1} \@@_text_super:n
+ }
+\cs_new_protected:Npn \@@_text_super:n #1
+ {
+ \@@_text_scripts:NnN
+ \textsuperscript {#1} \@@_text_sub:n
+ }
+\cs_new_protected:Npn \@@_text_scripts:NnN #1#2#3
+ {
+ \cs_set_protected:Npn \@@_text_scripts:
+ {
+ \if_meaning:w \l_peek_token #3
+ \exp_after:wN \@@_text_scripts_two:NnNn
+ \else:
+ \exp_after:wN \@@_text_scripts_one:Nn
+ \fi:
+ #1 {#2}
+ }
+ \peek_after:Nw \@@_text_scripts:
+ }
+\cs_new_protected:Npn \@@_text_scripts: { }
+% \end{macrocode}
+% In the simple case of one script item, we have to do a search-and-replace
+% to deal with anything inside the argument.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_text_scripts_one:Nn #1#2
+ {
+ \group_begin:
+ \tl_set:Nn \l_@@_tmp_tl {#2}
+ \@@_text_replace:N \l_@@_tmp_tl
+ \exp_args:NNV \group_end:
+ #1 \l_@@_tmp_tl
+ }
+% \end{macrocode}
+% For the two scripts case, we cannot use |\textsubscript|/|\textsuperscript|
+% as they don't stack directly. Instead, we sort out the ordering then use
+% an implementation for both parts that is the same as the kernel text
+% scripts.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_text_scripts_two:NnNn #1#2#3#4
+ {
+ \cs_if_eq:NNTF #1 \textsubscript
+ { \@@_text_scripts_two:nn {#4} {#2} }
+ { \@@_text_scripts_two:nn {#2} {#4} }
+ }
+\cs_new_protected:Npx \@@_text_scripts_two:nn #1#2
+ {
+ \group_begin:
+ \exp_not:N \m at th
+ \exp_not:N \ensuremath
+ {
+ ^ { \exp_not:N \@@_text_scripts_two:n {#1} }
+ \char_generate:nn { `\_ } { 8 }
+ { \exp_not:N \@@_text_scripts_two:n {#2} }
+ }
+ \group_end:
+ }
+\cs_new_protected:Npn \@@_text_scripts_two:n #1
+ {
+ \mbox
+ {
+ \fontsize \sf at size \z@ \selectfont
+ \@@_text_scripts_one:Nn \use:n {#1}
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Standard settings for module options}
+%
+% Some of these follow naturally from the point of definition
+% (\foreign{e.g.}~boolean variables are always |false| to begin with),
+% but for clarity everything is set here.
+% \begin{macrocode}
+\keys_set:nn { siunitx }
+ {
+ color = ,
+ mode = math ,
+ number-color = ,
+ number-mode = math ,
+ propagate-math-font = false ,
+ reset-math-version = true ,
+ reset-text-shape = true ,
+ reset-text-series = true ,
+ reset-text-family = true ,
+ text-family-to-math = false ,
+ text-weight-to-math = false ,
+ unit-color = ,
+ unit-mode = math
+ }
+% \end{macrocode}
+%
+% \begin{macrocode}
+%</package>
+% \end{macrocode}
+%
+% \end{implementation}
+%
+% \PrintIndex
\ No newline at end of file
Property changes on: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-print.dtx
___________________________________________________________________
Added: svn:eol-style
## -0,0 +1 ##
+native
\ No newline at end of property
Added: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-table.dtx
===================================================================
--- trunk/Master/texmf-dist/source/latex/siunitx/siunitx-table.dtx (rev 0)
+++ trunk/Master/texmf-dist/source/latex/siunitx/siunitx-table.dtx 2018-04-29 22:36:26 UTC (rev 47489)
@@ -0,0 +1,606 @@
+% \iffalse meta-comment
+%
+% File: siunitx-table.dtx Copyright (C) 2016-2017 Joseph Wright
+%
+% It may be distributed and/or modified under the conditions of the
+% LaTeX Project Public License (LPPL), either version 1.3c of this
+% license or (at your option) any later version. The latest version
+% of this license is in the file
+%
+% https://www.latex-project.org/lppl.txt
+%
+% This file is part of the "siunitx bundle" (The Work in LPPL)
+% and all files in that bundle must be distributed together.
+%
+% The released version of this bundle is available from CTAN.
+%
+% -----------------------------------------------------------------------
+%
+% The development version of the bundle can be found at
+%
+% https://github.com/josephwright/siunitx
+%
+% for those people who are interested.
+%
+% -----------------------------------------------------------------------
+%
+%<*driver>
+\documentclass{l3doc}
+% The next line is needed so that \GetFileInfo will be able to pick up
+% version data
+\usepackage{siunitx}
+\begin{document}
+ \DocInput{\jobname.dtx}
+\end{document}
+%</driver>
+% \fi
+%
+% \GetFileInfo{siunitx.sty}
+%
+% \title{^^A
+% \pkg{siunitx-table} -- Formatting numbers in tables^^A
+% \thanks{This file describes \fileversion,
+% last revised \filedate.}^^A
+% }
+%
+% \author{^^A
+% Joseph Wright^^A
+% \thanks{^^A
+% E-mail:
+% \href{mailto:joseph.wright at morningstar2.co.uk}
+% {joseph.wright at morningstar2.co.uk}^^A
+% }^^A
+% }
+%
+% \date{Released \filedate}
+%
+% \maketitle
+%
+% \begin{documentation}
+%
+% \end{documentation}
+%
+% \begin{implementation}
+%
+% \section{\pkg{siunitx-table} implementation}
+%
+% Start the \pkg{DocStrip} guards.
+% \begin{macrocode}
+%<*package>
+% \end{macrocode}
+%
+% Identify the internal prefix (\LaTeX3 \pkg{DocStrip} convention): only
+% internal material in this \emph{submodule} should be used directly.
+% \begin{macrocode}
+%<@@=siunitx_table>
+% \end{macrocode}
+%
+% \begin{variable}{\l_@@_tmp_tl}
+% Scratch space.
+% \begin{macrocode}
+\tl_new:N \l_@@_tmp_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \subsection{Interface functions}
+%
+% \begin{variable}{\l_@@_parse_bool}
+% One top-level setting applies to all number cells.
+% \begin{macrocode}
+\keys_define:nn { siunitx / table }
+ {
+ parse .bool_set:N = \l_@@_parse_bool
+ }
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\siunitx_cell_begin:, \siunitx_cell_end:}
+% \begin{macrocode}
+\cs_new_protected:Npn \siunitx_cell_begin:
+ {
+ \bool_if:NTF \l_@@_parse_bool
+ { \@@_collect_begin: }
+ { \@@_direct_begin: }
+ }
+\cs_new_protected:Npn \siunitx_cell_end:
+ {
+ \@@_collect_end:
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{Collecting tokens}
+%
+% \begin{variable}{\l_@@_collect_tl}
+% Space for tokens.
+% \begin{macrocode}
+\tl_new:N \l_@@_collect_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_collect_begin:}
+% Collecting a tabular cell means doing a token-by-token collection.
+% In previous versions of \pkg{siunitx} that was done along with picking
+% out the numerical part, but the code flow ends up very tricky. Here,
+% therefore, we just collect up the unchanged tokens first. Some issues
+% with the way that the \texttt{peek} functions are currently set up in
+% \pkg{expl3} mean that at present there is a need to do stuff by hand
+% here. The definition of \cs{cr} is used to allow collection of any tokens
+% inserted after the main content when dealing with the last cell of a row:
+% the \enquote{group} around it is needed to avoid issues with the underlying
+% |\halign|. (The approach is based on that in \pkg{collcell}.) Notice that
+% as each cell forms a group there is no need to reset the definition of
+% \cs{cr}.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_collect_begin:
+ {
+ \tl_clear:N \l_@@_collect_tl
+ \cs_set_protected:Npn \__peek_execute_branches:
+ {
+ \if_catcode:w \exp_not:N \l_peek_token \c_group_begin_token
+ \exp_after:wN \@@_collect_group:n
+ \else:
+ \exp_after:wN \@@_collect_token:N
+ \fi:
+ }
+ \if_false: { \fi:
+ \cs_set_protected:Npn \cr
+ {
+ \@@_collect_loop:
+ \tex_cr:D
+ }
+ \if_false: } \fi:
+ \@@_collect_loop:
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_collect_loop:}
+% \begin{macro}{\@@_collect_group:n}
+% \begin{macro}{\@@_collect_token:N}
+% \begin{macro}{\@@_collect_search:NnF}
+% \begin{macro}{\@@_collect_search_aux:NNn}
+% Collecting up the cell content needs a loop: this is done using
+% a |peek| approach as it's most natural. (A slower approach is possible
+% using something like the |\tl_lower_case:n| loop code.) The set of
+% possible tokens is somewhat limited compared to an arbitrary cell
+% (\foreign{cf.}~the approach in \pkg{collcell}): the special cases are
+% pulled out for manual handling. The flexible lookup approach is more-or-less
+% the same idea as in the kernel |case| functions. The |\relax| special case
+% covers the case where |\\| has been expanded in an empty cell.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_collect_loop:
+ { \peek_after:Nw \__peek_ignore_spaces_execute_branches: }
+\cs_new_protected:Npn \@@_collect_group:n #1
+ {
+ \tl_put_right:Nn \l_@@_collect_tl { {#1} }
+ \@@_collect_loop:
+ }
+\cs_new_protected:Npn \@@_collect_token:N #1
+ {
+ \@@_collect_search:NnF #1
+ {
+ \ignorespaces { \@@_collect_loop: }
+ \unskip { \@@_collect_loop: }
+ \end { \tabularnewline \end }
+ \relax { \relax }
+ \tabularnewline { \tabularnewline }
+ \siunitx_cell_end: { \siunitx_cell_end: }
+ }
+ {
+ \tl_put_right:Nn \l_@@_collect_tl {#1}
+ \@@_collect_loop:
+ }
+ }
+\AtBeginDocument
+ {
+ \@ifpackageloaded { mdwtab }
+ {
+ \cs_set_protected:Npn \@@_collect_token:N #1
+ {
+ \@@_collect_search:NnF #1
+ {
+ \@maybe at unskip { \@@_collect_loop: }
+ \ignorespaces { \@@_collect_loop: }
+ \tab at setcr { \@@_collect_loop: }
+ \unskip { \@@_collect_loop: }
+ \end { \tabularnewline \end }
+ \relax { \relax }
+ \tabularnewline { \tabularnewline }
+ \siunitx_cell_end: { \siunitx_cell_end: }
+ }
+ {
+ \tl_put_right:Nn \l_@@_collect_tl {#1}
+ \@@_collect_loop:
+ }
+ }
+ }
+ { }
+ }
+\cs_new_protected:Npn \@@_collect_search:NnF #1#2#3
+ {
+ \@@_collect_search_aux:NNn #1
+ #2
+ #1 {#3}
+ \q_stop
+ }
+\cs_new_protected:Npn \@@_collect_search_aux:NNn #1#2#3
+ {
+ \token_if_eq_meaning:NNTF #1 #2
+ { \use_i_delimit_by_q_stop:nw {#3} }
+ { \@@_collect_search_aux:NNn #1 }
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Separating collected material}
+%
+% The input needs to be divided into numerical tokens and those which appear
+% before and after them. This needs a second loop and validation.
+%
+% \begin{variable}{\l_@@_pre_tl, \l_@@_number_tl, \l_@@_post_tl}
+% Space for tokens.
+% \begin{macrocode}
+\tl_new:N \l_@@_pre_tl
+\tl_new:N \l_@@_number_tl
+\tl_new:N \l_@@_post_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_collect_end:}
+% At the end of the cell, expand all of the content as far as possible then
+% split it up into numerical and non-numerical parts.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_collect_end:
+ {
+ \protected at edef \l_@@_collect_tl
+ { \l_@@_collect_tl }
+ \tl_clear:N \l_@@_pre_tl
+ \tl_clear:N \l_@@_number_tl
+ \tl_clear:N \l_@@_post_tl
+ \cs_set_protected:Npn \__peek_execute_branches:
+ {
+ \if_catcode:w \exp_not:N \l_peek_token \c_group_begin_token
+ \exp_after:wN \@@_split_group:n
+ \else:
+ \exp_after:wN \@@_split_token:N
+ \fi:
+ }
+ \exp_after:wN \@@_split_loop: \l_@@_collect_tl
+ \q_recursion_tail \q_recursion_stop
+ \@@_split_tidy:N \l_@@_pre_tl
+ \@@_split_tidy:N \l_@@_post_tl
+ \tl_if_empty:NTF \l_@@_number_tl
+ { \@@_print_text:V \l_@@_pre_tl }
+ {
+ \@@_print:VVV
+ \l_@@_pre_tl
+ \l_@@_number_tl
+ \l_@@_post_tl
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_split_loop:}
+% \begin{macro}{\@@_split_group:n}
+% \begin{macro}{\@@_split_token:N}
+% Splitting into parts uses the fact that numbers cannot contain groups
+% and that we can track where we are up to based on the content of the
+% token lists.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_split_loop:
+ { \peek_after:Nw \__peek_ignore_spaces_execute_branches: }
+\cs_new_protected:Npn \@@_split_group:n #1
+ {
+ \tl_if_empty:NTF \l_@@_number_tl
+ { \tl_put_right:Nn \l_@@_pre_tl { {#1} } }
+ { \tl_put_right:Nn \l_@@_post_tl { {#1} } }
+ \@@_split_loop:
+ }
+\cs_new_protected:Npn \@@_split_token:N #1
+ {
+ \quark_if_recursion_tail_stop:N #1
+ \tl_if_empty:NTF \l_@@_post_tl
+ {
+ \siunitx_if_number_token:NTF #1
+ { \tl_put_right:Nn \l_@@_number_tl {#1} }
+ {
+ \tl_if_empty:NTF \l_@@_number_tl
+ { \tl_put_right \l_@@_pre_tl {#1} }
+ { \tl_put_right \l_@@_post_tl {#1} }
+ }
+ }
+ { \tl_put_right:Nn \l_@@_post_tl {#1} }
+ \@@_split_loop:
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_split_tidy:N}
+% \begin{macro}{\@@_split_tidy:Nn, \@@_split_tidy:NV}
+% A quick test for the entire content being surrounded by a set of braces:
+% rather than look explicitly, use the fact that a string comparison can
+% detect the same thing. The auxiliary is needed to avoid having to go
+% \foreign{via} a |:D| function (for the expansion behaviour).
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_split_tidy:N #1
+ {
+ \tl_if_empty:NF #1
+ { \@@_split_tidy:NV #1 #1 }
+ }
+\cs_new_protected:Npn \@@_split_tidy:Nn #1#2
+ {
+ \str_if_eq_x:nnT
+ { \exp_not:n {#2} }
+ { { \exp_not:o { \use:n #2 } } }
+ { \tl_set:No #1 { \use:n #2 } }
+ }
+\cs_generate_variant:Nn \@@_split_tidy:Nn { NV }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Printing numbers in cells: spacing}
+%
+% Getting the general alignment correct in tables is made more complex than one
+% would like by the \pkg{colortbl} package. In the original \LaTeXe{}
+% definition, cell material is centred by a construction of the (primitive)
+% form
+% \begin{verbatim}
+% \hfil
+% #
+% \hfil
+% \end{verbatim}
+% which only uses \texttt{fil} stretch. That is altered by \pkg{colortbl} to
+% broadly
+% \begin{verbatim}
+% \hskip 0pt plus 0.5fill
+% \kern 0pt
+% #
+% \hskip 0pt plus 0.5fill
+% \end{verbatim}
+% which means there is \texttt{fill} stretch to worry about and the kern as
+% well.
+%
+% \begin{macro}{\@@_skip:n}
+% To prevent combination of skips, a kern is inserted after each one.
+% This is best handled as a short auxiliary.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_skip:n #1
+ {
+ \skip_horizontal:n {#1}
+ \tex_kern:D \c_zero_skip
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{variable}{\l_@@_column_width_dim, \l_@@_fixed_width_bool}
+% Settings which apply to aligned columns in general.
+% \begin{macrocode}
+\keys_define:nn { siunitx / table }
+ {
+ column-width .dim_set:N =
+ \l_@@_column_width_dim ,
+ fixed-width .bool_set:N =
+ \l_@@_fixed_width_bool
+ }
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_align_center:n, \@@_align_left:n, \@@_align_right:n}
+% \begin{macro}{\@@_align_auxi:nn, \@@_align_auxii:nn}
+% The beginning and end of each table cell have to adjust the position of
+% the content using glue. When \pkg{colortbl} is loaded the glue is done in
+% two parts: one for our positioning and one to explicitly override that from
+% the package. Using a two-step auxiliary chain avoids needing to repeat any
+% code and the impact of the extra expansion should be trivial.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_align_center:n #1
+ { \@@_align_auxi:nn {#1} { 0pt plus 0.5fill } {#1} }
+\cs_new_protected:Npn \@@_align_left:n #1
+ { \@@_align_auxi:nn {#1} { 0pt } {#1} }
+\cs_new_protected:Npn \@@_align_right:n #1
+ { \@@_align_auxi:nn {#1} { 0pt plus 1fill } {#1} }
+\cs_new_protected:Npn \@@_align_auxi:nn #1#2
+ {
+ \bool_if:NTF \l_@@_fixed_width_bool
+ { \hbox_to_wd:nn \l_@@_column_width_dim }
+ { \use:n }
+ {
+ \@@_skip:n {#2}
+ #1
+ \@@_skip:n { 0pt plus 1fill - #2 }
+ }
+ }
+\AtBeginDocument
+ {
+ \@ifpackageloaded { colortbl }
+ {
+ \cs_new_eq:NN
+ \@@_align_auxii:nn
+ \@@_align_auxi:nn
+ \cs_set_protected:Npn \@@_align_auxi:nn #1#2
+ {
+ \@@_skip:n{ 0pt plus -0.5fill }
+ \@@_align_auxii:nn {#1} {#2}
+ \@@_skip:n { 0pt plus -0.5fill }
+ }
+ }
+ { }
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Printing just text}
+%
+% In cases where there is no numerical part, \pkg{siunitx} allows alignment
+% of the \enquote{escaped} text independent of the underlying column type.
+%
+% \begin{variable}{\l_@@_align_text_tl}
+% Alignment is handled using a |tl| as this allows a fast lookup at the
+% point of use.
+% \begin{macrocode}
+\keys_define:nn { siunitx / table }
+ {
+ text-alignment .choices:nn =
+ { center , left , right }
+ { \tl_set:Nn \l_@@_align_text_tl {#1} } ,
+ }
+\tl_new:N \l_@@_align_text_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_print_text:n, \@@_print_text:V}
+% Printing escaped text is easy: just place it in correctly in the
+% column.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_print_text:n #1
+ {
+ \use:c { @@_align_ \l_@@_align_text_tl :n } {#1}
+ }
+\cs_generate_variant:Nn \@@_print_text:n { V }
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{Reserving space: the table format}
+%
+% \begin{variable}{\l_@@_format_tl}
+% \begin{macrocode}
+\keys_define:nn { siunitx / table }
+ {
+ format .tl_set:N = \l_@@_format_tl
+ }
+% \end{macrocode}
+% \end{variable}
+%
+% \subsection{Directly printing without collection}
+%
+% Collecting the number allows for various effects but is not as fast
+% as simply aligning on the first token that is a decimal marker. The
+% strategy here is that used by \pkg{dcolumn}.
+%
+% \begin{macro}{\@@_direct_begin:}
+% \begin{macro}{\@@_direct_begin:w}
+% \begin{macro}{\@@_direct_begin_aux:}
+% After removing the |\ignorespaces| at the start of the cell, check to see
+% if there is a |{| and branch as appropriate.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_direct_begin:
+ { \@@_direct_begin:w }
+\cs_new_protected:Npn \@@_direct_begin:w \ignorespaces
+ {
+ \cs_set:Npn \__peek_execute_branches:
+ {
+ \if_catcode:w \exp_not:N \l_peek_token \c_group_begin_token
+ \exp_after:wN \@@_print_text:n
+ \else:
+ \m at th
+ \exp_after:wN \@@_direct_begin_aux:
+ \fi:
+ }
+ \peek_after:Nw \__peek_ignore_spaces_execute_branches:
+ }
+\cs_new_protected:Npn \@@_direct_begin_aux:
+ {
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Printing numbers in cells: main functions}
+%
+% \begin{variable}
+% {
+% \l_@@_align_comparator_bool ,
+% \l_@@_align_exponent_bool ,
+% \l_@@_align_uncertainty_bool ,
+% \l_@@_alignment_tl ,
+% \l_@@_parse_only_bool
+% }
+% Alignment is handled using a |tl| as this allows a fast lookup at the
+% point of use.
+% \begin{macrocode}
+\keys_define:nn { siunitx / table }
+ {
+ align-comparator .bool_set:N =
+ \l_@@_align_comparator_bool ,
+ align-exponent .bool_set:N =
+ \l_@@_align_exponent_bool ,
+ align-uncertainty .bool_set:N =
+ \l_@@_align_uncertainty_bool ,
+ alignment .choices:nn =
+ { center , left , right }
+ { \tl_set:Nn \l_@@_alignment_tl {#1} } ,
+ parse-only .bool_set:N =
+ \l_@@_parse_only_bool
+ }
+\tl_new:N \l_@@_alignment_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_print:nnn, \@@_print:VVV}
+% \begin{macro}{\@@_print_non_aligned:nnn, \@@_print_aligned:nnn}
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_print:nnn #1#2#3
+ {
+ \use:c { @@_align_ \l_@@_alignment_tl :n }
+ {
+ \bool_if:NTF \l_@@_parse_only_bool
+ { \@@_print_non_aligned:nnn }
+ { \@@_print_aligned:nnn }
+ {#1} {#2} {#3}
+ }
+ }
+\cs_generate_variant:Nn \@@_print:nnn { VVV }
+\cs_new_protected:Npn \@@_print_non_aligned:nnn #1#2#3
+ {
+ #1
+ \siunitx_number_format:nN {#2} \l_@@_tmp_tl
+ \siunitx_print:nV { number } \l_@@_tmp_tl
+ #3
+ }
+\cs_new_protected:Npn \@@_print_aligned:nnn #1#2#3
+ {
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Standard settings for module options}
+%
+% Some of these follow naturally from the point of definition
+% (\foreign{e.g.}~boolean variables are always |false| to begin with),
+% but for clarity everything is set here.
+% \begin{macrocode}
+\keys_set:nn { siunitx / table }
+ {
+ align-comparator = false ,
+ align-exponent = false ,
+ align-uncertainty = false ,
+ alignment = center ,
+ column-width = 0pt ,
+ fixed-width = false ,
+ format = ,
+ parse = true ,
+ parse-only = false ,
+ text-alignment = center
+ }
+% \end{macrocode}
+%
+%
+% \begin{macrocode}
+%</package>
+% \end{macrocode}
+%
+% \end{implementation}
+%
+% \PrintIndex
\ No newline at end of file
Property changes on: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-table.dtx
___________________________________________________________________
Added: svn:eol-style
## -0,0 +1 ##
+native
\ No newline at end of property
Added: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-unit.dtx
===================================================================
--- trunk/Master/texmf-dist/source/latex/siunitx/siunitx-unit.dtx (rev 0)
+++ trunk/Master/texmf-dist/source/latex/siunitx/siunitx-unit.dtx 2018-04-29 22:36:26 UTC (rev 47489)
@@ -0,0 +1,2516 @@
+% \iffalse meta-comment
+%
+% File: siunitx-unit.dtx Copyright (C) 2014-2018 Joseph Wright
+%
+% It may be distributed and/or modified under the conditions of the
+% LaTeX Project Public License (LPPL), either version 1.3c of this
+% license or (at your option) any later version. The latest version
+% of this license is in the file
+%
+% https://www.latex-project.org/lppl.txt
+%
+% This file is part of the "siunitx bundle" (The Work in LPPL)
+% and all files in that bundle must be distributed together.
+%
+% The released version of this bundle is available from CTAN.
+%
+% -----------------------------------------------------------------------
+%
+% The development version of the bundle can be found at
+%
+% https://github.com/josephwright/siunitx
+%
+% for those people who are interested.
+%
+% -----------------------------------------------------------------------
+%
+%<*driver>
+\documentclass{l3doc}
+% The next line is needed so that \GetFileInfo will be able to pick up
+% version data
+\usepackage{siunitx}
+\begin{document}
+ \DocInput{\jobname.dtx}
+\end{document}
+%</driver>
+% \fi
+%
+% \GetFileInfo{siunitx.sty}
+%
+% \title{^^A
+% \pkg{siunitx-unit} -- Parsing and formatting units^^A
+% \thanks{This file describes \fileversion,
+% last revised \filedate.}^^A
+% }
+%
+% \author{^^A
+% Joseph Wright^^A
+% \thanks{^^A
+% E-mail:
+% \href{mailto:joseph.wright at morningstar2.co.uk}
+% {joseph.wright at morningstar2.co.uk}^^A
+% }^^A
+% }
+%
+% \date{Released \filedate}
+%
+% \maketitle
+%
+% \begin{documentation}
+%
+% This submodule is dedicated to formatting physical units. The main function,
+% \cs{siunitx_unit_format:nN}, takes user input specify physical units and
+% converts it into a formatted token list suitable for typesetting in math
+% mode. While the formatter will deal correctly with \enquote{literal} user
+% input, the key strength of the module is providing a method to describe
+% physical units in a \enquote{symbolic} manner. The output format of these
+% symbolic units can then be controlled by a number of key--value options
+% made available by the module.
+%
+% A small number of \LaTeXe{} math mode commands are assumed to be available
+% as part of the formatted output. The \cs{mathchoice} command
+% (normally the \TeX{} primitive) is needed when using
+% |per-mode = symbol-or-fraction|. The command \cs{mathrm} is used for
+% wrapping the text (letter) part of units. The commands \cs{frac},
+% \cs{mbox}, \verb*|\ | and \cs{,} are used by the standard module settings,
+% and \cs{ensuremath}, \cs{hbar}, \cs{mathit} and \cs{mathrm} in some standard
+% unit definitions (for atomic and natural units). For the display of colored
+% (highlighted) and cancelled units, the commands \cs{textcolor} and
+% \cs{cancel} are assumed to be available.
+%
+% \section{Formatting units}
+%
+% \begin{function}{\siunitx_unit_format:nN}
+% \begin{syntax}
+% \cs{siunitx_unit_format:nN} \Arg{units} \meta{tl~var}
+% \end{syntax}
+% This function converts the input \meta{units} into a processed
+% \meta{tl~var} which can then be inserted in math mode to typeset the
+% material. Where the \meta{units} are given in symbolic form, described
+% elsewhere, this formatting process takes place in two stages: the
+% \meta{units} are parsed into a structured form before the generation
+% of the appropriate output form based on the active settings. When the
+% \meta{units} are given as literals, processing is minimal: the
+% characters |.| and |~| are converted to unit products (boundaries).
+% In both cases, the result is a series of tokens intended to be typeset
+% in math mode with appropriate choice of font for typesetting of the
+% textual parts.
+%
+% For example,
+% \begin{verbatim}
+% \siunitx_unit_format:nN { \kilo \metre \per \second } \l_tmpa_tl
+% \end{verbatim}
+% will, with standard settings, result in \cs{l_tmpa_tl} being set to
+% \begin{verbatim}
+% \mathrm{km}\,\mathrm{s}^{-1}
+% \end{verbatim}
+% \end{function}
+%
+% \begin{function}{\siunitx_unit_format:nNN}
+% \begin{syntax}
+% \cs{siunitx_unit_format:nNN} \Arg{units} \meta{tl~var} \meta{fp~var}
+% \end{syntax}
+% This function formats the \meta{units} in the same way as described for
+% \cs{siunitx_unit_format:nN}. When the input is given in symbolic form,
+% any decimal unit prefixes will be extracted and the overall power of
+% ten that these represent will be stored in the \meta{fp~var}.
+%
+% For example,
+% \begin{verbatim}
+% \siunitx_unit_format:nNN { \kilo \metre \per \second }
+% \l_tmpa_tl \l_tmpa_fp
+% \end{verbatim}
+% will, with standard settings, result in \cs{l_tmpa_tl} being set to
+% \begin{verbatim}
+% \mathrm{m}\,\mathrm{s}^{-1}
+% \end{verbatim}
+% with \cs{l_tmpa_fp} taking value~$3$. Note that the latter is a floating
+% point variable: it is possible for non-integer values to be obtained here.
+% \end{function}
+%
+% \section{Defining symbolic units}
+%
+% \begin{function}{\siunitx_declare_prefix:Nnn}
+% \begin{syntax}
+% \cs{siunitx_declare_prefix:Nnn} \meta{prefix} \Arg{symbol} \Arg{power}
+% \end{syntax}
+% Defines a symbolic \meta{prefix} (which should be a control sequence
+% such as |\kilo|) to be converted by the parser to the \meta{symbol}.
+% The latter should consist of literal content (\emph{e.g.}~|k|).
+% In literal mode the \meta{symbol} will be typeset directly. The prefix
+% should represent an integer \meta{power} of $10$, and this information
+% may be used to convert from one or more \meta{prefix} symbols to an
+% overall power applying to a unit. See also
+% \cs{siunitx_declare_prefix:Nn}.
+% \end{function}
+%
+% \begin{function}{\siunitx_declare_prefix:Nn}
+% \begin{syntax}
+% \cs{siunitx_declare_prefix:Nn} \meta{prefix} \Arg{symbol}
+% \end{syntax}
+% Defines a symbolic \meta{prefix} (which should be a control sequence
+% such as |\kilo|) to be converted by the parser to the \meta{symbol}.
+% The latter should consist of literal content (\emph{e.g.}~|k|).
+% In literal mode the \meta{symbol} will be typeset directly. In contrast
+% to \cs{siunitx_declare_prefix:Nnn}, there is no assumption about the
+% mathematical nature of the \meta{prefix}, \emph{i.e.}~the prefix may
+% represent a power of any base. As a result, no conversion of the
+% \meta{prefix} to a numerical power will be possible.
+% \end{function}
+%
+% \begin{function}{\siunitx_declare_power:NNn}
+% \begin{syntax}
+% \cs{siunitx_declare_power:NnN} \meta{pre-power} \meta{post-power} \Arg{value}
+% \end{syntax}
+% Defines \emph{two} symbolic \meta{powers} (which should be control
+% sequences such as |\squared|) to be converted by the parser to the
+% \meta{value}. The latter should be an integer or floating point number in
+% the format defined for \pkg{l3fp}. Powers may precede a unit or be give
+% after it: both forms are declared at once, as indicated by the argument
+% naming. In literal mode, the \meta{value} will be applied as
+% a superscript to either the next token in the input (for the
+% \meta{pre-power}) or appended to the previously-typeset material
+% (for the \meta{post-power}).
+% \end{function}
+%
+% \begin{function}{\siunitx_declare_qualifier:Nn}
+% \begin{syntax}
+% \cs{siunitx_declare_qualifier:Nn} \meta{qualifier} \Arg{meaning}
+% \end{syntax}
+% Defines a symbolic \meta{qualifier} (which should be a control sequence
+% such as |\catalyst|) to be converted by the parser to the \meta{meaning}.
+% The latter should consist of literal content (\emph{e.g.}~|cat|). In
+% literal mode the \meta{meaning} will be typeset following a space after
+% the unit to which it applies.
+% \end{function}
+%
+% \begin{function}{\siunitx_declare_unit:Nn, \siunitx_declare_unit:Nx}
+% \begin{syntax}
+% \cs{siunitx_declare_unit:Nn} \meta{unit} \Arg{meaning}
+% \end{syntax}
+% Defines a symbolic \meta{unit} (which should be a control sequence
+% such as |\kilogram|) to be converted by the parser to the \meta{meaning}.
+% The latter may consist of literal content (\emph{e.g.}~|kg|), other
+% symbolic unit commands (\emph{e.g.}~|\kilo\gram|) or a mixture of the two.
+% In literal mode the \meta{meaning} will be typeset directly.
+% \end{function}
+%
+% \begin{variable}{\l_siunitx_unit_symbolic_seq}
+% This sequence contains all of the symbolic \meta{unit} names defined :
+% these will be in the form of control sequences such as |\kilogram|.
+% The order of the sequence is unimportant.
+% \end{variable}
+%
+% \section{Pre-defined symbolic unit components}
+%
+% The unit parser is defined to recognise a number of pre-defined units,
+% prefixes and powers, and also interpret a small selection of
+% \enquote{generic} symbolic parts.
+%
+% Broadly, the pre-defined units are those defined by the \textsc{bipm} in the
+% documentation for the \emph{International System of Units} (SI)~\cite{BIPM}.
+% As far as possible, the names given to the command names for units are those
+% used by the \textsc{bipm}, omitting spaces and using only \textsc{ascii}
+% characters. The standard symbols are also taken from the same documentation.
+% In the following documentation, the order of the description of units
+% broadly follows the SI~Brochure.
+%
+% \begin{function}
+% {
+% \kilogram ,
+% \metre ,
+% \meter ,
+% \mole ,
+% \kelvin ,
+% \candela ,
+% \second ,
+% \ampere
+% }
+% The base units as defined in Section~2.1 of the SI Brochure~\cite{SI:2.1}.
+% Notice that \cs{meter} is defined as an alias for \cs{metre} as the former
+% spelling is common in the US (although the latter is the official spelling).
+% \end{function}
+%
+% \begin{function}{\gram}
+% The base unit \cs{kilogram} is defined using an SI prefix: as such the
+% (derived) unit \cs{gram} is required by the module to correctly produce
+% output for the \cs{kilogram}.
+% \end{function}
+%
+% \begin{function}
+% {
+% \yocto ,
+% \zepto ,
+% \atto ,
+% \femto ,
+% \pico ,
+% \nano ,
+% \micro ,
+% \milli ,
+% \centi ,
+% \deci ,
+% \deca ,
+% \deka ,
+% \hecto ,
+% \kilo ,
+% \mega ,
+% \giga ,
+% \tera ,
+% \peta ,
+% \exa ,
+% \zetta ,
+% \yotta
+% }
+% Prefixes, all of which are integer powers of $10$: the powers are stored
+% internally by the module and can be used for conversion from prefixes to
+% their numerical equivalent. These prefixes are documented in Section~3.1
+% of the SI~Brochure~\cite{SI:3.1}.
+%
+% Note that the \cs{kilo} prefix is required to
+% define the base \cs{kilogram} unit. Also note the two spellings available
+% for \cs{deca}/\cs{deka}.
+% \end{function}
+%
+% \begin{function}
+% {
+% \becquerel ,
+% \degreeCelsius ,
+% \coulomb ,
+% \farad ,
+% \gray ,
+% \hertz ,
+% \henry ,
+% \joule ,
+% \katal ,
+% \lumen ,
+% \lux ,
+% \newton ,
+% \ohm ,
+% \pascal ,
+% \radian ,
+% \siemens ,
+% \sievert ,
+% \steradian ,
+% \tesla ,
+% \volt ,
+% \watt ,
+% \weber
+% }
+% The defined SI~units with defined names and symbols, as given in
+% Section~2.2.2 of the SI~Brochure~\cite{SI:2.2.2}. Notice that the names
+% of the units are lower case with the exception of \cs{degreeCelsius}, and
+% that this unit name includes \enquote{degree}.
+% \end{function}
+%
+% \begin{function}
+% {
+% \day ,
+% \hectare ,
+% \hour ,
+% \litre ,
+% \liter ,
+% \minute ,
+% \tonne
+% }
+% Units accepted for use with the SI: here \cs{minute} is a unit of time
+% not of plane angle. These units are taken from Table~4.1 of the
+% SI~Brochure~\cite{SI:T6}.
+%
+% For the unit \cs{litre}, both |l| and |L| are listed
+% as acceptable symbols: the latter is the standard setting of the module.
+% The alternative spelling \cs{liter} is also given for this unit for US
+% users (as with \cs{metre}, the official spelling is \enquote{re}).
+% \end{function}
+%
+% \begin{function}
+% {
+% \arcminute ,
+% \arcsecond ,
+% \degree
+% }
+% Units for plane angles accepted for use with the SI: to avoid a clash
+% with units for time, here \cs{arcminute} and \cs{arcsecond} are used in
+% place of \cs{minute} and \cs{second}. These units are taken from Table~4.1
+% of the SI~Brochure~\cite{SI:T6}.
+% \end{function}
+%
+% \begin{function}
+% {
+% \astronomicalunit ,
+% \atomicmassunit ,
+% \auaction ,
+% \aucharge ,
+% \auenergy ,
+% \aulength ,
+% \aumass ,
+% \autime ,
+% \bohr ,
+% \dalton ,
+% \electronvolt ,
+% \hartree ,
+% \nuaction ,
+% \numass ,
+% \nuspeed ,
+% \nutime
+% }
+% Non-SI where values must be determined experimentally. These units are taken from
+% Table~7 of the SI~Brochure~\cite{SI:T7}. Where no better name is given for
+% the unit in the SI~Brochure, the prefixes |nu| (natural unit) and |au| (atomic
+% unit) are used.
+%
+% Note that the value of the natural unit of speed (the speed of light) is used
+% to define the second and is thus not determined by experiment: it is however
+% included in this set of units.
+% \end{function}
+%
+% \begin{function}
+% {
+% \angstrom ,
+% \bar ,
+% \barn ,
+% \bel ,
+% \decibel ,
+% \knot ,
+% \millimetremercury ,
+% \nauticalmile ,
+% \neper
+% }
+% Non-SI units accepted for use with the SI. These units are taken from
+% Table~8 of the SI~Brochure~\cite{SI:T8}.
+% \end{function}
+%
+% \begin{function}
+% {
+% \dyne ,
+% \erg ,
+% \gal ,
+% \gauss ,
+% \maxwell ,
+% \oersted ,
+% \phot ,
+% \poise ,
+% \stilb ,
+% \stokes
+% }
+% Non-SI units associated with the CGS and the CGS-Gaussian system of units.
+% These units are taken from Table~9 of the SI~Brochure~\cite{SI:T9}.
+% \end{function}
+%
+% \begin{function}{\percent}
+% The mathematical concept of percent, usable with the SI as detailed in
+% Section~5.3.7 of the SI~Brochure~\cite{SI:5.3.7}.
+% \end{function}
+%
+% \begin{function}{\square, \cubic}
+% \begin{syntax}
+% \cs{square} \meta{prefix} \meta{unit}
+% \cs{cubic} \meta{prefix} \meta{unit}
+% \end{syntax}
+% Pre-defined unit powers which apply to the next \meta{prefix}/\meta{unit}
+% combination.
+% \end{function}
+%
+% \begin{function}{\squared, \cubed}
+% \begin{syntax}
+% \meta{prefix} \meta{unit} \cs{squared}
+% \meta{prefix} \meta{unit} \cs{cubed}
+% \end{syntax}
+% Pre-defined unit powers which apply to the preceding
+% \meta{prefix}/\meta{unit} combination.
+% \end{function}
+%
+% \begin{function}{\per}
+% \begin{syntax}
+% \cs{per} \meta{prefix} \meta{unit} \meta{power}
+% \end{syntax}
+% Indicates that the next \meta{prefix}/\meta{unit}/\meta{power} combination
+% is reciprocal, \emph{i.e.}~raises it to the power $-1$. This symbolic
+% representation may be applied in addition to a \cs{power}, and will work
+% correctly if the \cs{power} itself is negative. In literal mode \cs{per}
+% will print a slash (\enquote{$/$}).
+% \end{function}
+%
+% \begin{function}{\cancel}
+% \begin{syntax}
+% \cs{cancel} \meta{prefix} \meta{unit} \meta{power}
+% \end{syntax}
+% Indicates that the next \meta{prefix}/\meta{unit}/\meta{power} combination
+% should be \enquote{cancelled out}. In the parsed output, the entire unit
+% combination will be given as the argument to a function \cs{cancel}, which
+% is assumed to be available at a higher level. In literal mode, the same
+% higher-level \cs{cancel} will be applied to the next token. It is the
+% responsibility of the calling code to provide an appropriate definition
+% for \cs{cancel} outside of the scope of the unit parser.
+% \end{function}
+%
+% \begin{function}{\highlight}
+% \begin{syntax}
+% \cs{highlight} \Arg{color} \meta{prefix} \meta{unit} \meta{power}
+% \end{syntax}
+% Indicates that the next \meta{prefix}/\meta{unit}/\meta{power} combination
+% should be highlighted in the specified \meta{color}. In the parsed output,
+% the entire unit combination will be given as the argument to a function
+% \cs{textcolor}, which is assumed to be available at a higher level. In
+% literal mode, the same higher-level \cs{textcolor} will be applied to the
+% next token. It is the responsibility of the calling code to provide an
+% appropriate definition for \cs{textcolor} outside of the scope of the unit
+% parser.
+% \end{function}
+%
+% \begin{function}{\of}
+% \begin{syntax}
+% \meta{prefix} \meta{unit} \meta{power} \cs{of} \Arg{qualifier}
+% \end{syntax}
+% Indicates that the \meta{qualifier} applies to the current
+% \meta{prefix}/\meta{unit}/\meta{power} combination. In parsed mode, the
+% display of the result will depend upon module options. In literal mode,
+% the \meta{qualifier} will be printed in parentheses following the preceding
+% \meta{unit} and a full-width space.
+% \end{function}
+%
+% \begin{function}{\raiseto, \tothe}
+% \begin{syntax}
+% \cs{raiseto} \Arg{power} \meta{prefix} \meta{unit}
+% \meta{prefix} \meta{unit} \cs{tothe} \Arg{power}
+% \end{syntax}
+% Indicates that the \meta{power} applies to the current
+% \meta{prefix}/\meta{unit} combination. As shown, \cs{raiseto} applies to
+% the next \meta{unit} whereas \cs{tothe} applies to the preceding unit. In
+% literal mode the \cs{power} will be printed as a superscript attached to
+% the next token (\cs{raiseto}) or preceding token (\cs{tothe}) as
+% appropriate.
+% \end{function}
+%
+% \subsection{Key--value options}
+%
+% The options defined by this submodule are available within the \pkg{l3keys}
+% |siunitx| tree.
+%
+% \begin{function}{bracket-denominator}
+% \begin{syntax}
+% |bracket-denominator| = |true|\verb"|"|false|
+% \end{syntax}
+% Switch to determine whether brackets are added to the denominator part of
+% a unit when printed using inline fractional form (with |per-mode| as
+% |repeated-symbol|, |symbol| or |symbol-or-fraction|). The standard setting
+% is |true|.
+% \end{function}
+%
+% \begin{function}{fraction-command}
+% \begin{syntax}
+% |fraction-command| = \meta{command}
+% \end{syntax}
+% Command used to create fractional output when |per-mode| is set to
+% |fraction|. The standard setting is |\frac|.
+% \end{function}
+%
+% \begin{function}{parse-units}
+% \begin{syntax}
+% |parse-units| = |true|\verb"|"|false|
+% \end{syntax}
+% Determines whether parsing of unit symbols is attempted or literal
+% mode is used directly. The standard setting is |true|.
+% \end{function}
+%
+% \begin{function}{per-mode}
+% \begin{syntax}
+% |per-mode| = \meta{choice}
+% \end{syntax}
+% Selects how the negative powers (\cs{per}) are formatted: a choice from
+% the options |fraction|, |power|, |power-positive-first|, |repeated-symbol|,
+% |symbol| and |symbol-or-fraction|. The option |fraction| generates
+% fractional output when appropriate using the command specified by
+% the |fraction-command| option. The setting |power| uses reciprocal powers
+% leaving the units in the order of input, while |power-positive-first| uses
+% the same display format but sorts units such that the positive powers
+% come before negative ones. The |symbol| setting uses a symbol (specified
+% by |per-symbol|) between positive and negative powers, while
+% |repeated-symbol| uses the same symbol but places it before \emph{every}
+% unit with a negative power (this is mathematically \enquote{wrong} but
+% often seen in real work). Finally, |symbol-or-fraction| acts like
+% |symbol| for inline output and like |fraction| when the output is used
+% in a display math environment. The standard setting is |power|.
+% \end{function}
+%
+% \begin{function}{per-symbol}
+% \begin{syntax}
+% |per-symbol| = \meta{symbol}
+% \end{syntax}
+% Specifies the symbol to be used to denote negative powers when the option
+% |per-mode| is set to |repeated-symbol|, |symbol| or |symbol-or-fraction|.
+% The standard setting is |/|.
+% \end{function}
+%
+% \begin{function}{qualifier-mode}
+% \begin{syntax}
+% |qualifier-mode| = \meta{choice}
+% \end{syntax}
+% Selects how qualifiers are formatted: a choice from the options |brackets|,
+% |combine|, |phrase| and |subscript|. The option |bracket| wraps
+% the qualifier in parenthesis, |combine| joins the qualifier with the unit
+% directly, |phrase| inserts the content stored by the option
+% |qualifier-phrase| between the unit and qualifier, and |subscript| formats
+% the qualifier as a subscript. The standard setting is |subscript|.
+% \end{function}
+%
+% \begin{function}{qualifier-phrase}
+% \begin{syntax}
+% |qualifier-phrase| = \meta{choice}
+% \end{syntax}
+% Defines the text to be inserted between a unit and qualifier when
+% |qualifier-mode| is set to |phrase|. This material is inserted without
+% any font control and so if text mode is required it should be
+% included in the setting, for example \verb*|\ \mbox{of}\ |. The
+% standard setting is a full width space (\verb*|\ |).
+% \end{function}
+%
+% \begin{function}{sticky-per}
+% \begin{syntax}
+% |sticky-per| = |true|\verb"|"|false|
+% \end{syntax}
+% Used to determine whether \cs{per} should be applied one a unit-by-unit
+% basis (when |false|) or should apply to all following units
+% (when |true|). The latter mode is somewhat akin conceptually to the
+% \TeX{} \cs{over} primitive. The standard setting is |false|.
+% \end{function}
+%
+% \begin{function}{unit-close-bracket}
+% \begin{syntax}
+% |unit-close-bracket| = \meta{symbol}
+% \end{syntax}
+% Bracket symbol used to close a matched pair around units when once is
+% required to maintain mathematical logic. The standard setting is ^^A (
+% |)|.
+% \end{function}
+%
+% \begin{function}{unit-open-bracket}
+% \begin{syntax}
+% |unit-open-bracket| = \meta{symbol}
+% \end{syntax}
+% Bracket symbol used to open a matched pair around units when once is
+% required to maintain mathematical logic. The standard setting is |(|. ^^A )
+% \end{function}
+%
+% \begin{function}{unit-product}
+% \begin{syntax}
+% |unit-product| = \meta{separator}
+% \end{syntax}
+% Inserted between unit combinations in parsed mode, and used to replace
+% |.| and |~| in literal mode. The standard setting is |\,|.
+% \end{function}
+%
+% \end{documentation}
+%
+% \begin{implementation}
+%
+% \section{\pkg{siunitx-unit} implementation}
+%
+% Start the \pkg{DocStrip} guards.
+% \begin{macrocode}
+%<*package>
+% \end{macrocode}
+%
+% Identify the internal prefix (\LaTeX3 \pkg{DocStrip} convention): only
+% internal material in this \emph{submodule} should be used directly.
+% \begin{macrocode}
+%<@@=siunitx_unit>
+% \end{macrocode}
+%
+% \subsection{Initial set up}
+%
+% The mechanisms defined here need a few variables to exist and to be
+% correctly set: these don't belong to one subsection and so are created
+% in a small general block.
+%
+% Variants not provided by \pkg{expl3}.
+% \begin{macrocode}
+\cs_generate_variant:Nn \tl_replace_all:Nnn { NnV }
+% \end{macrocode}
+%
+% \begin{variable}{\l_@@_tmp_fp}
+% \begin{variable}{\l_@@_tmp_int}
+% \begin{variable}{\l_@@_tmp_tl}
+% Scratch space.
+% \begin{macrocode}
+\fp_new:N \l_@@_tmp_fp
+\int_new:N \l_@@_tmp_int
+\tl_new:N \l_@@_tmp_tl
+% \end{macrocode}
+% \end{variable}
+% \end{variable}
+% \end{variable}
+%
+% \begin{variable}{\c_@@_math_subscript_tl}
+% Useful tokens with awkward category codes.
+% \begin{macrocode}
+\tl_const:Nx \c_@@_math_subscript_tl
+ { \char_generate:nn { `\_ } { 8 } }
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_parsing_bool}
+% A boolean is used to indicate when the symbolic unit functions should
+% produce symbolic or literal output. This is used when the symbolic names
+% are used along with literal input, and ensures that there is a sensible
+% fall-back for these cases.
+% \begin{macrocode}
+\bool_new:N \l_@@_parsing_bool
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_test_bool}
+% A switch used to indicate that the code is testing the input to find
+% if there is any typeset output from individual unit macros. This is needed
+% to allow the \enquote{base} macros to be found, and also to pick up the
+% difference between symbolic and literal unit input.
+% \begin{macrocode}
+\bool_new:N \l_@@_test_bool
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_if_symbolic:nTF}
+% The test for symbolic units is needed in two places. First, there is the
+% case of \enquote{pre-parsing} input to check if it can be parsed. Second,
+% when parsing there is a need to check if the current unit is built up
+% from others (symbolic) or is defined in terms of some literals. To do this,
+% the approach used is to set all of the symbolic unit commands expandable
+% and to do nothing, with the few special cases handled manually. If
+% an \texttt{f}-type definition then yields nothing at all then the
+% assumption is that the input is symbolic. (We use \texttt{f}-type
+% expansion since it will turn the symbolic unit names into nothing at
+% all but doesn't require them to be expandable.)
+% \begin{macrocode}
+\prg_new_protected_conditional:Npnn \@@_if_symbolic:n #1 { TF }
+ {
+ \group_begin:
+ \bool_set_true:N \l_@@_test_bool
+ \tl_set:Nf \l_@@_tmp_tl {#1}
+ \exp_args:NNV \group_end:
+ \tl_if_blank:nTF \l_@@_tmp_tl
+ { \prg_return_true: }
+ { \prg_return_false: }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{Defining symbolic unit}
+%
+% Unit macros and related support are created here. These exist only within
+% the scope of the unit processor code, thus not polluting document-level
+% namespace and allowing overlap with other areas in the case of useful short
+% names (for example \cs{pm}). Setting up the mechanisms to allow this requires
+% a few additional steps on top of simply saving the data given by the user
+% in creating the unit.
+%
+% \begin{variable}{\l_siunitx_unit_symbolic_seq}
+% A list of all of the symbolic units, \emph{etc.}, set up. This is needed
+% to allow the symbolic names to be defined within the scope of the unit
+% parser but not elsewhere using simple mappings.
+% \begin{macrocode}
+\seq_new:N \l_siunitx_unit_symbolic_seq
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_set_symbolic:Nnn}
+% \begin{macro}{\@@_set_symbolic:Npnn}
+% \begin{macro}{\@@_set_symbolic:NNpnn}
+% The majority of the work for saving each symbolic definition is the same
+% irrespective of the item being defined (unit, prefix, power, qualifier).
+% This is therefore all carried out in a single internal function which
+% does the common tasks. The three arguments here are the symbolic macro
+% name, the literal output and the code to insert when doing full unit
+% parsing. To allow for the \enquote{special cases} (where arguments are
+% required) the entire mechanism is set up in a two-part fashion allowing
+% for flexibility at the slight cost of additional functions.
+%
+% Importantly, notice that the unit macros are declared as expandable. This
+% is required so that literals can be correctly converted into a token list
+% of material which does not depend on local redefinitions for the unit
+% macros. That is required so that the unit formatting system can be grouped.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_set_symbolic:Nnn #1
+ { \@@_set_symbolic:NNnnn \cs_set:cpn #1 { } }
+\cs_new_protected:Npn \@@_set_symbolic:Npnn #1#2#
+ { \@@_set_symbolic:NNnnn \cs_set:cpn #1 {#2} }
+\cs_new_protected:Npn \@@_set_symbolic:NNnnn #1#2#3#4#5
+ {
+ \seq_put_right:Nn \l_siunitx_unit_symbolic_seq {#2}
+ #1 { units ~ > ~ \token_to_str:N #2 } #3
+ {
+ \bool_if:NF \l_@@_test_bool
+ {
+ \bool_if:NTF \l_@@_parsing_bool
+ {#5}
+ {#4}
+ }
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\siunitx_declare_power:NNn}
+% Powers can come either before or after the unit. As they always come
+% (logically) in matching, we handle this by declaring two commands,
+% and setting each up separately.
+% \begin{macrocode}
+\cs_new_protected:Npn \siunitx_declare_power:NNn #1#2#3
+ {
+ \@@_set_symbolic:Nnn #1
+ { \@@_literal_power:nN {#1} }
+ { \@@_parse_power:nnN {#1} {#3} \c_true_bool }
+ \@@_set_symbolic:Nnn #2
+ { ^ {#3} }
+ { \@@_parse_power:nnN {#2} {#3} \c_false_bool }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\siunitx_declare_prefix:Nn}
+% \begin{macro}{\siunitx_declare_prefix:Nnn}
+% \begin{variable}
+% {\l_@@_prefixes_forward_prop, \l_@@_prefixes_reverse_prop}
+% For prefixes there are a couple of options. In all cases, the basic
+% requirement is to set up to parse the prefix using the appropriate
+% internal function. For prefixes which are powers of $10$, there is also
+% the need to be able to do conversion to/from the numerical equivalent.
+% That is handled using two properly lists which can be used to supply
+% the conversion data later.
+% \begin{macrocode}
+\cs_new_protected:Npn \siunitx_declare_prefix:Nn #1#2
+ {
+ \@@_set_symbolic:Nnn #1
+ {#2}
+ { \@@_parse_prefix:Nn #1 {#2} }
+ }
+\cs_new_protected:Npn \siunitx_declare_prefix:Nnn #1#2#3
+ {
+ \siunitx_declare_prefix:Nn #1 {#2}
+ \prop_put:Nnn \l_@@_prefixes_forward_prop {#2} {#3}
+ \prop_put:Nnn \l_@@_prefixes_reverse_prop {#3} {#2}
+ }
+\prop_new:N \l_@@_prefixes_forward_prop
+\prop_new:N \l_@@_prefixes_reverse_prop
+% \end{macrocode}
+% \end{variable}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\siunitx_declare_qualifier:Nn}
+% Qualifiers are relatively easy to handle: nothing to do other than save
+% the input appropriately.
+% \begin{macrocode}
+\cs_new_protected:Npn \siunitx_declare_qualifier:Nn #1#2
+ {
+ \@@_set_symbolic:Nnn #1
+ { \ ( #2 ) }
+ { \@@_parse_qualifier:nn {#1} {#2} }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\siunitx_declare_unit:Nn, \siunitx_declare_unit:Nx}
+% For the unit parsing, allowing for variations in definition order requires
+% that a test is made for the output of each unit at point of use.
+% \begin{macrocode}
+\cs_new_protected:Npn \siunitx_declare_unit:Nn #1#2
+ {
+ \@@_set_symbolic:Nnn #1
+ {#2}
+ {
+ \@@_if_symbolic:nTF {#2}
+ {#2}
+ { \@@_parse_unit:Nn #1 {#2} }
+ }
+ }
+\cs_generate_variant:Nn \siunitx_declare_unit:Nn { Nx }
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{Non-standard symbolic units}
+%
+% A few of the symbolic units require non-standard definitions: these are
+% created here. They all use parts of the more general code but have particular
+% requirements which can only be addressed by hand. Some of these could in
+% principle be used in place of the dedicated definitions above, but at point
+% of use that would then require additional expansions for each unit parsed:
+% as the macro names would still be needed, this does not offer any real
+% benefits.
+%
+% \begin{macro}{\per}
+% The \cs{per} symbolic unit is a bit special: it has a mechanism entirely
+% different from everything else, so has to be set up by hand. In literal
+% mode it is represented by a very simple symbol!
+% \begin{macrocode}
+\@@_set_symbolic:Nnn \per
+ { / }
+ { \@@_parse_per: }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\cancel}
+% \begin{macro}{\highlight}
+% The two special cases, \cs{cancel} and \cs{highlight}, are easy to deal
+% with when parsing. When not parsing, a precaution is taken to ensure that
+% the user level equivalents always get a braced argument.
+% \begin{macrocode}
+\@@_set_symbolic:Npnn \cancel
+ { \@@_literal_special:nN { \cancel } }
+ { \@@_parse_special:n { \cancel } }
+\@@_set_symbolic:Npnn \highlight #1
+ { \@@_literal_special:nN { \textcolor {#1} } }
+ { \@@_parse_special:n { \textcolor {#1} } }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\of}
+% The generic qualifier is simply the same as the dedicated ones except for
+% needing to grab an argument.
+% \begin{macrocode}
+\@@_set_symbolic:Npnn \of #1
+ { \ ( #1 ) }
+ { \@@_parse_qualifier:nn { \of {#1} } {#1} }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\raiseto, \tothe}
+% Generic versions of the pre-defined power macros. These require an
+% argument and so cannot be handled using the general approach. Other than
+% that, the code here is very similar to that in
+% \cs{siunitx_unit_power_set:NnN}.
+% \begin{macrocode}
+\@@_set_symbolic:Npnn \raiseto #1
+ { \@@_literal_power:nN {#1} }
+ { \@@_parse_power:nnN { \raiseto {#1} } {#1} \c_true_bool }
+\@@_set_symbolic:Npnn \tothe #1
+ { ^ {#1} }
+ { \@@_parse_power:nnN { \tothe {#1} } {#1} \c_false_bool }
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{Main formatting routine}
+%
+% Unit input can take two forms, \enquote{literal} units (material to be
+% typeset directly) or \enquote{symbolic} units (macro-based). Before any
+% parsing or typesetting is carried out, a small amount of pre-parsing has to
+% be carried out to decide which of these cases applies.
+%
+% \begin{variable}{\l_@@_product_tl}
+% Options which apply to the main formatting routine, and so are not tied
+% to either symbolic or literal input.
+% \begin{macrocode}
+\keys_define:nn { siunitx }
+ {
+ unit-product .tl_set:N = \l_@@_product_tl
+ }
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_formatted_tl}
+% A token list for the final formatted result: may or may not be generated
+% by the parser, depending on the nature of the input.
+% \begin{macrocode}
+\tl_new:N \l_@@_formatted_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\siunitx_unit_format:nN}
+% \begin{macro}{\siunitx_unit_format:nNN}
+% \begin{macro}{\@@_format:nNN}
+% \begin{macro}{\@@_format_aux:}
+% Formatting parsed units can take place either with the prefixes printed or
+% separated out into a power of ten. This variation is handled using two
+% separate functions: as this submodule does not really deal with numbers,
+% formatting the numeral part here would be tricky and it is better therefore
+% to have a mechanism to return a simple numerical power. At the same time,
+% most uses will no want this more complex return format and so a version of
+% the code which does not do this is also provided.
+%
+% The main unit formatting routine groups all of the parsing/formatting, so
+% that the only value altered will be the return token list. As definitions
+% for the various unit macros are not globally created, the first step is to
+% map over the list of names and active the unit definitions: these do
+% different things depending on the switches set. There is then a decision to
+% be made: is the unit input one that can be parsed (\enquote{symbolic}), or
+% is is one containing one or more literals. In the latter case, there is a
+% still the need to convert the input into an expanded token list as some
+% parts of the input could still be using unit macros.
+%
+% Notice that for \cs{siunitx_unit_format:nN} a second return value from the
+% auxiliary has to be allowed for, but is simply discarded.
+% \begin{macrocode}
+\cs_new_protected:Npn \siunitx_unit_format:nN #1#2
+ {
+ \bool_set_false:N \l_@@_prefix_power_bool
+ \@@_format:nNN {#1} #2 \l_@@_tmp_fp
+ }
+\cs_new_protected:Npn \siunitx_unit_format:nNN #1#2#3
+ {
+ \bool_set_true:N \l_@@_prefix_power_bool
+ \@@_format:nNN {#1} #2 #3
+ }
+\cs_new_protected:Npn \@@_format:nNN #1#2#3
+ {
+ \group_begin:
+ \seq_map_inline:Nn \l_siunitx_unit_symbolic_seq
+ { \cs_set_eq:Nc ##1 { units ~ > ~ \token_to_str:N ##1 } }
+ \tl_clear:N \l_@@_formatted_tl
+ \fp_zero:N \l_@@_prefix_fp
+ \bool_if:NTF \l_@@_parse_bool
+ {
+ \@@_if_symbolic:nTF {#1}
+ {
+ \@@_parse:n {#1}
+ \prop_if_empty:NF \l_@@_parsed_prop
+ { \@@_format_parsed: }
+ }
+ { \@@_format_literal:n {#1} }
+ }
+ { \@@_format_literal:n {#1} }
+ \cs_set_protected:Npx \@@_format_aux:
+ {
+ \tl_set:Nn \exp_not:N #2
+ { \exp_not:V \l_@@_formatted_tl }
+ \fp_set:Nn \exp_not:N #3
+ { \fp_use:N \l_@@_prefix_fp }
+ }
+ \exp_after:wN \group_end:
+ \@@_format_aux:
+ }
+\cs_new_protected:Npn \@@_format_aux: { }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Formatting literal units}
+%
+% While in literal mode no parsing occurs, there is a need to provide a few
+% auxiliary functions to handle one or two special cases.
+%
+% \begin{macro}{\@@_literal_power:nN}
+% For printing literal units which are given before the unit they apply to,
+% there is a slight rearrangement.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_literal_power:nN #1#2 { #2 ^ {#1} }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_literal_special:nN}
+% When dealing with the special cases, there is an argument to absorb. This
+% should be braced to be passed up to the user level, which is dealt with
+% here.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_literal_special:nN #1#2 { #1 {#2} }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_format_literal:n}
+% \begin{macro}{\@@_format_literal_auxi:w}
+% \begin{macro}{\@@_format_literal_auxii:w}
+% \begin{macro}{\@@_format_literal_auxiii:w}
+% \begin{macro}{\@@_format_literal_auxiv:w}
+% \begin{macro}{\@@_format_literal_auxv:w}
+% \begin{variable}{\l_@@_separator_tl}
+% To format literal units, there are two tasks to do. The input is
+% \texttt{x}-type expanded to force any symbolic units to be converted into
+% their literal representation: this requires setting the appropriate
+% switch. In the resulting token list, all |.| and |~| tokens are then
+% replaced by the current unit product token list. To enable this to happen
+% correctly with a normal (active) |~|, a small amount of
+% \enquote{protection} is needed first.
+%
+% As with other code dealing with user input, \cs{protected at edef} is used
+% here rather than \cs{tl_set:Nx} as \LaTeXe{} robust commands may be
+% present.
+% \begin{macrocode}
+\group_begin:
+ \char_set_catcode_active:n { `\~ }
+ \cs_new_protected:Npx \@@_format_literal:n #1
+ {
+ \group_begin:
+ \exp_not:n { \bool_set_false:N \l_@@_parsing_bool }
+ \tl_set:Nn \exp_not:N \l_@@_tmp_tl {#1}
+ \tl_replace_all:Nnn \exp_not:N \l_@@_tmp_tl
+ { \exp_not:N ~ } { . }
+ \tl_replace_all:Nnn \exp_not:N \l_@@_tmp_tl
+ { \token_to_str:N ^ } { ^ }
+ \tl_replace_all:Nnn \exp_not:N \l_@@_tmp_tl
+ { \token_to_str:N _ } { \c_@@_math_subscript_tl }
+ \exp_not:n
+ {
+ \protected at edef \l_@@_tmp_tl
+ { \l_@@_tmp_tl }
+ \tl_clear:N \l_@@_formatted_tl
+ \tl_if_empty:NF \l_@@_tmp_tl
+ {
+ \exp_after:wN \@@_format_literal_auxi:w
+ \l_@@_tmp_tl .
+ \q_recursion_tail . \q_recursion_stop
+ }
+ \exp_args:NNNV \group_end:
+ \tl_set:Nn \l_@@_formatted_tl
+ \l_@@_formatted_tl
+ }
+ }
+\group_end:
+% \end{macrocode}
+% To introduce the font changing commands while still allowing for line
+% breaks in literal units, a loop is needed to replace one |.| at a time.
+% To also allow for division, a second loop is used within that to handle
+% |/|: as a result, the separator between parts has to be tracked.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_literal_auxi:w #1 .
+ {
+ \quark_if_recursion_tail_stop:n {#1}
+ \@@_format_literal_auxii:n {#1}
+ \tl_set_eq:NN \l_@@_separator_tl \l_@@_product_tl
+ \@@_format_literal_auxi:w
+ }
+\cs_set_protected:Npn \@@_format_literal_auxii:n #1
+ {
+ \@@_format_literal_auxiii:w
+ #1 / \q_recursion_tail / \q_recursion_stop
+ }
+\cs_new_protected:Npn \@@_format_literal_auxiii:w #1 /
+ {
+ \quark_if_recursion_tail_stop:n {#1}
+ \@@_format_literal_auxiv:w #1 ^ ^ \q_stop
+ \tl_set:Nn \l_@@_separator_tl { / }
+ \@@_format_literal_auxiii:w
+ }
+% \end{macrocode}
+% Within each unit any sub- and superscript parts need to be separated out:
+% wrapping everything in the font command is incorrect. The superscript part
+% is relatively easy as there is no catcode issue or font command to add,
+% while the subscript part needs a bit more work. As the user might have the
+% two parts in either order, picking up the subscript needs to look in two
+% places. We assume that only one is given: odd input here is simply accepted.
+% \begin{macrocode}
+\use:x
+ {
+ \cs_new_protected:Npn \exp_not:N \@@_format_literal_auxiv:w
+ ##1 ^ ##2 ^ ##3 \exp_not:N \q_stop
+ {
+ \exp_not:N \@@_format_literal_auxv:w
+ ##1
+ \c_@@_math_subscript_tl
+ \c_@@_math_subscript_tl
+ \exp_not:N \q_mark
+ ##2
+ \c_@@_math_subscript_tl
+ \c_@@_math_subscript_tl
+ \exp_not:N \q_stop
+ }
+ \cs_new_protected:Npn \exp_not:N \@@_format_literal_auxv:w
+ ##1 \c_@@_math_subscript_tl
+ ##2 \c_@@_math_subscript_tl ##3
+ \exp_not:N \q_mark
+ ##4 \c_@@_math_subscript_tl
+ ##5 \c_@@_math_subscript_tl ##6
+ \exp_not:N \q_stop
+ {
+ \tl_set:Nx \exp_not:N \l_@@_formatted_tl
+ {
+ \exp_not:N \exp_not:V
+ \exp_not:N \l_@@_formatted_tl
+ \exp_not:N \tl_if_empty:NF
+ \exp_not:N \l_@@_formatted_tl
+ {
+ \exp_not:N \exp_not:V
+ \exp_not:N \l_@@_separator_tl
+ }
+ \exp_not:N \tl_if_blank:nF {##1}
+ {
+ \exp_not:N \exp_not:N
+ \exp_not:N \mathrm
+ { \exp_not:N \exp_not:n {##1} }
+ }
+ \exp_not:N \tl_if_blank:nF {##4}
+ { ^ { \exp_not:N \exp_not:n {##4} } }
+ \exp_not:N \tl_if_blank:nF {##2##5}
+ {
+ \c_@@_math_subscript_tl
+ {
+ \exp_not:N \exp_not:N
+ \exp_not:N \mathrm
+ { \exp_not:N \exp_not:n {##2##5} }
+ }
+ }
+ }
+ }
+ }
+\tl_new:N \l_@@_separator_tl
+% \end{macrocode}
+% \end{variable}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Parsing symbolic units}
+%
+% Parsing units takes place by storing information about each unit in a
+% \texttt{prop}. As well as the unit itself, there are various other optional
+% data points, for example a prefix or a power. Some of these can come before
+% the unit, others only after. The parser therefore tracks the number of units
+% read and uses the current position to allocate data to individual units.
+%
+% The result of parsing is a property list (\cs{l_@@_parsed_prop}) which
+% contains one or more entries for each unit:
+% \begin{itemize}
+% \item \texttt{prefix-$n$} The symbol for the prefix which applies to this
+% unit, \emph{e.g.} for \cs{kilo} with (almost certainly) would be
+% |k|.
+% \item \texttt{unit-$n$} The symbol for the unit itself, \emph{e.g.}~for
+% \cs{metre} with (almost certainly) would be |m|.
+% \item \texttt{power-$n$} The power which a unit is raised to. During
+% initial parsing this will (almost certainly) be positive, but is combined
+% with \texttt{per-$n$} to give a \enquote{fully qualified} power before
+% any formatting takes place
+% \item \texttt{per-$n$} Indicates that \texttt{per} applies to the current
+% unit: stored during initial parsing then combined with \texttt{power-$n$}
+% (and removed from the list) before further work.
+% \item \texttt{qualifier-$n$} Any qualifier which applies to the current
+% unit.
+% \item \texttt{special-$n$} Any \enquote{special effect} to apply to the
+% current unit.
+% \end{itemize}
+%
+% \begin{variable}{\l_@@_sticky_per_bool}
+% There is one option when \emph{parsing} the input (as opposed to
+% \emph{formatting} for output): how to deal with \cs{per}.
+% \begin{macrocode}
+\keys_define:nn { siunitx }
+ {
+ sticky-per .bool_set:N = \l_@@_sticky_per_bool
+ }
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_parsed_prop}
+% \begin{variable}{\l_@@_per_bool}
+% \begin{variable}{\l_@@_position_int}
+% Parsing units requires a small number of variables are available: a
+% \texttt{prop} for the parsed units themselves, a \texttt{bool} to
+% indicate if \cs{per} is active and an \texttt{int} to track how many units
+% have be parsed.
+% \begin{macrocode}
+\prop_new:N \l_@@_parsed_prop
+\bool_new:N \l_@@_per_bool
+\int_new:N \l_@@_position_int
+% \end{macrocode}
+% \end{variable}
+% \end{variable}
+% \end{variable}
+%
+% \begin{macro}{\@@_parse:n}
+% The main parsing function is quite simple. After initialising the variables,
+% each symbolic unit is set up. The input is then simply inserted into the
+% input stream: the symbolic units themselves then do the real work of
+% placing data into the parsing system. There is then a bit of tidying up to
+% ensure that later stages can rely on the nature of the data here.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse:n #1
+ {
+ \prop_clear:N \l_@@_parsed_prop
+ \bool_set_true:N \l_@@_parsing_bool
+ \bool_set_false:N \l_@@_per_bool
+ \bool_set_false:N \l_@@_test_bool
+ \int_zero:N \l_@@_position_int
+ #1
+ \int_step_inline:nnnn 1 1 \l_@@_position_int
+ { \@@_parse_finalise:n {##1} }
+ \@@_parse_finalise:
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_add:nnnn}
+% In all cases, storing a data item requires setting a temporary \texttt{tl}
+% which will be used as the key, then using this to store the value. The
+% \texttt{tl} is set using \texttt{x}-type expansion as this will expand the
+% unit index and any additional calculations made for this.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_add:nnnn #1#2#3#4
+ {
+ \tl_set:Nx \l_@@_tmp_tl { #1 - #2 }
+ \prop_if_in:NVTF \l_@@_parsed_prop
+ \l_@@_tmp_tl
+ {
+ \msg_error:nnxx { siunitx } { unit / duplicate-part }
+ { \exp_not:n {#1} } { \token_to_str:N #3 }
+ }
+ {
+ \prop_put:NVn \l_@@_parsed_prop
+ \l_@@_tmp_tl {#4}
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_prefix:Nn}
+% \begin{macro}{\@@_parse_power:nnN}
+% \begin{macro}{\@@_parse_qualifier:nn}
+% \begin{macro}{\@@_parse_special:n}
+% Storage of the various optional items follows broadly the same pattern
+% in each case. The data to be stored is passed along with an appropriate
+% key name to the underlying storage system. The details for each type of
+% item should be relatively clear. For example, prefixes have to come before
+% their \enquote{parent} unit and so there is some adjustment to do to add
+% them to the correct unit.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_prefix:Nn #1#2
+ {
+ \int_set:Nn \l_@@_tmp_int { \l_@@_position_int + 1 }
+ \@@_parse_add:nnnn { prefix }
+ { \int_use:N \l_@@_tmp_int } {#1} {#2}
+ }
+\cs_new_protected:Npn \@@_parse_power:nnN #1#2#3
+ {
+ \tl_set:Nx \l_@@_tmp_tl
+ { unit- \int_use:N \l_@@_position_int }
+ \bool_lazy_or:nnTF
+ {#3}
+ {
+ \prop_if_in_p:NV
+ \l_@@_parsed_prop \l_@@_tmp_tl
+ }
+ {
+ \@@_parse_add:nnnn { power }
+ {
+ \int_eval:n
+ { \l_@@_position_int \bool_if:NT #3 { + 1 } }
+ }
+ {#1} {#2}
+ }
+ {
+ \msg_error:nnxx { siunitx }
+ { unit / part-before-unit } { power } { \token_to_str:N #1 }
+ }
+ }
+\cs_new_protected:Npn \@@_parse_qualifier:nn #1#2
+ {
+ \tl_set:Nx \l_@@_tmp_tl
+ { unit- \int_use:N \l_@@_position_int }
+ \prop_if_in:NVTF \l_@@_parsed_prop \l_@@_tmp_tl
+ {
+ \@@_parse_add:nnnn { qualifier }
+ { \int_use:N \l_@@_position_int } {#1} {#2}
+ }
+ {
+ \msg_error:nnnn { siunitx }
+ { unit / part-before-unit } { qualifier } { \token_to_str:N #1 }
+ }
+ }
+% \end{macrocode}
+% Special (exceptional) items should always come before the relevant units.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_special:n #1
+ {
+ \@@_parse_add:nnnn { special }
+ { \int_eval:n { \l_@@_position_int + 1 } }
+ {#1} {#1}
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_unit:Nn}
+% Parsing units is slightly more involved than the other cases: this is the
+% one place where the tracking value is incremented. If the switch
+% \cs{l_@@_per_bool} is set true then the current unit is also
+% reciprocal: this can only happen if \cs{l_@@_sticky_per_bool} is also
+% true, so only one test is required.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_unit:Nn #1#2
+ {
+ \int_incr:N \l_@@_position_int
+ \@@_parse_add:nnnn { unit }
+ { \int_use:N \l_@@_position_int }
+ {#1} {#2}
+ \bool_if:NT \l_@@_per_bool
+ {
+ \@@_parse_add:nnnn { per }
+ { \int_use:N \l_@@_position_int }
+ { \per } { true }
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_per:}
+% Storing the \cs{per} command requires adding a data item separate from
+% the power which applies: this makes later formatting much more
+% straight-forward. This data could in principle be combined with the
+% \texttt{power}, but depending on the output format required that may make
+% life more complex. Thus this information is stored separately for later
+% retrieval. If \cs{per} is set to be \enquote{sticky} then after parsing
+% the first occurrence, any further uses are in error.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_per:
+ {
+ \bool_if:NTF \l_@@_sticky_per_bool
+ {
+ \bool_set_true:N \l_@@_per_bool
+ \cs_set_protected:Npn \per
+ { \msg_error:nn { siunitx } { unit / duplicate-sticky-per } }
+ }
+ {
+ \@@_parse_add:nnnn
+ { per } { \int_eval:n { \l_@@_position_int + 1 } }
+ { \per } { true }
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_finalise:n}
+% If \cs{per} applies to the current unit, the power needs to be multiplied
+% by $-1$. That is done using an \texttt{fp} operation so that non-integer
+% powers are supported. The flag for \cs{per} is also removed as this means
+% we don't have to check that the original power was positive. To be on
+% the safe side, there is a check for a trivial power at this stage.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_finalise:n #1
+ {
+ \tl_set:Nx \l_@@_tmp_tl { per- #1 }
+ \prop_if_in:NVT \l_@@_parsed_prop \l_@@_tmp_tl
+ {
+ \prop_remove:NV \l_@@_parsed_prop
+ \l_@@_tmp_tl
+ \tl_set:Nx \l_@@_tmp_tl { power- #1 }
+ \prop_get:NVNTF
+ \l_@@_parsed_prop
+ \l_@@_tmp_tl
+ \l_@@_part_tl
+ {
+ \tl_set:Nx \l_@@_part_tl
+ { \fp_eval:n { \l_@@_part_tl * -1 } }
+ \fp_compare:nNnTF \l_@@_part_tl = 1
+ {
+ \prop_remove:NV \l_@@_parsed_prop
+ \l_@@_tmp_tl
+ }
+ {
+ \prop_put:NVV \l_@@_parsed_prop
+ \l_@@_tmp_tl \l_@@_part_tl
+ }
+ }
+ {
+ \prop_put:NVn \l_@@_parsed_prop
+ \l_@@_tmp_tl { -1 }
+ }
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_parse_finalise:}
+% The final task is to check that there is not a \enquote{dangling} power
+% or prefix: these are added to the \enquote{next} unit so are easy to
+% test for.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_parse_finalise:
+ {
+ \clist_map_inline:nn { per , power , prefix }
+ {
+ \tl_set:Nx \l_@@_tmp_tl
+ { ##1 - \int_eval:n { \l_@@_position_int + 1 } }
+ \prop_if_in:NVT \l_@@_parsed_prop \l_@@_tmp_tl
+ { \msg_error:nnn { siunitx } { unit / dangling-part } { ##1 } }
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{Formatting parsed units}
+%
+% \begin{variable}
+% {
+% \l_@@_denominator_bracket_bool ,
+% \l_@@_fraction_function_tl ,
+% \l_@@_bracket_close_tl ,
+% \l_@@_bracket_open_tl ,
+% \l_@@_parse_bool ,
+% \l_@@_per_symbol_tl ,
+% \l_@@_qualifier_mode_tl ,
+% \l_@@_qualifier_phrase_tl
+% }
+% Set up the options which apply to formatting.
+% \begin{macrocode}
+\keys_define:nn { siunitx }
+ {
+ bracket-denominator .bool_set:N =
+ \l_@@_denominator_bracket_bool ,
+ fraction-command .tl_set:N =
+ \l_@@_fraction_function_tl ,
+ parse-units .bool_set:N =
+ \l_@@_parse_bool ,
+ per-mode .choice: ,
+ per-mode / fraction .code:n =
+ {
+ \bool_set_false:N \l_@@_autofrac_bool
+ \bool_set_false:N \l_@@_per_symbol_bool
+ \bool_set_true:N \l_@@_powers_positive_bool
+ \bool_set_true:N \l_@@_two_part_bool
+ } ,
+ per-mode / power .code:n =
+ {
+ \bool_set_false:N \l_@@_autofrac_bool
+ \bool_set_false:N \l_@@_per_symbol_bool
+ \bool_set_false:N \l_@@_powers_positive_bool
+ \bool_set_false:N \l_@@_two_part_bool
+ } ,
+ per-mode / power-positive-first .code:n =
+ {
+ \bool_set_false:N \l_@@_autofrac_bool
+ \bool_set_false:N \l_@@_per_symbol_bool
+ \bool_set_false:N \l_@@_powers_positive_bool
+ \bool_set_true:N \l_@@_two_part_bool
+ } ,
+ per-mode / repeated-symbol .code:n =
+ {
+ \bool_set_false:N \l_@@_autofrac_bool
+ \bool_set_true:N \l_@@_per_symbol_bool
+ \bool_set_true:N \l_@@_powers_positive_bool
+ \bool_set_false:N \l_@@_two_part_bool
+ } ,
+ per-mode / symbol .code:n =
+ {
+ \bool_set_false:N \l_@@_autofrac_bool
+ \bool_set_true:N \l_@@_per_symbol_bool
+ \bool_set_true:N \l_@@_powers_positive_bool
+ \bool_set_true:N \l_@@_two_part_bool
+ } ,
+ per-mode / symbol-or-fraction .code:n =
+ {
+ \bool_set_true:N \l_@@_autofrac_bool
+ \bool_set_true:N \l_@@_per_symbol_bool
+ \bool_set_true:N \l_@@_powers_positive_bool
+ \bool_set_true:N \l_@@_two_part_bool
+ } ,
+ per-symbol .tl_set:N =
+ \l_@@_per_symbol_tl ,
+ qualifier-mode .choice: ,
+ qualifier-mode / bracket .code:n =
+ { \tl_set:Nn \l_@@_qualifier_mode_tl { bracket } } ,
+ qualifier-mode / combine .code:n =
+ { \tl_set:Nn \l_@@_qualifier_mode_tl { combine } } ,
+ qualifier-mode / phrase .code:n =
+ { \tl_set:Nn \l_@@_qualifier_mode_tl { phrase } } ,
+ qualifier-mode / subscript .code:n =
+ { \tl_set:Nn \l_@@_qualifier_mode_tl { subscript } } ,
+ qualifier-phrase .tl_set:N =
+ \l_@@_qualifier_phrase_tl ,
+ unit-close-bracket .tl_set:N =
+ \l_@@_bracket_close_tl ,
+ unit-open-bracket .tl_set:N =
+ \l_@@_bracket_open_tl
+ }
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_bracket_bool}
+% A flag to indicate that the unit currently under construction will require
+% brackets if a power is added.
+% \begin{macrocode}
+\bool_new:N \l_@@_bracket_bool
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_font_bool}
+% A flag to control when font wrapping is applied to the output.
+% \begin{macrocode}
+\bool_new:N \l_@@_font_bool
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}
+% {
+% \l_@@_autofrac_bool ,
+% \l_@@_powers_positive_bool ,
+% \l_@@_per_symbol_bool ,
+% \l_@@_two_part_bool
+% }
+% Dealing with the various ways that reciprocal (\cs{per}) can be handled
+% requires a few different switches.
+% \begin{macrocode}
+\bool_new:N \l_@@_autofrac_bool
+\bool_new:N \l_@@_per_symbol_bool
+\bool_new:N \l_@@_powers_positive_bool
+\bool_new:N \l_@@_two_part_bool
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_numerator_bool}
+% Indicates that the current unit should go into the numerator when splitting
+% into two parts (fractions or other \enquote{sorted} styles).
+% \begin{macrocode}
+\bool_new:N \l_@@_numerator_bool
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_qualifier_mode_tl}
+% For storing the text of options which are best handled by picking
+% function names.
+% \begin{macrocode}
+\tl_new:N \l_@@_qualifier_mode_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_prefix_power_bool}
+% Used to determine if prefixes are converted into powers. Note that
+% while this may be set as an option \enquote{higher up}, at this point it
+% is handled as an internal switch (see the two formatting interfaces for
+% reasons).
+% \begin{macrocode}
+\bool_new:N \l_@@_prefix_power_bool
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_prefix_fp}
+% When converting prefixes to powers, the calculations are done as an
+% \texttt{fp}.
+% \begin{macrocode}
+\fp_new:N \l_@@_prefix_fp
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_current_tl, \l_@@_part_tl}
+% Building up the (partial) formatted unit requires some token list storage.
+% Each part of the unit combination that is recovered also has to be
+% placed in a token list: this is a dedicated one to leave the scratch
+% variables available.
+% \begin{macrocode}
+\tl_new:N \l_@@_current_tl
+\tl_new:N \l_@@_part_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_denominator_tl}
+% For fraction-like units, space is needed for the denominator as well as
+% the numerator (which is handled using \cs{l_@@_formatted_tl}).
+% \begin{macrocode}
+\tl_new:N \l_@@_denominator_tl
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{variable}{\l_@@_total_int}
+% The formatting routine needs to know both the total number of units and
+% the current unit. Thus an \texttt{int} is required in addition to
+% \cs{l_@@_position_int}.
+% \begin{macrocode}
+\int_new:N \l_@@_total_int
+% \end{macrocode}
+% \end{variable}
+%
+% \begin{macro}{\@@_format_parsed:}
+% \begin{macro}{\@@_format_parsed_aux:n}
+% The main formatting routine is essentially a loop over each position,
+% reading the various parts of the unit to build up complete unit
+% combination.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_parsed:
+ {
+ \int_set_eq:NN \l_@@_total_int \l_@@_position_int
+ \tl_clear:N \l_@@_denominator_tl
+ \tl_clear:N \l_@@_formatted_tl
+ \fp_zero:N \l_@@_prefix_fp
+ \int_zero:N \l_@@_position_int
+ \int_do_while:nNnn
+ \l_@@_position_int < \l_@@_total_int
+ {
+ \bool_set_false:N \l_@@_bracket_bool
+ \tl_clear:N \l_@@_current_tl
+ \bool_set_false:N \l_@@_font_bool
+ \bool_set_true:N \l_@@_numerator_bool
+ \int_incr:N \l_@@_position_int
+ \clist_map_inline:nn { prefix , unit , qualifier , power , special }
+ { \@@_format_parsed_aux:n {##1} }
+ \@@_format_output:
+ }
+ \@@_format_finalise:
+ }
+\cs_new_protected:Npn \@@_format_parsed_aux:n #1
+ {
+ \tl_set:Nx \l_@@_tmp_tl
+ { #1 - \int_use:N \l_@@_position_int }
+ \prop_get:NVNT \l_@@_parsed_prop
+ \l_@@_tmp_tl \l_@@_part_tl
+ { \use:c { @@_format_ #1 : } }
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}[EXP]{\@@_format_bracket:N}
+% A quick utility function which wraps up a token list variable in brackets
+% if they are required.
+% \begin{macrocode}
+\cs_new:Npn \@@_format_bracket:N #1
+ {
+ \bool_if:NTF \l_@@_bracket_bool
+ {
+ \exp_not:V \l_@@_bracket_open_tl
+ \exp_not:V #1
+ \exp_not:V \l_@@_bracket_close_tl
+ }
+ { \exp_not:V #1 }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_format_power:}
+% \begin{macro}[EXP]{\@@_format_power_aux:wTF}
+% \begin{macro}
+% {
+% \@@_format_power_positive: ,
+% \@@_format_power_negative:
+% }
+% \begin{macro}[EXP]{\@@_format_power_negative_aux:w}
+% \begin{macro}{\@@_format_power_superscript:}
+% Formatting powers requires a test for negative numbers and depending on
+% output format requests some adjustment to the stored value. This could be
+% done using an \texttt{fp} function, but that would be slow compared to
+% a dedicated if lower-level approach based on delimited arguments.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_power:
+ {
+ \@@_format_font:
+ \exp_after:wN \@@_format_power_aux:wTF
+ \l_@@_part_tl - \q_stop
+ { \@@_format_power_negative: }
+ { \@@_format_power_positive: }
+ }
+\cs_new:Npn \@@_format_power_aux:wTF #1 - #2 \q_stop
+ { \tl_if_empty:nTF {#1} }
+% \end{macrocode}
+% In the case of positive powers, there is little to do: add the power
+% as a subscript (must be required as the parser ensures it's $\neq 1$).
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_power_positive:
+ { \@@_format_power_superscript: }
+% \end{macrocode}
+% Dealing with negative powers starts by flipping the switch used to track
+% where in the final output the current part should get added to. For the
+% case where the output is fraction-like, strip off the |~| then ensure that
+% the result is not the trivial power~$1$. Assuming all is well, addition
+% to the current unit combination goes ahead.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_power_negative:
+ {
+ \bool_set_false:N \l_@@_numerator_bool
+ \bool_if:NTF \l_@@_powers_positive_bool
+ {
+ \tl_set:Nx \l_@@_part_tl
+ {
+ \exp_after:wN \@@_format_power_negative_aux:w
+ \l_@@_part_tl \q_stop
+ }
+ \str_if_eq_x:nnF { \exp_not:V \l_@@_part_tl } { 1 }
+ { \@@_format_power_superscript: }
+ }
+ { \@@_format_power_superscript: }
+ }
+\cs_new:Npn \@@_format_power_negative_aux:w - #1 \q_stop
+ { \exp_not:n {#1} }
+% \end{macrocode}
+% Adding the power as a superscript has the slight complication that there
+% is the possibility of needing some brackets. The superscript itself uses
+% \cs{sp} as that avoids any category code issues and also allows redirection
+% at a higher level more readily.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_power_superscript:
+ {
+ \tl_set:Nx \l_@@_current_tl
+ {
+ \@@_format_bracket:N \l_@@_current_tl
+ ^ { \exp_not:V \l_@@_part_tl }
+ }
+ \bool_set_false:N \l_@@_bracket_bool
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_format_prefix:}
+% \begin{macro}{\@@_format_prefix_power:, \@@_format_prefix_symbol:}
+% Formatting for prefixes depends on whether they are to be expressed as
+% symbols or collected up to be returned as a power of $10$. The latter
+% case requires a bit of processing, which includes checking that the
+% conversion is possible and allowing for any power that applies to the
+% current unit.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_prefix:
+ {
+ \bool_if:NTF \l_@@_prefix_power_bool
+ { \@@_format_prefix_power: }
+ { \@@_format_prefix_symbol: }
+ }
+\cs_new_protected:Npn \@@_format_prefix_power:
+ {
+ \prop_get:NVNTF \l_@@_prefixes_forward_prop
+ \l_@@_part_tl \l_@@_part_tl
+ {
+ \tl_set:Nx \l_@@_tmp_tl
+ { power- \int_use:N \l_@@_position_int }
+ \prop_get:NVNF \l_@@_parsed_prop
+ \l_@@_tmp_tl \l_@@_tmp_tl
+ { \tl_set:Nn \l_@@_tmp_tl { 1 } }
+ \fp_add:Nn \l_@@_prefix_fp
+ { \l_@@_tmp_tl * \l_@@_part_tl }
+ }
+ { \@@_format_prefix_symbol: }
+ }
+\cs_new_protected:Npn \@@_format_prefix_symbol:
+ { \tl_set_eq:NN \l_@@_current_tl \l_@@_part_tl }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_format_qualifier:}
+% \begin{macro}
+% {
+% \@@_format_qualifier_bracket : ,
+% \@@_format_qualifier_combine: ,
+% \@@_format_qualifier_phrase: ,
+% \@@_format_qualifier_subscript:
+% }
+% There are various ways that a qualifier can be added to the output. The
+% idea here is to modify the \enquote{base} text appropriately and then add
+% to the current unit. In the case that a linking phrase is in use, the
+% addition of spaces means that the unit may end up ambiguous, and brackets
+% are therefore required \emph{if} there is a power. Notice that when the
+% qualifier is just treated as \enquote{text}, the auxiliary is actually
+% a no-op.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_qualifier:
+ {
+ \use:c
+ {
+ @@_format_qualifier_
+ \l_@@_qualifier_mode_tl :
+ }
+ \tl_put_right:NV \l_@@_current_tl \l_@@_part_tl
+ }
+\cs_new_protected:Npn \@@_format_qualifier_bracket:
+ {
+ \@@_format_font:
+ \tl_set:Nx \l_@@_part_tl
+ {
+ \exp_not:V \l_@@_bracket_open_tl
+ \exp_not:N \mathrm
+ { \exp_not:V \l_@@_part_tl }
+ \exp_not:V \l_@@_bracket_close_tl
+ }
+ }
+\cs_new_protected:Npn \@@_format_qualifier_combine: { }
+\cs_new_protected:Npn \@@_format_qualifier_phrase:
+ {
+ \@@_format_font:
+ \tl_set:Nx \l_@@_part_tl
+ {
+ \exp_not:V \l_@@_qualifier_phrase_tl
+ \exp_not:N \mathrm
+ { \exp_not:V \l_@@_part_tl }
+ }
+ }
+\cs_new_protected:Npn \@@_format_qualifier_subscript:
+ {
+ \@@_format_font:
+ \tl_set:Nx \l_@@_part_tl
+ {
+ \c_@@_math_subscript_tl
+ {
+ \exp_not:N \mathrm
+ { \exp_not:V \l_@@_part_tl }
+ }
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_format_special:}
+% Any special odds and ends are handled by simply making the current
+% combination into an argument for the recovered code.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_special:
+ {
+ \tl_set:Nx \l_@@_current_tl
+ {
+ \exp_not:V \l_@@_part_tl
+ { \exp_not:V \l_@@_current_tl }
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_format_unit:}
+% A very simple task: add the unit to the output currently being
+% constructed.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_unit:
+ {
+ \tl_put_right:NV
+ \l_@@_current_tl \l_@@_part_tl
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_format_output:}
+% \begin{macro}
+% {\@@_format_output_aux:, \@@_format_output_denominator:}
+% \begin{macro}
+% {
+% \@@_format_output_aux:nn ,
+% \@@_format_output_aux:nV ,
+% \@@_format_output_aux:nv
+% }
+% The first step here is to make a choice based on whether the current
+% part should be stored as part of the numerator or denominator of a
+% fraction. In all cases, if the switch \cs{l_@@_numerator_bool} is
+% true then life is simple: add the current part to the numerator with
+% a standard separator
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_output:
+ {
+ \@@_format_font:
+ \bool_set_false:N \l_@@_bracket_bool
+ \use:c
+ {
+ @@_format_output_
+ \bool_if:NTF \l_@@_numerator_bool
+ { aux: }
+ { denominator: }
+ }
+ }
+\cs_new_protected:Npn \@@_format_output_aux:
+ {
+ \@@_format_output_aux:nV { formatted }
+ \l_@@_product_tl
+ }
+% \end{macrocode}
+% There are a few things to worry about at this stage if the current part
+% is in the denominator. Powers have already been dealt with and some
+% formatting outcomes only need a branch at the final point of building
+% the entire unit. That means that there are three possible outcomes here:
+% if collecting two separate parts, add to the denominator with a product
+% separator, or if only building one token list there may be a need to use
+% a symbol separator. When the |repeated-symbol| option is in use there may
+% be a need to add a leading |1| to the output in the case where the
+% first unit is in the denominator: that can be picked up by looking for
+% empty output in combination with the flag for using a symbol in the output
+% but not a two-part strategy.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_output_denominator:
+ {
+ \bool_if:NTF \l_@@_two_part_bool
+ {
+ \bool_lazy_and:nnT
+ { \l_@@_denominator_bracket_bool }
+ { ! \tl_if_empty_p:N \l_@@_denominator_tl }
+ { \bool_set_true:N \l_@@_bracket_bool }
+ \@@_format_output_aux:nV { denominator }
+ \l_@@_product_tl
+ }
+ {
+ \bool_lazy_and:nnT
+ { \l_@@_per_symbol_bool }
+ { \tl_if_empty_p:N \l_@@_formatted_tl }
+ { \tl_set:Nn \l_@@_formatted_tl { 1 } }
+ \@@_format_output_aux:nv { formatted }
+ {
+ l_@@_
+ \bool_if:NTF \l_@@_per_symbol_bool
+ { per_symbol }
+ { product }
+ _tl
+ }
+ }
+ }
+\cs_new_protected:Npn \@@_format_output_aux:nn #1#2
+ {
+ \tl_set:cx { l_@@_ #1 _tl }
+ {
+ \exp_not:v { l_@@_ #1 _tl }
+ \tl_if_empty:cF { l_@@_ #1 _tl }
+ { \exp_not:n {#2} }
+ \exp_not:V \l_@@_current_tl
+ }
+ }
+\cs_generate_variant:Nn \@@_format_output_aux:nn { nV , nv }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\@@_format_font:}
+% A short auxiliary which checks if the font has been applied to the
+% main part of the output: if not, add it and set the flag.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_font:
+ {
+ \bool_if:NF \l_@@_font_bool
+ {
+ \tl_set:Nx \l_@@_current_tl
+ {
+ \exp_not:N \mathrm
+ { \exp_not:V \l_@@_current_tl }
+ }
+ \bool_set_true:N \l_@@_font_bool
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\@@_format_finalise:}
+% \begin{macro}
+% {
+% \@@_format_finalise_autofrac: ,
+% \@@_format_finalise_fractional: ,
+% \@@_format_finalise_power:
+% }
+% Finalising the unit format is really about picking up the cases involving
+% fractions: these require assembly of the parts with the need to add
+% additional material in some cases
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_finalise:
+ {
+ \tl_if_empty:NF \l_@@_denominator_tl
+ {
+ \bool_if:NTF \l_@@_powers_positive_bool
+ { \@@_format_finalise_fractional: }
+ { \@@_format_finalise_power: }
+ }
+ }
+% \end{macrocode}
+% For fraction-like output, there are three possible choices and two
+% actual styles. In all cases, if the numerator is empty then it is set
+% here to |1|. To deal with the \enquote{auto-format} case, the two
+% styles (fraction and symbol) are handled in auxiliaries: this allows both
+% to be used at the same time! Beyond that, the key here is to use a
+% single \cs{tl_set:Nx} to keep down the number of assignments.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_finalise_fractional:
+ {
+ \tl_if_empty:NT \l_@@_formatted_tl
+ { \tl_set:Nn \l_@@_formatted_tl { 1 } }
+ \bool_if:NTF \l_@@_autofrac_bool
+ { \@@_format_finalise_autofrac: }
+ {
+ \bool_if:NTF \l_@@_per_symbol_bool
+ { \@@_format_finalise_symbol: }
+ { \@@_format_finalise_fraction: }
+ }
+ }
+% \end{macrocode}
+% For the \enquote{auto-selected} fraction method, the two other auxiliary
+% functions are used to do both forms of formatting. So that everything
+% required is available, this needs one group so that the second auxiliary
+% receives the correct input. After that it is just a case of applying
+% \cs{mathchoice} to the formatted output.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_finalise_autofrac:
+ {
+ \group_begin:
+ \@@_format_finalise_fraction:
+ \exp_args:NNNV \group_end:
+ \tl_set:Nn \l_@@_tmp_tl \l_@@_formatted_tl
+ \@@_format_finalise_symbol:
+ \tl_set:Nx \l_@@_formatted_tl
+ {
+ \mathchoice
+ { \exp_not:V \l_@@_tmp_tl }
+ { \exp_not:V \l_@@_formatted_tl }
+ { \exp_not:V \l_@@_formatted_tl }
+ { \exp_not:V \l_@@_formatted_tl }
+ }
+ }
+% \end{macrocode}
+% When using a fraction function the two parts are now assembled.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_finalise_fraction:
+ {
+ \tl_set:Nx \l_@@_formatted_tl
+ {
+ \exp_not:V \l_@@_fraction_function_tl
+ { \exp_not:V \l_@@_formatted_tl }
+ { \exp_not:V \l_@@_denominator_tl }
+ }
+ }
+\cs_new_protected:Npn \@@_format_finalise_symbol:
+ {
+ \tl_set:Nx \l_@@_formatted_tl
+ {
+ \exp_not:V \l_@@_formatted_tl
+ \exp_not:V \l_@@_per_symbol_tl
+ \@@_format_bracket:N \l_@@_denominator_tl
+ }
+ }
+% \end{macrocode}
+% In the case of sorted powers, there is a test to make sure there was
+% at least one positive power, and if so a simple join of the two parts
+% with the appropriate product.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_format_finalise_power:
+ {
+ \tl_if_empty:NTF \l_@@_formatted_tl
+ {
+ \tl_set_eq:NN
+ \l_@@_formatted_tl
+ \l_@@_denominator_tl
+ }
+ {
+ \tl_set:Nx \l_@@_formatted_tl
+ {
+ \exp_not:V \l_@@_formatted_tl
+ \exp_not:V \l_@@_product_tl
+ \exp_not:V \l_@@_denominator_tl
+ }
+ }
+ }
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Pre-defined unit components}
+%
+% Quite a number of units can be predefined: while this is a code-level module,
+% there is little point having a unit parser which does not start off able to
+% parse any units!
+%
+% \begin{macro}
+% {
+% \kilogram ,
+% \metre ,
+% \meter ,
+% \mole ,
+% \kelvin ,
+% \candela ,
+% \second ,
+% \ampere
+% }
+% The basic SI units: technically the correct spelling is \cs{metre} but
+% US users tend to use \cs{meter}.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \kilogram { \kilo \gram }
+\siunitx_declare_unit:Nn \metre { m }
+\siunitx_declare_unit:Nn \meter { \metre }
+\siunitx_declare_unit:Nn \mole { mol }
+\siunitx_declare_unit:Nn \second { s }
+\siunitx_declare_unit:Nn \ampere { A }
+\siunitx_declare_unit:Nn \kelvin { K }
+\siunitx_declare_unit:Nn \candela { cd }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\gram}
+% The gram is an odd unit as it is needed for the base unit kilogram.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \gram { g }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}
+% {
+% \yocto ,
+% \zepto ,
+% \atto ,
+% \femto ,
+% \pico ,
+% \nano ,
+% \micro ,
+% \milli ,
+% \centi ,
+% \deci
+% }
+% The various SI multiple prefixes are defined here: first the small
+% ones.
+% \begin{macrocode}
+\siunitx_declare_prefix:Nnn \yocto { y } { -24 }
+\siunitx_declare_prefix:Nnn \zepto { z } { -21 }
+\siunitx_declare_prefix:Nnn \atto { a } { -18 }
+\siunitx_declare_prefix:Nnn \femto { f } { -15 }
+\siunitx_declare_prefix:Nnn \pico { p } { -12 }
+\siunitx_declare_prefix:Nnn \nano { n } { -9 }
+\siunitx_declare_prefix:Nnn \micro { [micro] } { -6 }
+\siunitx_declare_prefix:Nnn \milli { m } { -3 }
+\siunitx_declare_prefix:Nnn \centi { c } { -2 }
+\siunitx_declare_prefix:Nnn \deci { d } { -1 }
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}
+% {
+% \deca ,
+% \deka ,
+% \hecto ,
+% \kilo ,
+% \mega ,
+% \giga ,
+% \tera ,
+% \peta ,
+% \exa ,
+% \zetta ,
+% \yotta
+% }
+% Now the large ones.
+% \begin{macrocode}
+\siunitx_declare_prefix:Nnn \deca { da } { 1 }
+\siunitx_declare_prefix:Nnn \deka { da } { 1 }
+\siunitx_declare_prefix:Nnn \hecto { h } { 2 }
+\siunitx_declare_prefix:Nnn \kilo { k } { 3 }
+\siunitx_declare_prefix:Nnn \mega { M } { 6 }
+\siunitx_declare_prefix:Nnn \giga { G } { 9 }
+\siunitx_declare_prefix:Nnn \tera { T } { 12 }
+\siunitx_declare_prefix:Nnn \peta { P } { 15 }
+\siunitx_declare_prefix:Nnn \exa { E } { 18 }
+\siunitx_declare_prefix:Nnn \zetta { Z } { 21 }
+\siunitx_declare_prefix:Nnn \yotta { Y } { 24 }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}
+% {
+% \becquerel ,
+% \degreeCelsius ,
+% \coulomb ,
+% \farad ,
+% \gray ,
+% \hertz ,
+% \henry ,
+% \joule ,
+% \katal ,
+% \lumen ,
+% \lux
+% }
+% Named derived units: first half of alphabet.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \becquerel { Bq }
+\siunitx_declare_unit:Nn \degreeCelsius
+ { \ensuremath { { } ^ { \circ } } \kern -\scriptspace C }
+\siunitx_declare_unit:Nn \coulomb { C }
+\siunitx_declare_unit:Nn \farad { F }
+\siunitx_declare_unit:Nn \gray { Gy }
+\siunitx_declare_unit:Nn \hertz { Hz }
+\siunitx_declare_unit:Nn \henry { H }
+\siunitx_declare_unit:Nn \joule { J }
+\siunitx_declare_unit:Nn \katal { kat }
+\siunitx_declare_unit:Nn \lumen { lm }
+\siunitx_declare_unit:Nn \lux { lx }
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}
+% {
+% \newton ,
+% \ohm ,
+% \pascal ,
+% \radian ,
+% \siemens ,
+% \sievert ,
+% \steradian ,
+% \tesla ,
+% \volt ,
+% \watt ,
+% \weber
+% }
+% Named derived units: second half of alphabet.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \newton { N }
+\siunitx_declare_unit:Nn \ohm { \ensuremath { \Omega } }
+\siunitx_declare_unit:Nn \pascal { Pa }
+\siunitx_declare_unit:Nn \radian { rad }
+\siunitx_declare_unit:Nn \siemens { S }
+\siunitx_declare_unit:Nn \sievert { Sv }
+\siunitx_declare_unit:Nn \steradian { sr }
+\siunitx_declare_unit:Nn \tesla { T }
+\siunitx_declare_unit:Nn \volt { V }
+\siunitx_declare_unit:Nn \watt { W }
+\siunitx_declare_unit:Nn \weber { Wb }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}
+% {
+% \day ,
+% \hectare ,
+% \hour ,
+% \litre ,
+% \liter ,
+% \minute ,
+% \tonne
+% }
+% Non-SI, but accepted for general use. Once again there are two
+% spellings, here for litre and with different output in this case.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \day { d }
+\siunitx_declare_unit:Nn \hectare { ha }
+\siunitx_declare_unit:Nn \hour { h }
+\siunitx_declare_unit:Nn \litre { L }
+\siunitx_declare_unit:Nn \liter { \litre }
+\siunitx_declare_unit:Nn \minute { min }
+\siunitx_declare_unit:Nn \tonne { t }
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}
+% {
+% \arcminute ,
+% \arcsecond ,
+% \degree
+% }
+% Arc units: again, non-SI, but accepted for general use.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \arcminute { \ensuremath { ^ { \prime } } }
+\siunitx_declare_unit:Nn \arcsecond { \ensuremath { ^ { \prime \prime } } }
+\siunitx_declare_unit:Nn \degree { \ensuremath { ^ { \circ } } }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\astronomicalunit, \atomicmassunit, \dalton, \electronvolt}
+% A few units based on physical measurements exist: these ones are accepted
+% for use with the International System.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \astronomicalunit { au }
+\siunitx_declare_unit:Nn \atomicmassunit { u }
+\siunitx_declare_unit:Nn \dalton { Da }
+\siunitx_declare_unit:Nn \electronvolt { eV }
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\nuaction, \numass, \nuspeed, \nutime}
+% Natural units based on physical constants.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \nuaction { \ensuremath { \mathit { \hbar } } }
+\siunitx_declare_unit:Nx \numass
+ {
+ \exp_not:N \ensuremath
+ {
+ \exp_not:N \mathit { m }
+ \c_@@_math_subscript_tl { \exp_not:N \mathrm { e } }
+ }
+ }
+\siunitx_declare_unit:Nx \nuspeed
+ {
+ \exp_not:N \ensuremath
+ { \exp_not:N \mathit { c } \c_@@_math_subscript_tl { 0 } }
+ }
+\siunitx_declare_unit:Nn \nutime
+ { \numass \per \numass \per \nuspeed \squared }
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}
+% {
+% \auaction ,
+% \aucharge ,
+% \auenergy ,
+% \aulength ,
+% \aumass ,
+% \autime ,
+% \bohr ,
+% \hartree
+% }
+% Atomic units based on physical constants.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \auaction { \ensuremath { \mathit { \hbar } } }
+\siunitx_declare_unit:Nn \aucharge { \ensuremath { \mathit { e } } }
+\siunitx_declare_unit:Nx \auenergy
+ {
+ \exp_not:N \ensuremath
+ {
+ \exp_not:N \mathit { E }
+ \c_@@_math_subscript_tl { \exp_not:N \mathrm { h } }
+ }
+ }
+\siunitx_declare_unit:Nx \aulength
+ {
+ \exp_not:N \ensuremath
+ { \exp_not:N \mathit { a } \c_@@_math_subscript_tl { 0 } }
+ }
+\siunitx_declare_unit:Nx \aumass
+ {
+ \exp_not:N \ensuremath
+ {
+ \exp_not:N \mathit { m }
+ \c_@@_math_subscript_tl { \exp_not:N \mathrm { e } }
+ }
+ }
+\siunitx_declare_unit:Nn \autime { \auaction \per \auenergy }
+\siunitx_declare_unit:Nn \bohr { \aulength }
+\siunitx_declare_unit:Nn \hartree { \auenergy }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}
+% {
+% \angstrom ,
+% \bar ,
+% \barn ,
+% \bel ,
+% \decibel ,
+% \knot ,
+% \millimetremercury ,
+% \nauticalmile ,
+% \neper
+% }
+% There are then some day-to-day units which are accepted for use
+% with SI, but are not part of the official specification.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \angstrom { \mbox { \AA } }
+\siunitx_declare_unit:Nn \bar { bar }
+\siunitx_declare_unit:Nn \barn { b }
+\siunitx_declare_unit:Nn \bel { B }
+\siunitx_declare_unit:Nn \decibel { \deci \bel }
+\siunitx_declare_unit:Nn \knot { kn }
+\siunitx_declare_unit:Nn \millimetremercury { mmHg }
+\siunitx_declare_unit:Nn \nauticalmile { M }
+\siunitx_declare_unit:Nn \neper { Np }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}
+% {
+% \dyne ,
+% \erg ,
+% \gal ,
+% \gauss ,
+% \maxwell ,
+% \oersted ,
+% \phot ,
+% \poise ,
+% \stilb ,
+% \stokes
+% }
+% \textsc{cgs} units: similar to the set immediately above, these may be used
+% for specific applications.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \dyne { dyn }
+\siunitx_declare_unit:Nn \erg { erg }
+\siunitx_declare_unit:Nn \gal { Gal }
+\siunitx_declare_unit:Nn \gauss { G }
+\siunitx_declare_unit:Nn \maxwell { Mx }
+\siunitx_declare_unit:Nn \oersted { Oe }
+\siunitx_declare_unit:Nn \phot { ph }
+\siunitx_declare_unit:Nn \poise { P }
+\siunitx_declare_unit:Nn \stilb { sb }
+\siunitx_declare_unit:Nn \stokes { St }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\percent}
+% For percent, the raw character is the most flexible way of handling output.
+% \begin{macrocode}
+\siunitx_declare_unit:Nn \percent { \char "25 ~ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\square, \squared, \cubic, \cubed}
+% Basic powers.
+% \begin{macrocode}
+\siunitx_declare_power:NNn \square \squared { 2 }
+\siunitx_declare_power:NNn \cubic \cubed { 3 }
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{Messages}
+%
+% \begin{macrocode}
+\msg_new:nnnn { siunitx } { unit / dangling-part }
+ { Found~#1~part~with~no~unit. }
+ {
+ Each~#1~part~must~be~associated~with~a~unit:~a~#1~part~was~found~
+ but~no~following~unit~was~given.
+ }
+\msg_new:nnnn { siunitx } { unit / duplicate-part }
+ { Duplicate~#1~part:~#2. }
+ {
+ Each~unit~may~have~only~one~#1:\\
+ the~additional~#1~part~'#2'~will~be~ignored.
+ }
+\msg_new:nnnn { siunitx } { unit / duplicate-sticky-per }
+ { Duplicate~\token_to_str:N \per. }
+ {
+ When~the~'sticky-per'~option~is~active,~only~one~
+ \token_to_str:N \per \ may~appear~in~a~unit.
+ }
+\msg_new:nnnn { siunitx } { unit / part-before-unit }
+ { Found~#1~part~before~first~unit:~#2. }
+ {
+ The~#1~part~'#2'~must~follow~after~a~unit:~
+ it~cannot~appear~before~any~units~and~will~therefore~be~ignored.
+ }
+% \end{macrocode}
+%
+% \subsection{Standard settings for module options}
+%
+% Some of these follow naturally from the point of definition
+% (\emph{e.g.}~boolean variables are always |false| to begin with),
+% but for clarity everything is set here.
+% \begin{macrocode}
+\keys_set:nn { siunitx }
+ {
+ bracket-denominator = true ,
+ fraction-command = \frac ,
+ parse-units = true ,
+ per-mode = power ,
+ per-symbol = / ,
+ qualifier-mode = subscript ,
+ qualifier-phrase = \ ,
+ sticky-per = false ,
+ unit-close-bracket = ) , % (
+ unit-open-bracket = ( , % )
+ unit-product = \,
+ }
+% \end{macrocode}
+%
+% \begin{macrocode}
+%</package>
+% \end{macrocode}
+%
+% \end{implementation}
+%
+% \begin{thebibliography}{1}
+% \bibitem{BIPM}
+% \emph{The International System of Units (SI)},
+% \url{https://www.bipm.org/en/measurement-units/}.
+% \bibitem{SI:2.1}
+% \emph{SI base units},
+% \url{https://www.bipm.org/en/publications/si-brochure/section2-1.html}.
+% \bibitem{SI:2.2.2}
+% \emph{Units with special names and symbols; units that
+% incorporate special names and symbols},
+% \url{https://www.bipm.org/en/publications/si-brochure/section2-2-2.html}.
+% \bibitem{SI:3.1}
+% \emph{SI Prefixes},
+% \url{https://www.bipm.org/en/publications/si-brochure/chapter3.html}.
+% \bibitem{SI:5.3.7}
+% \emph{Stating values of dimensionless quantities, or quantities of
+% dimension one},
+% \url{https://www.bipm.org/en/publications/si-brochure/section5-3-7.html}.
+% \bibitem{SI:T6}
+% \emph{Non-SI units accepted for use with the International
+% System of Units},
+% \url{https://www.bipm.org/en/publications/si-brochure/table6.html}.
+% \bibitem{SI:T7}
+% \emph{Non-SI units whose values in SI units must be obtained
+% experimentally},
+% \url{https://www.bipm.org/en/publications/si-brochure/table7.html}.
+% \bibitem{SI:T8}
+% \emph{Other non-SI units},
+% \url{https://www.bipm.org/en/publications/si-brochure/table8.html}.
+% \bibitem{SI:T9}
+% \emph{Non-SI units associated with the CGS and the CGS-Gaussian
+% system of units},
+% \url{https://www.bipm.org/en/publications/si-brochure/table9.html}.
+% \end{thebibliography}
+%
+% \PrintIndex
Property changes on: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-unit.dtx
___________________________________________________________________
Added: svn:eol-style
## -0,0 +1 ##
+native
\ No newline at end of property
Added: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v1.dtx
===================================================================
--- trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v1.dtx (rev 0)
+++ trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v1.dtx 2018-04-29 22:36:26 UTC (rev 47489)
@@ -0,0 +1,85 @@
+% \iffalse meta-comment
+%
+% File: siunitx-v1.dtx Copyright (C) 2016-2017 Joseph Wright
+%
+% It may be distributed and/or modified under the conditions of the
+% LaTeX Project Public License (LPPL), either version 1.3c of this
+% license or (at your option) any later version. The latest version
+% of this license is in the file
+%
+% https://www.latex-project.org/lppl.txt
+%
+% This file is part of the "siunitx bundle" (The Work in LPPL)
+% and all files in that bundle must be distributed together.
+%
+% The released version of this bundle is available from CTAN.
+%
+% -----------------------------------------------------------------------
+%
+% The development version of the bundle can be found at
+%
+% https://github.com/josephwright/siunitx
+%
+% for those people who are interested.
+%
+% -----------------------------------------------------------------------
+%
+%<*driver>
+\documentclass{l3doc}
+% The next line is needed so that \GetFileInfo will be able to pick up
+% version data
+\usepackage{siunitx}
+\begin{document}
+ \DocInput{\jobname.dtx}
+\end{document}
+%</driver>
+% \fi
+%
+% \GetFileInfo{siunitx.sty}
+%
+% \title{^^A
+% \pkg{siunitx-v1} -- Version~1 compatibility^^A
+% \thanks{This file describes \fileversion,
+% last revised \filedate.}^^A
+% }
+%
+% \author{^^A
+% Joseph Wright^^A
+% \thanks{^^A
+% E-mail:
+% \href{mailto:joseph.wright at morningstar2.co.uk}
+% {joseph.wright at morningstar2.co.uk}^^A
+% }^^A
+% }
+%
+% \date{Released \filedate}
+%
+% \maketitle
+%
+% \begin{documentation}
+%
+% \end{documentation}
+%
+% \begin{implementation}
+%
+% \section{\pkg{siunitx-v1} implementation}
+%
+% Start the \pkg{DocStrip} guards.
+% \begin{macrocode}
+%<*cfg>
+% \end{macrocode}
+%
+% Identify the internal prefix (\LaTeX3 \pkg{DocStrip} convention). In contrast
+% to other parts of the bundle, the functions here may need to redefine those
+% from various submodules.
+% \begin{macrocode}
+%<@@=siunitx>
+% \end{macrocode}
+%
+% \begin{macrocode}
+%</cfg>
+% \end{macrocode}
+%
+% \end{implementation}
+%
+% \PrintIndex
\ No newline at end of file
Property changes on: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v1.dtx
___________________________________________________________________
Added: svn:eol-style
## -0,0 +1 ##
+native
\ No newline at end of property
Added: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v2.dtx
===================================================================
--- trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v2.dtx (rev 0)
+++ trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v2.dtx 2018-04-29 22:36:26 UTC (rev 47489)
@@ -0,0 +1,97 @@
+% \iffalse meta-comment
+%
+% File: siunitx-v2.dtx Copyright (C) 2016-2017 Joseph Wright
+%
+% It may be distributed and/or modified under the conditions of the
+% LaTeX Project Public License (LPPL), either version 1.3c of this
+% license or (at your option) any later version. The latest version
+% of this license is in the file
+%
+% https://www.latex-project.org/lppl.txt
+%
+% This file is part of the "siunitx bundle" (The Work in LPPL)
+% and all files in that bundle must be distributed together.
+%
+% The released version of this bundle is available from CTAN.
+%
+% -----------------------------------------------------------------------
+%
+% The development version of the bundle can be found at
+%
+% https://github.com/josephwright/siunitx
+%
+% for those people who are interested.
+%
+% -----------------------------------------------------------------------
+%
+%<*driver>
+\documentclass{l3doc}
+% The next line is needed so that \GetFileInfo will be able to pick up
+% version data
+\usepackage{siunitx}
+\begin{document}
+ \DocInput{\jobname.dtx}
+\end{document}
+%</driver>
+% \fi
+%
+% \GetFileInfo{siunitx.sty}
+%
+% \title{^^A
+% \pkg{siunitx-v2} -- Version~2 compatibility^^A
+% \thanks{This file describes \fileversion,
+% last revised \filedate.}^^A
+% }
+%
+% \author{^^A
+% Joseph Wright^^A
+% \thanks{^^A
+% E-mail:
+% \href{mailto:joseph.wright at morningstar2.co.uk}
+% {joseph.wright at morningstar2.co.uk}^^A
+% }^^A
+% }
+%
+% \date{Released \filedate}
+%
+% \maketitle
+%
+% \begin{documentation}
+%
+% \end{documentation}
+%
+% \begin{implementation}
+%
+% \section{\pkg{siunitx-v2} implementation}
+%
+% Start the \pkg{DocStrip} guards.
+% \begin{macrocode}
+%<*cfg>
+% \end{macrocode}
+%
+% Identify the internal prefix (\LaTeX3 \pkg{DocStrip} convention). In contrast
+% to other parts of the bundle, the functions here may need to redefine those
+% from various submodules.
+% \begin{macrocode}
+%<@@=siunitx>
+% \end{macrocode}
+%
+% The old |s| column type is handled by using the functionality of
+% \pkg{collcell}.
+% \begin{macrocode}
+\RequirePackage { collcell }
+\AtBeginDocument
+ {
+ \@@_declare_column:Nnn s
+ { \collectcell \unit }
+ { \endcollectcell }
+ }
+% \end{macrocode}
+%
+% \begin{macrocode}
+%</cfg>
+% \end{macrocode}
+%
+% \end{implementation}
+%
+% \PrintIndex
\ No newline at end of file
Property changes on: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v2.dtx
___________________________________________________________________
Added: svn:eol-style
## -0,0 +1 ##
+native
\ No newline at end of property
Added: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v3.dtx
===================================================================
--- trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v3.dtx (rev 0)
+++ trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v3.dtx 2018-04-29 22:36:26 UTC (rev 47489)
@@ -0,0 +1,347 @@
+% \iffalse meta-comment
+%
+% File: siunitx-v3.dtx Copyright (C) 2016-2017 Joseph Wright
+%
+% It may be distributed and/or modified under the conditions of the
+% LaTeX Project Public License (LPPL), either version 1.3c of this
+% license or (at your option) any later version. The latest version
+% of this license is in the file
+%
+% https://www.latex-project.org/lppl.txt
+%
+% This file is part of the "siunitx bundle" (The Work in LPPL)
+% and all files in that bundle must be distributed together.
+%
+% The released version of this bundle is available from CTAN.
+%
+% -----------------------------------------------------------------------
+%
+% The development version of the bundle can be found at
+%
+% https://github.com/josephwright/siunitx
+%
+% for those people who are interested.
+%
+% -----------------------------------------------------------------------
+%
+%<*driver>
+\documentclass{l3doc}
+% The next line is needed so that \GetFileInfo will be able to pick up
+% version data
+\usepackage{siunitx}
+\begin{document}
+ \DocInput{\jobname.dtx}
+\end{document}
+%</driver>
+% \fi
+%
+% \GetFileInfo{siunitx.sty}
+%
+% \title{^^A
+% \pkg{siunitx} -- Overall set up^^A
+% \thanks{This file describes \fileversion,
+% last revised \filedate.}^^A
+% }
+%
+% \author{^^A
+% Joseph Wright^^A
+% \thanks{^^A
+% E-mail:
+% \href{mailto:joseph.wright at morningstar2.co.uk}
+% {joseph.wright at morningstar2.co.uk}^^A
+% }^^A
+% }
+%
+% \date{Released \filedate}
+%
+% \maketitle
+%
+% \begin{documentation}
+%
+% \end{documentation}
+%
+% \begin{implementation}
+%
+% \section{\pkg{siunitx} implementation}
+%
+% Start the \pkg{DocStrip} guards.
+% \begin{macrocode}
+%<*package>
+% \end{macrocode}
+%
+% Identify the internal prefix (\LaTeX3 \pkg{DocStrip} convention).
+% \begin{macrocode}
+%<@@=siunitx>
+% \end{macrocode}
+%
+% \subsection{Initial set up}
+%
+% Load only the essential support (\pkg{expl3}) \enquote{up-front}.
+% \begin{macrocode}
+\RequirePackage{expl3}
+% \end{macrocode}
+%
+% Make sure that the version of \pkg{l3kernel} in use is sufficiently new.
+% This will also trap any problems with \pkg{l3packages} (as the two are now
+% tied together, version-wise).
+% \begin{macrocode}
+\@ifpackagelater {expl3}{2015/11/15}
+ {}
+ {%
+ \PackageError{siunitx} {Support package expl3 too old}
+ {%
+ You need to update your installation of the bundles 'l3kernel' and
+ 'l3packages'.\MessageBreak
+ Loading~siunitx~will~abort!%
+ }%
+ \endinput
+ }%
+% \end{macrocode}
+%
+% Identify the package and give the over all version information.
+% \begin{macrocode}
+\ProvidesExplPackage {siunitx} {2017/11/26} {3.0alpha}
+ {A comprehensive (SI) units package}
+% \end{macrocode}
+%
+% \subsection{Safety checks}
+%
+% \begin{macro}{\@@_load_check:}
+% There are a number of packages that are incompatible with \pkg{siunitx}
+% as they cover the same concepts and in some cases define the same command
+% names. These are all tested at the point of loading to try to trap
+% issues, and a couple are also tested later as it's possible for them to
+% load without an obvious error if \pkg{siunitx} was loaded first.
+%
+% The testing here is done in a group so that the tests do not add
+% anything to the hash table.
+% \begin{macrocode}
+\msg_new:nnnn { siunitx } { incompatible-package }
+ { Package~'#1'~incompatible. }
+ { The~#1~package~and~siunitx~are~incompatible. }
+\cs_new_protected:Npn \@@_load_check:n #1
+ {
+ \group_begin:
+ \@ifpackageloaded {#1}
+ {
+ \group_end:
+ \msg_error:nnx { siunitx } { incompatible-package } {#1}
+ }
+ { \group_end: }
+ }
+\clist_map_function:nN
+ { SIunits , sistyle , unitsdef , fancyunits }
+ \@@_load_check:n
+\AtBeginDocument
+ {
+ \clist_map_function:nN { SIunits , sistyle }
+ \@@_load_check:n
+ }
+% \end{macrocode}
+%\end{macro}
+%
+% \subsection{Top-level scratch space}
+%
+% \begin{macro}{\l_@@_tmp_tl}
+% Scratch space for the interfaces.
+% \begin{macrocode}
+\tl_new:N \l_@@_tmp_tl
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{User interfaces}
+%
+% The user interfaces are defined in terms of documented code-level ones.
+% This is all done here, and will appear in the \file{.sty} file before the
+% relevant code. Things could be re-arranged by \pkg{DocStrip} but there is no
+% advantage.
+%
+% User level interfaces are all created by \pkg{xparse}
+% \begin{macrocode}
+\RequirePackage { xparse }
+% \end{macrocode}
+%
+% \subsubsection{Preamble commands}
+%
+% \begin{macro}
+% {
+% \DeclareSIPower ,
+% \DeclareSIPrefix ,
+% \DeclareSIQualifier ,
+% \DeclareSIUnit
+% }
+% Pass data to the code layer.
+% \begin{macrocode}
+\NewDocumentCommand \DeclareSIPower { +m +m m }
+ {
+ \siunitx_declare_power:NNn #1 #2 {#3}
+ }
+\NewDocumentCommand \DeclareSIPrefix { +m m m }
+ {
+ \siunitx_declare_prefix:Nnn #1 {#2} {#3}
+ }
+\NewDocumentCommand \DeclareSIQualifier { +m m }
+ {
+ \siunitx_declare_qualifier:Nn #1 {#2}
+ }
+\NewDocumentCommand \DeclareSIUnit { O { } +m m }
+ {
+ \siunitx_declare_unit:Nn #2 {#3}
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \subsubsection{Document commands}
+%
+% \begin{macro}{\qty}
+% \begin{macrocode}
+\NewDocumentCommand \qty { O { } m m }
+ {
+ \leavevmode
+ \group_begin:
+ \keys_set:nn { siunitx } {#1}
+ \siunitx_number_format:nN {#2} \l_@@_tmp_tl
+ \siunitx_print:nV { number } \l_@@_tmp_tl
+ \, \nobreak % TEMP
+ \siunitx_unit_format:nN {#3} \l_@@_tmp_tl
+ \siunitx_print:nV { unit } \l_@@_tmp_tl
+ \group_end:
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\num, \unit}
+% All of a standard form: start a paragraph (if required), set local
+% key values, do the formatting, print the result.
+% \begin{macrocode}
+\NewDocumentCommand \num { O { } m }
+ {
+ \leavevmode
+ \group_begin:
+ \keys_set:nn { siunitx } {#1}
+ \siunitx_number_format:nN {#2} \l_@@_tmp_tl
+ \siunitx_print:nV { number } \l_@@_tmp_tl
+ \group_end:
+ }
+\NewDocumentCommand \unit { O { } m }
+ {
+ \leavevmode
+ \group_begin:
+ \keys_set:nn { siunitx } {#1}
+ \siunitx_unit_format:nN {#2} \l_@@_tmp_tl
+ \siunitx_print:nV { unit } \l_@@_tmp_tl
+ \group_end:
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\sisetup}
+% A very thin wrapper.
+% \begin{macrocode}
+\NewDocumentCommand \sisetup { m }
+ { \keys_set:nn { siunitx } {#1} }
+% \end{macrocode}
+% \end{macro}
+%
+%
+% \subsubsection{Table column}
+%
+% User interfaces in tabular constructs are provided using the mechanisms from
+% the \pkg{array} package.
+% \begin{macrocode}
+\RequirePackage { array }
+% \end{macrocode}
+%
+% There is a slight problem with the \pkg{cellspace} package: it also uses
+% \texttt{S} for a column type Here, \texttt{S} seems to make more sense for
+% \pkg{siunitx}, with \texttt{C} used for \pkg{cellspace}. To enable this to
+% work well, the column rewriting code happens \cs{AtBeginDocument}. The
+% \texttt{S} column is deleted from \cs{NC at list} so that no warning appears.
+% \begin{macrocode}
+\AtBeginDocument
+ {
+ \@ifpackageloaded { cellspace }
+ {
+ \newcolumntype { C } [ 1 ]
+ { > { \bcolumn #1 \@nil } #1 < { \ecolumn } }
+ \cs_set:Npn \@@_tmp:w #1 \NC at do S #2 \q_stop
+ { \NC at list { #1 #2 } }
+ \exp_after:wN \@@_tmp:w \tex_the:D \NC at list \q_stop
+ \cs_undefine:N \NC at find@S
+ \msg_warning:nnn { siunitx } { moved-cellspace-column } { C }
+ \ifcellspace at m
+ \def \env at matrix
+ {
+ \hskip -\arraycolsep
+ \let \@ifnextchar \new at ifnextchar
+ \array
+ {
+ * { \c at MaxMatrixCols }
+ { > { \bcolumn c \@nil $ } c < { $ \ecolumn } } @ { }
+ }
+ }
+ \fi
+ }
+ { }
+ }
+\msg_new:nnn { siunitx } { moved-cellspace-column }
+ { Column~type~for~cellspace~package~moved~to~'#1'. }
+% \end{macrocode}
+%
+% \begin{macro}{\@@_declare_column:Nnn}
+% Creating numerical columns requires that these are declared before anything
+% else in \cs{NC at list}: this is necessary to work with optional arguments.
+% This means a bit of manual effort after the simple declaration of a new
+% column type. The token assigned to the column type is not fixed as this
+% allows the same code to be used in compatibility with version~$2$.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_declare_column:Nnn #1#2#3
+ {
+ \newcolumntype {#1} { }
+ \cs_set_protected:Npn \@@_tmp:w \NC at do ##1##2 \NC at do #1
+ { \NC at list { \NC at do ##1 \NC at do #1 ##2 } }
+ \exp_after:wN \@@_tmp:w \tex_the:D \NC at list
+ \exp_args:NNc \renewcommand * { NC at rewrite@ #1 } [ 1 ] [ ]
+ {
+ \@temptokena \expandafter
+ {
+ \the \@temptokena
+ > {#2} c < {#3}
+ }
+ \NC at find
+ }
+ }
+% \end{macrocode}
+% When \pkg{mdwtab} is loaded the syntax required is slightly different.
+% \begin{macrocode}
+\AtBeginDocument
+ {
+ \@ifpackageloaded { mdwtab }
+ {
+ \cs_set_protected:Npn \@@_declare_column:Nnn #1#2#3
+ {
+ \newcolumntype {#1} [ 1 ] [ ]
+ { > {#2} c < {#3} }
+ }
+ }
+ { }
+ }
+\AtBeginDocument
+ {
+ \@@_declare_column:Nnn n
+ {
+ \keys_set:nn { siunitx } {#1}
+ \siunitx_cell_begin:
+ }
+ { \siunitx_cell_end: }
+ }
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macrocode}
+%</package>
+% \end{macrocode}
+%
+% \end{implementation}
+%
+% \PrintIndex
Property changes on: trunk/Master/texmf-dist/source/latex/siunitx/siunitx-v3.dtx
___________________________________________________________________
Added: svn:eol-style
## -0,0 +1 ##
+native
\ No newline at end of property
Modified: trunk/Master/texmf-dist/source/latex/siunitx/siunitx.dtx
===================================================================
--- trunk/Master/texmf-dist/source/latex/siunitx/siunitx.dtx 2018-04-29 22:36:02 UTC (rev 47488)
+++ trunk/Master/texmf-dist/source/latex/siunitx/siunitx.dtx 2018-04-29 22:36:26 UTC (rev 47489)
@@ -76,7 +76,7 @@
Maintained by Joseph Wright
E-mail: joseph.wright at morningstar2.co.uk
Released under the LaTeX Project Public License v1.3c or later
-See http://www.latex-project.org/lppl.txt
+See https://www.latex-project.org/lppl.txt
---------------------------------------------------------------
\endpreamble
@@ -88,7 +88,7 @@
the LaTeX Project Public License (LPPL), either version 1.3c of
this license or (at your option) any later version. The latest
version of this license is in the file:
- http://www.latex-project.org/lppl.txt
+ https://www.latex-project.org/lppl.txt
This work is "maintained" (as per LPPL maintenance status) by
Joseph Wright.
@@ -670,6 +670,8 @@
% \changes{v2.7i}{2017/07/23}{Rename an incorrect variable}
% \changes{v2.7k}{2017/11/01}{Fix issues when units are not parsed}
% \changes{v2.7l}{2018/03/11}{Track \pkg{unicode-math} changes}
+% \changes{v2.7n}{2018/04/27}{Fix issue with symbolic input
+% and digit grouping}
%
%\begin{abstract}
% Physical quantities have both numbers and units, and each physical
@@ -5218,7 +5220,7 @@
%
% The usual preliminaries.
% \begin{macrocode}
-\ProvidesExplPackage {siunitx} {2018-03-20} {2.7m}
+\ProvidesExplPackage {siunitx} {2018-04-27} {2.7n}
{A comprehensive (SI) units package}
% \end{macrocode}
%
@@ -7263,7 +7265,7 @@
{
\prop_put:Nnn \l_@@_number_in_prop { symbolic }
{ true }
- \@@_number_in_digit:n { \ensuremath {#1} }
+ \@@_number_in_digit:n {#1}
}
{
\tl_if_in:NnF \l_@@_input_ignore_tl {#1}
Added: trunk/Master/texmf-dist/source/latex/siunitx/siunitx.ins
===================================================================
--- trunk/Master/texmf-dist/source/latex/siunitx/siunitx.ins (rev 0)
+++ trunk/Master/texmf-dist/source/latex/siunitx/siunitx.ins 2018-04-29 22:36:26 UTC (rev 47489)
@@ -0,0 +1,83 @@
+\iffalse meta-comment
+
+File: siunitx.ins Copyright (C) 2014-2017 Joseph Wright
+
+It may be distributed and/or modified under the conditions of the
+LaTeX Project Public License (LPPL), either version 1.3c of this
+license or (at your option) any later version. The latest version
+of this license is in the file
+
+ https://www.latex-project.org/lppl.txt
+
+This file is part of the "siunitx bundle" (The Work in LPPL)
+and all files in that bundle must be distributed together.
+
+The released version of this bundle is available from CTAN.
+
+-----------------------------------------------------------------------
+
+The development version of the bundle can be found at
+
+ https://github.com/josephwright/siunitx
+
+for those people who are interested.
+
+-----------------------------------------------------------------------
+
+\fi
+
+\input l3docstrip.tex
+\askforoverwritefalse
+
+% stop DocStrip adding rather wordy text
+\preamble
+\endpreamble
+\postamble
+Copyright (C) 2008-2017 by
+ Joseph Wright <joseph.wright at morningstar2.co.uk>
+
+It may be distributed and/or modified under the conditions of
+the LaTeX Project Public License (LPPL), either version 1.3c of
+this license or (at your option) any later version. The latest
+version of this license is in the file:
+
+ https://www.latex-project.org/lppl.txt
+
+This work is "maintained" (as per LPPL maintenance status) by
+ Joseph Wright.
+
+This work consists of the files siunitx.dtx,
+ siunitx.ins,
+ siunitx-v3.dtx,
+ siunitx-angle.dtx,
+ siunitx-number.dtx,
+ siunitx-print.dtx,
+ siunitx-table.dtx,
+ siunitx-unit.dtx,
+ siunitx-v1.dtx and
+ siunitx-v2.dtx,
+ and the derived files siunitx.pdf,
+ siunitx.sty,
+ siunitx-v1.cfg and
+ siunitx-v2.cfg.
+
+\endpostamble
+
+\keepsilent
+
+\generate
+ {%
+ \file{siunitx.sty}
+ {%
+ \from{siunitx-v3.dtx} {package}
+ \from{siunitx-angle.dtx} {package}
+ \from{siunitx-number.dtx}{package}
+ \from{siunitx-print.dtx} {package}
+ \from{siunitx-table.dtx} {package}
+ \from{siunitx-unit.dtx} {package}
+ }%
+ \file{siunitx-v2.cfg}
+ {\from{siunitx-v2.dtx}{cfg}}%
+ }
+
+\endbatchfile
Modified: trunk/Master/texmf-dist/tex/latex/siunitx/siunitx.sty
===================================================================
--- trunk/Master/texmf-dist/tex/latex/siunitx/siunitx.sty 2018-04-29 22:36:02 UTC (rev 47488)
+++ trunk/Master/texmf-dist/tex/latex/siunitx/siunitx.sty 2018-04-29 22:36:26 UTC (rev 47489)
@@ -15,7 +15,7 @@
%%
\RequirePackage{expl3}[2015/09/11]
\RequirePackage{xparse}
-\ProvidesExplPackage {siunitx} {2018-03-20} {2.7m}
+\ProvidesExplPackage {siunitx} {2018-04-27} {2.7n}
{A comprehensive (SI) units package}
\@ifpackagelater { expl3 } { 2015/11/15 }
{ }
@@ -1270,7 +1270,7 @@
{
\prop_put:Nnn \l__siunitx_number_in_prop { symbolic }
{ true }
- \__siunitx_number_in_digit:n { \ensuremath {#1} }
+ \__siunitx_number_in_digit:n {#1}
}
{
\tl_if_in:NnF \l__siunitx_input_ignore_tl {#1}
More information about the tex-live-commits
mailing list