[latex3-commits] [git/LaTeX3-latex3-latex2e] gh253: implementation for #253 (bb252af8)

[Frank Mittelbach]

Repository : https://github.com/latex3/latex2e
On branch  : gh253
Link       : https://github.com/latex3/latex2e/commit/bb252af8f6ef14fc8de0f4253a3907a096acd203

>---------------------------------------------------------------

commit bb252af8f6ef14fc8de0f4253a3907a096acd203
Author: Frank Mittelbach <frank.mittelbach at latex-project.org>
Date:   Tue Oct 5 11:55:08 2021 +0200

    implementation for #253


>---------------------------------------------------------------

bb252af8f6ef14fc8de0f4253a3907a096acd203
 base/doc/ltnews34.tex             | 17 +++++++++++
 base/ltoutenc.dtx                 | 60 ++++++++++++++++++++++++++++-----------
 base/testfiles-TU/github-0253.lvt | 31 ++++++++++++++++++++
 base/testfiles-TU/github-0253.tlg | 26 +++++++++++++++++
 4 files changed, 117 insertions(+), 17 deletions(-)

diff --git a/base/doc/ltnews34.tex b/base/doc/ltnews34.tex
index 57e101e8..f4d78bdd 100644
--- a/base/doc/ltnews34.tex
+++ b/base/doc/ltnews34.tex
@@ -481,6 +481,23 @@ whether used or not.
 
 
 
+\subsection{Cleanup of the Unicode declaration interface}
+
+When declaring encoding specific commands for the Unicode (TU)
+encoding some declarations, such as \cs{DeclareUnicodeComposite} do
+not have the encoding name as an explicit argument but use
+\cs{UnicodeEncodingName} implicitly. Others, such as
+\cs{DeclareUnicodeAccent} require it as an explicit argument.  This
+inconsistency has now been removed and the encoding name is always
+implicit. To avoid a breaking change for a few packages on CTAN
+\cs{DeclareUnicodeAccent} still accepts three arguments if the second
+argument is \texttt{TU} or \cs{UnicodeEncodingName}. Once all packages
+have been updated this code branch will get removed.
+%
+\githubissue{253}
+
+
+
 \subsection{???}
 
 %
diff --git a/base/ltoutenc.dtx b/base/ltoutenc.dtx
index 14518d2f..58e63788 100644
--- a/base/ltoutenc.dtx
+++ b/base/ltoutenc.dtx
@@ -3016,8 +3016,31 @@
   \char#1\relax}
 %    \end{macrocode}
 %
-%    \begin{macrocode}
-\def\DeclareUnicodeAccent#1#2#3{%
+%    In its original implementation \cs{DeclareUnicodeAccent} was
+%    given 3 arguments (with second the ``Unicode encoding'' a.k.a.,
+%    \cs{UnicodeEncodingName}) while in other places, e.g.,
+%    \cs{DeclareUnicodeComposite}, we always made encoding implicit. So
+%    we now change it here to implicit too so that the interfaces
+%    become a bit more consistent. To avoid making that a
+%    breaking change (even though it only affects two packages on
+%    CTAN) we test for \verb=#2= being \cs{UnicodeEncodingName}. This
+%    would not catch if somebody used
+%    \verb|\DeclareUnicodeAccent{\=}{TU-sub}{"0304}|
+%    but that fortunately hasn't happened. With the implicit argument
+%    you would need to change \cs{UnicodeEncodingName} instead, as you
+%    have to do anyway for the other interface commands.
+%    \begin{macrocode}
+\def\DeclareUnicodeAccent#1#2{%
+  \edef\reserved at a{#2}%
+  \edef\reserved at b{\UnicodeEncodingName}%
+  \ifx\reserved at a\reserved at b
+    \def\reserved at a{\DeclareUnicodeAccent@{#1}}%
+  \else
+    \def\reserved at a{\DeclareUnicodeAccent@{#1}\UnicodeEncodingName}%
+  \fi
+  \reserved at a{#2}%
+}  
+\def\DeclareUnicodeAccent@#1#2#3{%
   \DeclareTextCommand{#1}{#2}{\add at unicode@accent{#3}}%
 }
 %    \end{macrocode}
@@ -3264,21 +3287,24 @@
 %    \end{macrocode}
 % Accents must be declared before the composites that use them.
 %    \begin{macrocode}
-\DeclareUnicodeAccent{\`}                \UnicodeEncodingName{"0300}
-\DeclareUnicodeAccent{\'}                \UnicodeEncodingName{"0301}
-\DeclareUnicodeAccent{\^}                \UnicodeEncodingName{"0302}
-\DeclareUnicodeAccent{\~}                \UnicodeEncodingName{"0303}
-\DeclareUnicodeAccent{\=}                \UnicodeEncodingName{"0304}
-\DeclareUnicodeAccent{\u}                \UnicodeEncodingName{"0306}
-\DeclareUnicodeAccent{\.}                \UnicodeEncodingName{"0307}
-\DeclareUnicodeAccent{\"}                \UnicodeEncodingName{"0308}
-\DeclareUnicodeAccent{\r}                \UnicodeEncodingName{"030A}
-\DeclareUnicodeAccent{\H}                \UnicodeEncodingName{"030B}
-\DeclareUnicodeAccent{\v}                \UnicodeEncodingName{"030C}
-\DeclareUnicodeAccent{\b}                \UnicodeEncodingName{"0332}
-\DeclareUnicodeAccent{\d}                \UnicodeEncodingName{"0323}
-\DeclareUnicodeAccent{\c}                \UnicodeEncodingName{"0327}
-\DeclareUnicodeAccent{\k}                \UnicodeEncodingName{"0328}
+\DeclareUnicodeAccent{\`}{"0300}
+\DeclareUnicodeAccent{\'}{"0301}
+\DeclareUnicodeAccent{\^}{"0302}
+\DeclareUnicodeAccent{\~}{"0303}
+\DeclareUnicodeAccent{\=}{"0304}
+\DeclareUnicodeAccent{\u}{"0306}
+\DeclareUnicodeAccent{\.}{"0307}
+\DeclareUnicodeAccent{\"}{"0308}
+\DeclareUnicodeAccent{\r}{"030A}
+\DeclareUnicodeAccent{\H}{"030B}
+\DeclareUnicodeAccent{\v}{"030C}
+\DeclareUnicodeAccent{\b}{"0332}
+\DeclareUnicodeAccent{\d}{"0323}
+\DeclareUnicodeAccent{\c}{"0327}
+\DeclareUnicodeAccent{\k}{"0328}
+%    \end{macrocode}
+%    The odd one out:
+%    \begin{macrocode}
 \DeclareTextCommand\textcommabelow       \UnicodeEncodingName[1]
   {\hmode at bgroup\ooalign{\null#1\crcr\hidewidth\raise-.31ex
    \hbox{\check at mathfonts\fontsize\ssf at size\z@
diff --git a/base/testfiles-TU/github-0253.lvt b/base/testfiles-TU/github-0253.lvt
new file mode 100644
index 00000000..f09b24a7
--- /dev/null
+++ b/base/testfiles-TU/github-0253.lvt
@@ -0,0 +1,31 @@
+\documentclass{article}
+
+\pagestyle{empty}
+
+\input{test2e}
+
+\START
+
+\ShowCommand\=
+\expandafter\show
+\csname TU\string\=\endcsname
+
+\DeclareUnicodeAccent{\=}\UnicodeEncodingName{"0300}
+
+% composites change the internals, so we have to apply at least one again
+\DeclareUnicodeComposite{\=}             {A}{"0100}
+
+\ShowCommand\=
+\expandafter\show
+\csname TU\string\=\endcsname
+
+
+\DeclareUnicodeAccent{\=}{"0304}           % new syntax
+\relax          
+\DeclareUnicodeComposite{\=}             {A}{"0100}
+
+\ShowCommand\=
+\expandafter\show
+\csname TU\string\=\endcsname
+
+\END
diff --git a/base/testfiles-TU/github-0253.tlg b/base/testfiles-TU/github-0253.tlg
new file mode 100644
index 00000000..054fa5e5
--- /dev/null
+++ b/base/testfiles-TU/github-0253.tlg
@@ -0,0 +1,26 @@
+This is a generated file for the LaTeX2e validation system.
+Don't change this file in any respect.
+> \==macro:
+->\TU-cmd \=\TU\= .
+<argument> \=
+l. ...\ShowCommand\=
+> \TU\==macro:
+#1->\@text at composite \TU\= #1\@empty \@text at composite {\add at unicode@accent {"0304}{#1}}.
+<recently read> \TU\= 
+l. ...\csname TU\string\=\endcsname
+> \==macro:
+->\TU-cmd \=\TU\= .
+<argument> \=
+l. ...\ShowCommand\=
+> \TU\==macro:
+#1->\@text at composite \TU\= #1\@empty \@text at composite {\add at unicode@accent {"0300}{#1}}.
+<recently read> \TU\= 
+l. ...\csname TU\string\=\endcsname
+> \==macro:
+->\TU-cmd \=\TU\= .
+<argument> \=
+l. ...\ShowCommand\=
+> \TU\==macro:
+#1->\@text at composite \TU\= #1\@empty \@text at composite {\add at unicode@accent {"0304}{#1}}.
+<recently read> \TU\= 
+l. ...\csname TU\string\=\endcsname