texlive[64507] trunk: texfot (25sep22)
commits+karl at tug.org
commits+karl at tug.org
Sun Sep 25 22:14:15 CEST 2022
Revision: 64507
http://tug.org/svn/texlive?view=revision&revision=64507
Author: karl
Date: 2022-09-25 22:14:15 +0200 (Sun, 25 Sep 2022)
Log Message:
-----------
texfot (25sep22)
Modified Paths:
--------------
trunk/Build/source/texk/texlive/linked_scripts/texfot/texfot.pl
trunk/Master/texmf-dist/doc/man/man1/texfot.1
trunk/Master/texmf-dist/doc/man/man1/texfot.man1.pdf
trunk/Master/texmf-dist/scripts/texfot/texfot.pl
Modified: trunk/Build/source/texk/texlive/linked_scripts/texfot/texfot.pl
===================================================================
--- trunk/Build/source/texk/texlive/linked_scripts/texfot/texfot.pl 2022-09-25 20:13:55 UTC (rev 64506)
+++ trunk/Build/source/texk/texlive/linked_scripts/texfot/texfot.pl 2022-09-25 20:14:15 UTC (rev 64507)
@@ -1,5 +1,5 @@
#!/usr/bin/env perl
-# $Id: texfot,v 1.45 2022/02/17 18:42:02 karl Exp $
+# $Id: texfot,v 1.47 2022/09/24 16:38:13 karl Exp $
# Invoke a TeX command, filtering all but interesting terminal output;
# do not look at the log or check any output files.
# Exit status is that of the subprogram.
@@ -8,7 +8,7 @@
#
# Public domain. Originally written 2014 by Karl Berry.
-my $ident = '$Id: texfot,v 1.45 2022/02/17 18:42:02 karl Exp $';
+my $ident = '$Id: texfot,v 1.47 2022/09/24 16:38:13 karl Exp $';
(my $prg = $0) =~ s,^.*/,,;
select STDERR; $| = 1; # no buffering
select STDOUT; $| = 1;
@@ -21,6 +21,7 @@
use Getopt::Long qw(:config require_order);
use Pod::Usage;
+my @opt_accept = ();
my $opt_debug = 0;
my @opt_ignore = ();
my $opt_interactive = 0;
@@ -36,6 +37,7 @@
#
sub main {
my $ret = GetOptions (
+ "accept=s" => \@opt_accept,
"debug!" => \$opt_debug,
"ignore=s" => \@opt_ignore,
"interactive!" => \$opt_interactive,
@@ -90,9 +92,9 @@
# Be sure everything is drained.
&debug ("starting waitpid() for $pid") ;
- waitpid ($pid, 0) || die "$prog: waitpid($pid) failed: $!\n";
+ waitpid ($pid, 0) || die "$prg: waitpid($pid) failed: $!\n";
my $child_exit_status = $? >> 8;
- &debug ("child exit status = $exit_status\n");
+ &debug ("child exit status = $child_exit_status\n");
&debug ("processing stderr from child");
seek (TEXERR, 0, 0) || warn "seek(stderr) failed: $!";
@@ -129,6 +131,17 @@
next;
}
+ # don't anchor user accept patterns, leave it up to them.
+ for my $user_accept (@opt_accept) {
+ &debug ("checking user accept '$user_accept'\n");
+ if (/${user_accept}/) {
+ &debug (" found user accept ($user_accept)\n");
+ print $prefix;
+ print $line;
+ next;
+ }
+ }
+
&debug ("checking ignores\n");
next if /^(
LaTeX\ Warning:\ You\ have\ requested\ package
@@ -148,13 +161,13 @@
|!\ $
)/x;
- # don't anchor user ignores, leave it up to them.
+ # don't anchor user ignore patterns either.
for my $user_ignore (@opt_ignore) {
&debug ("checking user ignore '$user_ignore'\n");
next LINE if /${user_ignore}/;
}
- &debug ("checking for print_next\n");
+ &debug ("checking for print_next pattern\n");
if (/^(
.*?:[0-9]+: # usual file:lineno: form
|! # usual ! form
@@ -189,6 +202,7 @@
|all\ text\ was\ ignored\ after\ line
|.*Fatal\ error
|.*for\ symbol.*on\ input\ line
+ |\#\#
)/x) {
&debug (" matched for showing ($1)\n");
print $prefix;
@@ -225,8 +239,8 @@
=head1 DESCRIPTION
C<texfot> invokes I<texcmd> with the given I<texarg> arguments,
-filtering the online output for ``interesting'' messages. Its exit
-value is that of I<texcmd>. Examples:
+filtering the online output for ``interesting'' messages. Its exit
+value is that of I<texcmd>. Examples:
# Sample basic invocation:
texfot pdflatex file.tex
@@ -238,15 +252,22 @@
# Example of more complex engine invocation:
texfot xelatex --recorder '\nonstopmode\input file'
+Here is an example of what the output looks like (in its entirety) on an
+error-free run:
+
+ /path/to/texfot: invoking: pdflatex hello.ltx
+ This is pdfTeX, Version 3.141592653-2.6-1.40.24 (TeX Live 2022) (preloaded format=pdflatex)
+ Output written on hello.pdf (1 page, 94415 bytes).
+
Aside from its own options, described below, C<texfot> just runs the
given command with the given arguments (same approach to command line
-syntax as C<env>, C<nice>, C<time>, C<timeout>, etc.). Thus, C<texfot>
-works with any engine and any command line options.
+syntax as C<env>, C<nice>, C<timeout>, etc.). Thus, C<texfot> works with
+any engine and any command line options.
C<texfot> does not look at the log file or any other possible output
file(s); it only looks at the standard output and standard error from
-the command. stdout is processed first, then stderr. Lines from stderr
-have an identifying prefix. C<texfot> writes all accepted lines to its
+the command. stdout is processed first, then stderr. Lines from stderr
+have an identifying prefix. C<texfot> writes all accepted lines to its
stdout.
The messages shown are intended to be those which likely need action by
@@ -265,32 +286,34 @@
=item 2.
+Otherwise, if the line matches any user-supplied list of regexps to
+accept (given with C<--accept>, see below), in that order, print it.
+
+=item 3.
+
Otherwise, if the line matches the built-in list of regexps to ignore,
or any user-supplied list of regexps to ignore (given with C<--ignore>,
see below), in that order, ignore it.
-=item 3.
+=item 4.
Otherwise, if the line matches the list of regexps for which the next
line (two lines in all) should be shown, show this line and set the
-``next line'' flag for the next time around the loop. Examples are the
+``next line'' flag for the next time around the loop. Examples are the
common C<!> and C<filename:lineno:> error messages, which are generally
followed by a line with specific detail about the error.
-=item 4.
+=item 5.
Otherwise, if the line matches the list of regexps to show, show it.
-=item 5.
+=item 6.
Otherwise, the default: if the line came from stdout, ignore it; if the
-line came from stderr, print it (to stdout). This distinction is made
+line came from stderr, print it (to stdout). This distinction is made
because TeX engines write relatively few messages to stderr, and it's
likely that any such should be considered.
-It would be easy to add more options to allow for user additions to the
-various regex lists, if that ever seems useful. Or email me (see end).
-
=back
Once a particular check matches, the program moves on to process the
@@ -303,8 +326,8 @@
Incidentally, although nothing in this basic operation is specific to
TeX engines, all the regular expressions included in the program are
-specific to TeX. So in practice the program isn't useful except with
-TeX engines, although it would be easy enough to adapt it (if there was
+specific to TeX. So in practice the program isn't useful except with TeX
+engines, although it would be easy enough to adapt it (if there was
anything else as verbose as TeX to make that useful).
=head1 OPTIONS
@@ -315,28 +338,35 @@
The first non-option terminates C<texfot>'s option parsing, and the
remainder of the command line is invoked as the TeX command, without
-further parsing. For example, C<texfot --debug tex
+further parsing. For example, C<texfot --debug tex
--debug> will output debugging information from both C<texfot> and
C<tex>.
Options may start with either - or --, and may be unambiguously
-abbreviated. It is best to use the full option name in scripts, though,
+abbreviated. It is best to use the full option name in scripts, though,
to avoid possible collisions with new options in the future.
=over 4
+=item C<--accept> I<regexp>
+
+Accept lines in the TeX output matching (Perl) I<regexp>. Can be
+repeated. This list is checked first, so any and all matches will be
+shown, regardless of other options. These regexps are not automatically
+anchored (or otherwise altered), simply used as-is.
+
=item C<--debug>
=item C<--no-debug>
-Output (or not) what is being done on standard error. Off by default.
+Output (or not) what the program is doing to standard error; off by default.
=item C<--ignore> I<regexp>
-Ignore lines in the TeX output matching (Perl) I<regexp>. Can be
-repeated. Adds to the default set of ignore regexps rather than
-replacing. These regexps are not automatically anchored (or otherwise
-altered), simply used as-is.
+Ignore lines in the TeX output matching (Perl) I<regexp>. Can be
+repeated. Adds to the default set of ignore regexps rather than
+replacing. Like the acceptance regexps, these are not automatically
+anchored (or otherwise altered).
=item C<--interactive>
@@ -344,14 +374,14 @@
By default, standard input to the TeX process is closed so that TeX's
interactive mode (waiting for input upon error, the C<*> prompt, etc.)
-is never entered. Giving C<--interactive> allows interaction to happen.
+is never entered. Giving C<--interactive> allows interaction to happen.
=item C<--quiet>
=item C<--no-quiet>
-By default, the TeX command being invoked is reported on standard output.
-C<--quiet> omits that reporting. To get a completely silent run,
+By default, the TeX command being invoked is reported on standard
+output. C<--quiet> omits that reporting. To get a completely silent run,
redirect standard output: S<C<texfot ... E<gt>/dev/null>>. (The only
messages to standard error should be errors from C<texfot> itself, so it
shouldn't be necessary to redirect that, but of course that can be done
@@ -362,13 +392,13 @@
=item C<--no-stderr>
The default is for C<texfot> to report everything written to stderr by
-the TeX command (on stdout). C<--no-stderr> omits that reporting.
-(Some programs, C<dvisvgm> is one, can be rather verbose on stderr.)
+the TeX command (on stdout). C<--no-stderr> omits that reporting. (Some
+programs, C<dvisvgm> is one, can be rather verbose on stderr.)
=item C<--tee> I<file>
By default, the output being filtered is C<tee>-ed, before filtering, to
-make it easy to check the full output in case of problems.
+make it easy to check the full output in case of problems.
The default I<file> is C<$TMPDIR/fot.>I<uid>; if C<TMPDIR> is not set,
C<TMP> is used if set; if neither is set, the default directory is
@@ -393,19 +423,18 @@
I wrote this because, in my work as a TUGboat editor
(L<https://tug.org/TUGboat>, journal submissions always welcome!), I run
-and rerun many documents, many times each. It was easy to lose
-warnings I needed to see in the mass of unvarying and uninteresting
-output from TeX, such as style files being read and fonts being used. I
-wanted to see all and only those messages which needed some action by
-me.
+and rerun many documents, many times each. It was easy to lose warnings
+I needed to see in the mass of unvarying and uninteresting output from
+TeX, such as style files being read and fonts being used. I wanted to
+see all and only those messages which needed some action by me.
I found some other programs of a similar nature, the LaTeX package
C<silence>, and plenty of other (La)TeX wrappers, but it seemed none of
-them did what I wanted. Either they read the log file (I wanted the
-online output only), or they output more or less than I wanted, or they
-required invoking TeX differently (I wanted to keep my build process
-exactly the same, most critically the TeX invocation, which can get
-complicated). Hence I wrote this.
+them did what I wanted. Either they read the log file (I wanted to look
+at only the online output), or they output more or less than I wanted,
+or they required invoking TeX differently (I wanted to keep my build
+process exactly the same, most critically the TeX invocation, which can
+get complicated). Hence I wrote this little script.
Here are some keywords if you want to explore other options:
texloganalyser, pydflatex, logfilter, latexmk, rubber, arara, and
@@ -417,17 +446,17 @@
The name comes from the C<trip.fot> and C<trap.fot> files that are part
of Knuth's trip and trap torture tests, which record the online output
-from the programs. I am not sure what "fot" stands for in trip and
-trap, but I can pretend that it stands for "filter online transcript" in
-the present S<case :).>
+from the programs. I am not sure what "fot" stands for in trip and trap,
+but I can pretend that it stands for "filter online transcript" in the
+present S<case :).>
=head1 AUTHORS AND COPYRIGHT
This script and its documentation were written by Karl Berry and both
-are released to the public domain. Email C<karl at freefriends.org> with
-bug reports. It has no home page beyond the package on CTAN:
+are released to the public domain. Email C<karl at freefriends.org> with
+bug reports. It has no home page beyond the package page on CTAN:
L<https://ctan.org/pkg/texfot>.
- $Id: texfot,v 1.45 2022/02/17 18:42:02 karl Exp $
+ $Id: texfot,v 1.47 2022/09/24 16:38:13 karl Exp $
=cut
Modified: trunk/Master/texmf-dist/doc/man/man1/texfot.1
===================================================================
--- trunk/Master/texmf-dist/doc/man/man1/texfot.1 2022-09-25 20:13:55 UTC (rev 64506)
+++ trunk/Master/texmf-dist/doc/man/man1/texfot.1 2022-09-25 20:14:15 UTC (rev 64507)
@@ -133,7 +133,7 @@
.\" ========================================================================
.\"
.IX Title "TEXFOT 1"
-.TH TEXFOT 1 "2022-02-17" "texfot" "Karl Berry"
+.TH TEXFOT 1 "2022-09-24" "texfot" "Karl Berry"
.\" For nroff, turn off justification. Always turn off hyphenation; it makes
.\" way too many mistakes in technical documents.
.if n .ad l
@@ -146,8 +146,8 @@
.SH "DESCRIPTION"
.IX Header "DESCRIPTION"
\&\f(CW\*(C`texfot\*(C'\fR invokes \fItexcmd\fR with the given \fItexarg\fR arguments,
-filtering the online output for ``interesting'' messages. Its exit
-value is that of \fItexcmd\fR. Examples:
+filtering the online output for ``interesting'' messages. Its exit
+value is that of \fItexcmd\fR. Examples:
.PP
.Vb 2
\& # Sample basic invocation:
@@ -161,15 +161,24 @@
\& texfot xelatex \-\-recorder \*(Aq\enonstopmode\einput file\*(Aq
.Ve
.PP
+Here is an example of what the output looks like (in its entirety) on an
+error-free run:
+.PP
+.Vb 3
+\& /path/to/texfot: invoking: pdflatex hello.ltx
+\& This is pdfTeX, Version 3.141592653\-2.6\-1.40.24 (TeX Live 2022) (preloaded format=pdflatex)
+\& Output written on hello.pdf (1 page, 94415 bytes).
+.Ve
+.PP
Aside from its own options, described below, \f(CW\*(C`texfot\*(C'\fR just runs the
given command with the given arguments (same approach to command line
-syntax as \f(CW\*(C`env\*(C'\fR, \f(CW\*(C`nice\*(C'\fR, \f(CW\*(C`time\*(C'\fR, \f(CW\*(C`timeout\*(C'\fR, etc.). Thus, \f(CW\*(C`texfot\*(C'\fR
-works with any engine and any command line options.
+syntax as \f(CW\*(C`env\*(C'\fR, \f(CW\*(C`nice\*(C'\fR, \f(CW\*(C`timeout\*(C'\fR, etc.). Thus, \f(CW\*(C`texfot\*(C'\fR works with
+any engine and any command line options.
.PP
\&\f(CW\*(C`texfot\*(C'\fR does not look at the log file or any other possible output
file(s); it only looks at the standard output and standard error from
-the command. stdout is processed first, then stderr. Lines from stderr
-have an identifying prefix. \f(CW\*(C`texfot\*(C'\fR writes all accepted lines to its
+the command. stdout is processed first, then stderr. Lines from stderr
+have an identifying prefix. \f(CW\*(C`texfot\*(C'\fR writes all accepted lines to its
stdout.
.PP
The messages shown are intended to be those which likely need action by
@@ -181,25 +190,25 @@
.IP "1." 4
If the ``next line'' needs to be printed (see below), print it.
.IP "2." 4
+Otherwise, if the line matches any user-supplied list of regexps to
+accept (given with \f(CW\*(C`\-\-accept\*(C'\fR, see below), in that order, print it.
+.IP "3." 4
Otherwise, if the line matches the built-in list of regexps to ignore,
or any user-supplied list of regexps to ignore (given with \f(CW\*(C`\-\-ignore\*(C'\fR,
see below), in that order, ignore it.
-.IP "3." 4
+.IP "4." 4
Otherwise, if the line matches the list of regexps for which the next
line (two lines in all) should be shown, show this line and set the
-``next line'' flag for the next time around the loop. Examples are the
+``next line'' flag for the next time around the loop. Examples are the
common \f(CW\*(C`!\*(C'\fR and \f(CW\*(C`filename:lineno:\*(C'\fR error messages, which are generally
followed by a line with specific detail about the error.
-.IP "4." 4
+.IP "5." 4
Otherwise, if the line matches the list of regexps to show, show it.
-.IP "5." 4
+.IP "6." 4
Otherwise, the default: if the line came from stdout, ignore it; if the
-line came from stderr, print it (to stdout). This distinction is made
+line came from stderr, print it (to stdout). This distinction is made
because TeX engines write relatively few messages to stderr, and it's
likely that any such should be considered.
-.Sp
-It would be easy to add more options to allow for user additions to the
-various regex lists, if that ever seems useful. Or email me (see end).
.PP
Once a particular check matches, the program moves on to process the
next line.
@@ -211,8 +220,8 @@
.PP
Incidentally, although nothing in this basic operation is specific to
TeX engines, all the regular expressions included in the program are
-specific to TeX. So in practice the program isn't useful except with
-TeX engines, although it would be easy enough to adapt it (if there was
+specific to TeX. So in practice the program isn't useful except with TeX
+engines, although it would be easy enough to adapt it (if there was
anything else as verbose as TeX to make that useful).
.SH "OPTIONS"
.IX Header "OPTIONS"
@@ -222,13 +231,20 @@
.PP
The first non-option terminates \f(CW\*(C`texfot\*(C'\fR's option parsing, and the
remainder of the command line is invoked as the TeX command, without
-further parsing. For example, \f(CW\*(C`texfot \-\-debug tex
+further parsing. For example, \f(CW\*(C`texfot \-\-debug tex
\&\-\-debug\*(C'\fR will output debugging information from both \f(CW\*(C`texfot\*(C'\fR and
\&\f(CW\*(C`tex\*(C'\fR.
.PP
Options may start with either \- or \-\-, and may be unambiguously
-abbreviated. It is best to use the full option name in scripts, though,
+abbreviated. It is best to use the full option name in scripts, though,
to avoid possible collisions with new options in the future.
+.ie n .IP """\-\-accept"" \fIregexp\fR" 4
+.el .IP "\f(CW\-\-accept\fR \fIregexp\fR" 4
+.IX Item "--accept regexp"
+Accept lines in the TeX output matching (Perl) \fIregexp\fR. Can be
+repeated. This list is checked first, so any and all matches will be
+shown, regardless of other options. These regexps are not automatically
+anchored (or otherwise altered), simply used as-is.
.ie n .IP """\-\-debug""" 4
.el .IP "\f(CW\-\-debug\fR" 4
.IX Item "--debug"
@@ -237,14 +253,14 @@
.el .IP "\f(CW\-\-no\-debug\fR" 4
.IX Item "--no-debug"
.PD
-Output (or not) what is being done on standard error. Off by default.
+Output (or not) what the program is doing to standard error; off by default.
.ie n .IP """\-\-ignore"" \fIregexp\fR" 4
.el .IP "\f(CW\-\-ignore\fR \fIregexp\fR" 4
.IX Item "--ignore regexp"
-Ignore lines in the TeX output matching (Perl) \fIregexp\fR. Can be
-repeated. Adds to the default set of ignore regexps rather than
-replacing. These regexps are not automatically anchored (or otherwise
-altered), simply used as-is.
+Ignore lines in the TeX output matching (Perl) \fIregexp\fR. Can be
+repeated. Adds to the default set of ignore regexps rather than
+replacing. Like the acceptance regexps, these are not automatically
+anchored (or otherwise altered).
.ie n .IP """\-\-interactive""" 4
.el .IP "\f(CW\-\-interactive\fR" 4
.IX Item "--interactive"
@@ -255,7 +271,7 @@
.PD
By default, standard input to the TeX process is closed so that TeX's
interactive mode (waiting for input upon error, the \f(CW\*(C`*\*(C'\fR prompt, etc.)
-is never entered. Giving \f(CW\*(C`\-\-interactive\*(C'\fR allows interaction to happen.
+is never entered. Giving \f(CW\*(C`\-\-interactive\*(C'\fR allows interaction to happen.
.ie n .IP """\-\-quiet""" 4
.el .IP "\f(CW\-\-quiet\fR" 4
.IX Item "--quiet"
@@ -264,8 +280,8 @@
.el .IP "\f(CW\-\-no\-quiet\fR" 4
.IX Item "--no-quiet"
.PD
-By default, the TeX command being invoked is reported on standard output.
-\&\f(CW\*(C`\-\-quiet\*(C'\fR omits that reporting. To get a completely silent run,
+By default, the TeX command being invoked is reported on standard
+output. \f(CW\*(C`\-\-quiet\*(C'\fR omits that reporting. To get a completely silent run,
redirect standard output: \f(CW\*(C`texfot\ ...\ >/dev/null\*(C'\fR. (The only
messages to standard error should be errors from \f(CW\*(C`texfot\*(C'\fR itself, so it
shouldn't be necessary to redirect that, but of course that can be done
@@ -279,8 +295,8 @@
.IX Item "--no-stderr"
.PD
The default is for \f(CW\*(C`texfot\*(C'\fR to report everything written to stderr by
-the TeX command (on stdout). \f(CW\*(C`\-\-no\-stderr\*(C'\fR omits that reporting.
-(Some programs, \f(CW\*(C`dvisvgm\*(C'\fR is one, can be rather verbose on stderr.)
+the TeX command (on stdout). \f(CW\*(C`\-\-no\-stderr\*(C'\fR omits that reporting. (Some
+programs, \f(CW\*(C`dvisvgm\*(C'\fR is one, can be rather verbose on stderr.)
.ie n .IP """\-\-tee"" \fIfile\fR" 4
.el .IP "\f(CW\-\-tee\fR \fIfile\fR" 4
.IX Item "--tee file"
@@ -306,19 +322,18 @@
.IX Header "RATIONALE"
I wrote this because, in my work as a TUGboat editor
(<https://tug.org/TUGboat>, journal submissions always welcome!), I run
-and rerun many documents, many times each. It was easy to lose
-warnings I needed to see in the mass of unvarying and uninteresting
-output from TeX, such as style files being read and fonts being used. I
-wanted to see all and only those messages which needed some action by
-me.
+and rerun many documents, many times each. It was easy to lose warnings
+I needed to see in the mass of unvarying and uninteresting output from
+TeX, such as style files being read and fonts being used. I wanted to
+see all and only those messages which needed some action by me.
.PP
I found some other programs of a similar nature, the LaTeX package
\&\f(CW\*(C`silence\*(C'\fR, and plenty of other (La)TeX wrappers, but it seemed none of
-them did what I wanted. Either they read the log file (I wanted the
-online output only), or they output more or less than I wanted, or they
-required invoking TeX differently (I wanted to keep my build process
-exactly the same, most critically the TeX invocation, which can get
-complicated). Hence I wrote this.
+them did what I wanted. Either they read the log file (I wanted to look
+at only the online output), or they output more or less than I wanted,
+or they required invoking TeX differently (I wanted to keep my build
+process exactly the same, most critically the TeX invocation, which can
+get complicated). Hence I wrote this little script.
.PP
Here are some keywords if you want to explore other options:
texloganalyser, pydflatex, logfilter, latexmk, rubber, arara, and
@@ -330,16 +345,16 @@
.PP
The name comes from the \f(CW\*(C`trip.fot\*(C'\fR and \f(CW\*(C`trap.fot\*(C'\fR files that are part
of Knuth's trip and trap torture tests, which record the online output
-from the programs. I am not sure what \*(L"fot\*(R" stands for in trip and
-trap, but I can pretend that it stands for \*(L"filter online transcript\*(R" in
-the present case\ :).
+from the programs. I am not sure what \*(L"fot\*(R" stands for in trip and trap,
+but I can pretend that it stands for \*(L"filter online transcript\*(R" in the
+present case\ :).
.SH "AUTHORS AND COPYRIGHT"
.IX Header "AUTHORS AND COPYRIGHT"
This script and its documentation were written by Karl Berry and both
-are released to the public domain. Email \f(CW\*(C`karl at freefriends.org\*(C'\fR with
-bug reports. It has no home page beyond the package on \s-1CTAN:\s0
+are released to the public domain. Email \f(CW\*(C`karl at freefriends.org\*(C'\fR with
+bug reports. It has no home page beyond the package page on \s-1CTAN:\s0
<https://ctan.org/pkg/texfot>.
.PP
.Vb 1
-\& $Id: texfot,v 1.45 2022/02/17 18:42:02 karl Exp $
+\& $Id: texfot,v 1.47 2022/09/24 16:38:13 karl Exp $
.Ve
Modified: trunk/Master/texmf-dist/doc/man/man1/texfot.man1.pdf
===================================================================
(Binary files differ)
Modified: trunk/Master/texmf-dist/scripts/texfot/texfot.pl
===================================================================
--- trunk/Master/texmf-dist/scripts/texfot/texfot.pl 2022-09-25 20:13:55 UTC (rev 64506)
+++ trunk/Master/texmf-dist/scripts/texfot/texfot.pl 2022-09-25 20:14:15 UTC (rev 64507)
@@ -1,5 +1,5 @@
#!/usr/bin/env perl
-# $Id: texfot,v 1.45 2022/02/17 18:42:02 karl Exp $
+# $Id: texfot,v 1.47 2022/09/24 16:38:13 karl Exp $
# Invoke a TeX command, filtering all but interesting terminal output;
# do not look at the log or check any output files.
# Exit status is that of the subprogram.
@@ -8,7 +8,7 @@
#
# Public domain. Originally written 2014 by Karl Berry.
-my $ident = '$Id: texfot,v 1.45 2022/02/17 18:42:02 karl Exp $';
+my $ident = '$Id: texfot,v 1.47 2022/09/24 16:38:13 karl Exp $';
(my $prg = $0) =~ s,^.*/,,;
select STDERR; $| = 1; # no buffering
select STDOUT; $| = 1;
@@ -21,6 +21,7 @@
use Getopt::Long qw(:config require_order);
use Pod::Usage;
+my @opt_accept = ();
my $opt_debug = 0;
my @opt_ignore = ();
my $opt_interactive = 0;
@@ -36,6 +37,7 @@
#
sub main {
my $ret = GetOptions (
+ "accept=s" => \@opt_accept,
"debug!" => \$opt_debug,
"ignore=s" => \@opt_ignore,
"interactive!" => \$opt_interactive,
@@ -90,9 +92,9 @@
# Be sure everything is drained.
&debug ("starting waitpid() for $pid") ;
- waitpid ($pid, 0) || die "$prog: waitpid($pid) failed: $!\n";
+ waitpid ($pid, 0) || die "$prg: waitpid($pid) failed: $!\n";
my $child_exit_status = $? >> 8;
- &debug ("child exit status = $exit_status\n");
+ &debug ("child exit status = $child_exit_status\n");
&debug ("processing stderr from child");
seek (TEXERR, 0, 0) || warn "seek(stderr) failed: $!";
@@ -129,6 +131,17 @@
next;
}
+ # don't anchor user accept patterns, leave it up to them.
+ for my $user_accept (@opt_accept) {
+ &debug ("checking user accept '$user_accept'\n");
+ if (/${user_accept}/) {
+ &debug (" found user accept ($user_accept)\n");
+ print $prefix;
+ print $line;
+ next;
+ }
+ }
+
&debug ("checking ignores\n");
next if /^(
LaTeX\ Warning:\ You\ have\ requested\ package
@@ -148,13 +161,13 @@
|!\ $
)/x;
- # don't anchor user ignores, leave it up to them.
+ # don't anchor user ignore patterns either.
for my $user_ignore (@opt_ignore) {
&debug ("checking user ignore '$user_ignore'\n");
next LINE if /${user_ignore}/;
}
- &debug ("checking for print_next\n");
+ &debug ("checking for print_next pattern\n");
if (/^(
.*?:[0-9]+: # usual file:lineno: form
|! # usual ! form
@@ -189,6 +202,7 @@
|all\ text\ was\ ignored\ after\ line
|.*Fatal\ error
|.*for\ symbol.*on\ input\ line
+ |\#\#
)/x) {
&debug (" matched for showing ($1)\n");
print $prefix;
@@ -225,8 +239,8 @@
=head1 DESCRIPTION
C<texfot> invokes I<texcmd> with the given I<texarg> arguments,
-filtering the online output for ``interesting'' messages. Its exit
-value is that of I<texcmd>. Examples:
+filtering the online output for ``interesting'' messages. Its exit
+value is that of I<texcmd>. Examples:
# Sample basic invocation:
texfot pdflatex file.tex
@@ -238,15 +252,22 @@
# Example of more complex engine invocation:
texfot xelatex --recorder '\nonstopmode\input file'
+Here is an example of what the output looks like (in its entirety) on an
+error-free run:
+
+ /path/to/texfot: invoking: pdflatex hello.ltx
+ This is pdfTeX, Version 3.141592653-2.6-1.40.24 (TeX Live 2022) (preloaded format=pdflatex)
+ Output written on hello.pdf (1 page, 94415 bytes).
+
Aside from its own options, described below, C<texfot> just runs the
given command with the given arguments (same approach to command line
-syntax as C<env>, C<nice>, C<time>, C<timeout>, etc.). Thus, C<texfot>
-works with any engine and any command line options.
+syntax as C<env>, C<nice>, C<timeout>, etc.). Thus, C<texfot> works with
+any engine and any command line options.
C<texfot> does not look at the log file or any other possible output
file(s); it only looks at the standard output and standard error from
-the command. stdout is processed first, then stderr. Lines from stderr
-have an identifying prefix. C<texfot> writes all accepted lines to its
+the command. stdout is processed first, then stderr. Lines from stderr
+have an identifying prefix. C<texfot> writes all accepted lines to its
stdout.
The messages shown are intended to be those which likely need action by
@@ -265,32 +286,34 @@
=item 2.
+Otherwise, if the line matches any user-supplied list of regexps to
+accept (given with C<--accept>, see below), in that order, print it.
+
+=item 3.
+
Otherwise, if the line matches the built-in list of regexps to ignore,
or any user-supplied list of regexps to ignore (given with C<--ignore>,
see below), in that order, ignore it.
-=item 3.
+=item 4.
Otherwise, if the line matches the list of regexps for which the next
line (two lines in all) should be shown, show this line and set the
-``next line'' flag for the next time around the loop. Examples are the
+``next line'' flag for the next time around the loop. Examples are the
common C<!> and C<filename:lineno:> error messages, which are generally
followed by a line with specific detail about the error.
-=item 4.
+=item 5.
Otherwise, if the line matches the list of regexps to show, show it.
-=item 5.
+=item 6.
Otherwise, the default: if the line came from stdout, ignore it; if the
-line came from stderr, print it (to stdout). This distinction is made
+line came from stderr, print it (to stdout). This distinction is made
because TeX engines write relatively few messages to stderr, and it's
likely that any such should be considered.
-It would be easy to add more options to allow for user additions to the
-various regex lists, if that ever seems useful. Or email me (see end).
-
=back
Once a particular check matches, the program moves on to process the
@@ -303,8 +326,8 @@
Incidentally, although nothing in this basic operation is specific to
TeX engines, all the regular expressions included in the program are
-specific to TeX. So in practice the program isn't useful except with
-TeX engines, although it would be easy enough to adapt it (if there was
+specific to TeX. So in practice the program isn't useful except with TeX
+engines, although it would be easy enough to adapt it (if there was
anything else as verbose as TeX to make that useful).
=head1 OPTIONS
@@ -315,28 +338,35 @@
The first non-option terminates C<texfot>'s option parsing, and the
remainder of the command line is invoked as the TeX command, without
-further parsing. For example, C<texfot --debug tex
+further parsing. For example, C<texfot --debug tex
--debug> will output debugging information from both C<texfot> and
C<tex>.
Options may start with either - or --, and may be unambiguously
-abbreviated. It is best to use the full option name in scripts, though,
+abbreviated. It is best to use the full option name in scripts, though,
to avoid possible collisions with new options in the future.
=over 4
+=item C<--accept> I<regexp>
+
+Accept lines in the TeX output matching (Perl) I<regexp>. Can be
+repeated. This list is checked first, so any and all matches will be
+shown, regardless of other options. These regexps are not automatically
+anchored (or otherwise altered), simply used as-is.
+
=item C<--debug>
=item C<--no-debug>
-Output (or not) what is being done on standard error. Off by default.
+Output (or not) what the program is doing to standard error; off by default.
=item C<--ignore> I<regexp>
-Ignore lines in the TeX output matching (Perl) I<regexp>. Can be
-repeated. Adds to the default set of ignore regexps rather than
-replacing. These regexps are not automatically anchored (or otherwise
-altered), simply used as-is.
+Ignore lines in the TeX output matching (Perl) I<regexp>. Can be
+repeated. Adds to the default set of ignore regexps rather than
+replacing. Like the acceptance regexps, these are not automatically
+anchored (or otherwise altered).
=item C<--interactive>
@@ -344,14 +374,14 @@
By default, standard input to the TeX process is closed so that TeX's
interactive mode (waiting for input upon error, the C<*> prompt, etc.)
-is never entered. Giving C<--interactive> allows interaction to happen.
+is never entered. Giving C<--interactive> allows interaction to happen.
=item C<--quiet>
=item C<--no-quiet>
-By default, the TeX command being invoked is reported on standard output.
-C<--quiet> omits that reporting. To get a completely silent run,
+By default, the TeX command being invoked is reported on standard
+output. C<--quiet> omits that reporting. To get a completely silent run,
redirect standard output: S<C<texfot ... E<gt>/dev/null>>. (The only
messages to standard error should be errors from C<texfot> itself, so it
shouldn't be necessary to redirect that, but of course that can be done
@@ -362,13 +392,13 @@
=item C<--no-stderr>
The default is for C<texfot> to report everything written to stderr by
-the TeX command (on stdout). C<--no-stderr> omits that reporting.
-(Some programs, C<dvisvgm> is one, can be rather verbose on stderr.)
+the TeX command (on stdout). C<--no-stderr> omits that reporting. (Some
+programs, C<dvisvgm> is one, can be rather verbose on stderr.)
=item C<--tee> I<file>
By default, the output being filtered is C<tee>-ed, before filtering, to
-make it easy to check the full output in case of problems.
+make it easy to check the full output in case of problems.
The default I<file> is C<$TMPDIR/fot.>I<uid>; if C<TMPDIR> is not set,
C<TMP> is used if set; if neither is set, the default directory is
@@ -393,19 +423,18 @@
I wrote this because, in my work as a TUGboat editor
(L<https://tug.org/TUGboat>, journal submissions always welcome!), I run
-and rerun many documents, many times each. It was easy to lose
-warnings I needed to see in the mass of unvarying and uninteresting
-output from TeX, such as style files being read and fonts being used. I
-wanted to see all and only those messages which needed some action by
-me.
+and rerun many documents, many times each. It was easy to lose warnings
+I needed to see in the mass of unvarying and uninteresting output from
+TeX, such as style files being read and fonts being used. I wanted to
+see all and only those messages which needed some action by me.
I found some other programs of a similar nature, the LaTeX package
C<silence>, and plenty of other (La)TeX wrappers, but it seemed none of
-them did what I wanted. Either they read the log file (I wanted the
-online output only), or they output more or less than I wanted, or they
-required invoking TeX differently (I wanted to keep my build process
-exactly the same, most critically the TeX invocation, which can get
-complicated). Hence I wrote this.
+them did what I wanted. Either they read the log file (I wanted to look
+at only the online output), or they output more or less than I wanted,
+or they required invoking TeX differently (I wanted to keep my build
+process exactly the same, most critically the TeX invocation, which can
+get complicated). Hence I wrote this little script.
Here are some keywords if you want to explore other options:
texloganalyser, pydflatex, logfilter, latexmk, rubber, arara, and
@@ -417,17 +446,17 @@
The name comes from the C<trip.fot> and C<trap.fot> files that are part
of Knuth's trip and trap torture tests, which record the online output
-from the programs. I am not sure what "fot" stands for in trip and
-trap, but I can pretend that it stands for "filter online transcript" in
-the present S<case :).>
+from the programs. I am not sure what "fot" stands for in trip and trap,
+but I can pretend that it stands for "filter online transcript" in the
+present S<case :).>
=head1 AUTHORS AND COPYRIGHT
This script and its documentation were written by Karl Berry and both
-are released to the public domain. Email C<karl at freefriends.org> with
-bug reports. It has no home page beyond the package on CTAN:
+are released to the public domain. Email C<karl at freefriends.org> with
+bug reports. It has no home page beyond the package page on CTAN:
L<https://ctan.org/pkg/texfot>.
- $Id: texfot,v 1.45 2022/02/17 18:42:02 karl Exp $
+ $Id: texfot,v 1.47 2022/09/24 16:38:13 karl Exp $
=cut
More information about the tex-live-commits
mailing list.