From c88a7b89c9b0fb65a94c28f07ba277decc224f74 Mon Sep 17 00:00:00 2001 From: Karl Berry Date: Thu, 28 May 2020 22:03:10 +0000 Subject: latexdiff (28may20) git-svn-id: svn://tug.org/texlive/trunk@55328 c570f23f-e606-0410-a88d-b1316a301751 --- .../linked_scripts/latexdiff/latexdiff-vc.pl | 23 +- .../texlive/linked_scripts/latexdiff/latexdiff.pl | 320 +++++++++++++++------ .../linked_scripts/latexdiff/latexrevise.pl | 58 ++-- .../texk/texlive/linked_scripts/texlive/fmtutil.pl | 14 +- Master/texmf-dist/doc/man/man1/latexdiff-vc.1 | 20 +- .../texmf-dist/doc/man/man1/latexdiff-vc.man1.pdf | Bin 34898 -> 34060 bytes Master/texmf-dist/doc/man/man1/latexdiff.1 | 20 +- Master/texmf-dist/doc/man/man1/latexdiff.man1.pdf | Bin 72323 -> 68379 bytes Master/texmf-dist/doc/man/man1/latexrevise.1 | 9 +- .../texmf-dist/doc/man/man1/latexrevise.man1.pdf | Bin 29948 -> 29354 bytes Master/texmf-dist/doc/support/latexdiff/Makefile | 12 +- Master/texmf-dist/doc/support/latexdiff/README | 6 +- .../doc/support/latexdiff/contrib/latexdiff-wrap | 4 +- .../doc/support/latexdiff/doc/latexdiff-man.pdf | Bin 274601 -> 282069 bytes .../doc/support/latexdiff/doc/latexdiff-man.tex | 6 +- Master/texmf-dist/doc/support/latexdiff/latexdiff | 318 ++++++++++++++------ .../doc/support/latexdiff/latexdiff-fast | 320 +++++++++++++++------ .../texmf-dist/scripts/latexdiff/latexdiff-vc.pl | 23 +- Master/texmf-dist/scripts/latexdiff/latexdiff.pl | 320 +++++++++++++++------ Master/texmf-dist/scripts/latexdiff/latexrevise.pl | 58 ++-- 20 files changed, 1099 insertions(+), 432 deletions(-) diff --git a/Build/source/texk/texlive/linked_scripts/latexdiff/latexdiff-vc.pl b/Build/source/texk/texlive/linked_scripts/latexdiff/latexdiff-vc.pl index 21bf1f3adf3..9843238c457 100755 --- a/Build/source/texk/texlive/linked_scripts/latexdiff/latexdiff-vc.pl +++ b/Build/source/texk/texlive/linked_scripts/latexdiff/latexdiff-vc.pl @@ -26,18 +26,17 @@ # Detailed usage information at the end of the file # # TODO/IDEAS: - option to call external pre-processing codes -# - choose type of latex processor / bibtex (luatex, xelatex etc) # version 1.3.0 ((7 October 2018) # - option --only-changes with hyperref will suppress hyperrefs (pull request jprotze)_ # - option --only-changes now moves (rather than copies) file with only changes # version 1.2.1 (22 June 2017): -# - update: use qpdf (instead of pdftk) to select particular pages of output (pull request #102 submited by Tom Scogland via github). This was necessary because pdftk is being deprecated. pdftk is still used as a fall-back +# - update: use qpdf (instead of pdftk) to select particular pages of output (pull request #102 submitted by Tom Scogland via github). This was necessary because pdftk is being deprecated. pdftk is still used as a fall-back # - bug fix: --hg option was not recognised (partially fixes github issue #93 ) # - wrap passed-through options to latexdiff in quotation marks (fix github issue #58 ) # - program names for latexdiff, latex, dvips, bibtex configurable (fixes issue #40) # # version 1.2.0: -# - depracation fix: left brace in RegEx now needs to be escaped +# - deprecation fix: left brace in RegEx now needs to be escaped # # version 1.1.1: # - better detection of RCS system @@ -72,8 +71,8 @@ use strict ; use warnings ; my $versionstring=< -r F ... -compares F with the most recent version checked into RCS. +compares F with the most recently checked-in version checked. =item B -r F F ... @@ -685,7 +684,7 @@ compares F with revision F. compares revisions F and F of F. Multiple files can be specified for all of the above options. All files must have the -extension C<.tex>, though. +extensions C<.tex>, C<.bbl>, or C<.flt>, though. =item B F F @@ -701,7 +700,7 @@ reported to stdout. Rather than appending the string C and optionally the version numbers given to the output-file, this will prepend a directory name C to the -original filename, creating the directory and subdirectories should they not exist already. This is particularly useful in order to clone a +original filename, creating the directory and sub-directories should they not exist already. This is particularly useful in order to clone a complete directory hierarchy. Optionally, a pathname F can be specified, which is prepended instead of C. =item B<--flatten,--flatten=keep-intermediate> @@ -713,9 +712,6 @@ The generic usage of this function is : C, the intermediate revision snapshots are kept in the current directory (Default is to store them in a temporary directory and delete them after generating the diff file.) -=item B<--config var1=val1,var2=val2,...> or B<-c var1=val1,..> - - =item B<--only-changes> @@ -738,6 +734,7 @@ run latex command on diff file after generation of diff file. run latex and dvixxx commands after generation of diff file. =item B<-c configfile> +=item B<--config var1=val1,var2=val2,...> or B<-c var1=val1,..> Set configuration variables for latexdiff and latexdiff-vc. The option can be repeated to set different variables (as an alternative to the comma-separated list). @@ -755,7 +752,7 @@ Available variables for latexdiff-vc: =back -All other config variables are passed to latexdiff. Explicity set configuration changes always override implicit +All other config variables are passed to latexdiff. Explicitly set configuration changes always override implicit changes by the following shortcut options B<--fast>, B<--so>, B<--ps> and B<--pdf>. =item B<--fast> or B<--so> @@ -803,7 +800,7 @@ L =head1 PORTABILITY I uses external commands and is therefore dependent on the system architecture; it has been -tested mainly on Unix-like systems. It also requires the a version control +tested mainly on Unix-like systems. It also requires a version control system and latex to be installed on the system to make use of all features. Modules from Perl 5.8 or higher are required. diff --git a/Build/source/texk/texlive/linked_scripts/latexdiff/latexdiff.pl b/Build/source/texk/texlive/linked_scripts/latexdiff/latexdiff.pl index 6d07c9a88f8..81387135c58 100755 --- a/Build/source/texk/texlive/linked_scripts/latexdiff/latexdiff.pl +++ b/Build/source/texk/texlive/linked_scripts/latexdiff/latexdiff.pl @@ -23,6 +23,25 @@ # Detailed usage information at the end of the file # + +# Version 1.3.1 +# Bug fixes: +# - remove some uninitialised variable $2 warnings in string substitution in flatten function in case included file is not found +# - add minimal postprocessing to diff processing of preamble commands (replace \RIGHTBRACE by \} ) +# - pre-processing: replace (contributed) routine take_comments_and_enter_from_frac() with take_comments_and_newline_from_frac(), which does the same thing +# (remove whitespace characters and comments between the argument of \frac commands) in an easier and more robust way. In addition, it +# will replace commands like \frac12 with \frac{1}{2} as pre-processing step. Fixes issue #184 +# - add "intertext" to list of unsafe math commands @UNSAFEMATHCMD . Fixes issue #179 +# - provide citation command patterns for biblatex and protect them with mbox'es. Fixes issue #199 +# - hardcode number of parameters for \href and \url commands to allow spaces between commands and arguments even if --allow-spaces option is not used (this +# is needed because some bibliography styles add such in-command-sequence spaces) Fixes issues: #178 #198 +# - bibitem is now kept even in deleted blocks such that deleted references show up properly (this implies that the actual numbers in numerical referencing schemes will change) +# (this is implemented by introducing a new class of commands KEEPCMD , which are kept as is in deleted environments (no effect in added environments). Currently +# \bibitem is hardwired to be the only member of this class (fixes issue #194, #174) +# Features: +# - add some special processing for revtex bibliography commands, so that the spaces between bibliography commands \bibfield and \bibinfo and their arguments are ignored. +# (fixes issue #194, should fix #174) +# # Version 1.3.0 (7 October 2018): # - treat options to \documentclass as potential package names (some packages allow implicit loading of or imply selected packages # - improved pattern matching: now allows nested angular brackets, and is no longer confused by escaped curly braces @@ -34,6 +53,9 @@ # - added to textcmds: \intertext # - new config variable CUSTOMDIFCMD to allow defining special versions of commands in added or deleted blocks (Pull request by github user jprotze) # - added option -no-links (mostly for use by latexdiff-vc in only-changes modes) (Pull request by github user jprotze) +# - new option --filter-script to run both input through a pre-processing script (PR jasonmccsmith #167) +# new option --no-filter-stderr to hide stderr output from filter-script (potentially dangerous, as this might hide malfunctioning of filter scripts) +# - --flatten now can deal with imports made using the import package {PR jasonmccsmith #173) # Bug fixes: # - pattern matching of \verb and \lstinline commands had an error which meant they would trigger on commands beginning with \verb. # - In description environments, mark up item descriptions by effectively reating the insides of item commannds as text commands (fixes #161) @@ -176,7 +198,7 @@ # package and use the standard latexdiff version instead # (current distribution available from http://search.cpan.org/~nedkonz/Algorithm-Diff-1.15 # the most recent version can be found via http://search.cpan.org/search?module=Algorithm::Diff ) -# Please note that the LICENCSE for Algorithm::Diff : +# Please note the LICENCE for Algorithm::Diff : # "Copyright (c) 2000-2002 Ned Konz. All rights reserved. # This program is free software; # you can redistribute it and/or modify it under the same terms @@ -630,11 +652,13 @@ use strict ; use warnings; use utf8 ; +use File::Spec ; + my ($algodiffversion)=split(/ /,$Algorithm::Diff::VERSION); my ($versionstring)=< undef, # COUNTERCMD textcmds which are associated with a counter # If any of these commands occur in a deleted block - # they will be succeeded by an \addtocounter{...}{-1} + # they will be followed by an \addtocounter{...}{-1} # for the associated counter such that the overall numbers # should be the same as in the new file LISTENV => undef , # list making environments - they will generally be kept @@ -704,7 +728,7 @@ my ($ARRENV, my $LABELCMD='(?:label)'; # matching commands are disabled within deleted blocks - mostly useful for maths mode, as otherwise it would be fine to just not add those to SAFECMDLIST -my @UNSAFEMATHCMD=('qedhere'); # Commands which are definitely unsafe for marking up in math mode (amsmath qedhere only tested to not work with UNDERLINE markup) (only affects WHOLE and COARSE math markup modes). Note that unlike text mode (or FINE math mode0 deleted unsafe commands are not deleted but simply taken outside \DIFdel +my @UNSAFEMATHCMD=('qedhere','intertext'); # Commands which are definitely unsafe for marking up in math mode (amsmath qedhere only tested to not work with UNDERLINE markup) (only affects WHOLE and COARSE math markup modes). Note that unlike text mode (or FINE math mode0 deleted unsafe commands are not deleted but simply taken outside \DIFdel my $MBOXINLINEMATH=0; # if set to 1 then surround marked-up inline maths expression with \mbox ( to get around compatibility # problems between some maths packages and ulem package @@ -759,6 +783,8 @@ our @SAFECMDEXCL=(); my @MBOXCMDLIST=(); # patterns for commands which are in principle safe but which need to be surrounded by an \mbox my @MBOXCMDEXCL=(); # all the patterns in MBOXCMDLIST will be appended to SAFECMDLIST +my @KEEPCMDLIST=( qr/^bibitem$/ ); # patterns for commands which should not be deleted in nominally delete text passages +my @KEEPCMDEXCL=(); my ($i,$j,$l); my ($old,$new); @@ -772,6 +798,7 @@ my ($oldpost, $newpost); my ($diffall); # Option names my ($type,$subtype,$floattype,$config,$preamblefile,$encoding,$nolabel,$visiblelabel, + $filterscript,$ignorefilterstderr, $showpreamble,$showsafe,$showtext,$showconfig,$showall, $replacesafe,$appendsafe,$excludesafe, $replacetext,$appendtext,$excludetext, @@ -873,6 +900,8 @@ GetOptions('type|t=s' => \$type, 'ignore-warnings' => \$ignorewarnings, 'driver=s'=> \$driver, 'flatten' => \$flatten, + 'filter-script=s' => \$filterscript, + 'ignore-filter-stderr' => \$ignorefilterstderr, 'no-links' => \$nolinks, 'version' => \$version, 'help|h' => \$help, @@ -917,7 +946,8 @@ if (defined($mathmarkup)) { # else use numerical value } - +# Give filterscript a default empty string +$filterscript="" unless defined($filterscript); # setting extra preamble commands if (defined($preamblefile)) { @@ -1075,17 +1105,20 @@ if ( $mathmarkup == COARSE || $mathmarkup == WHOLE ) { + + foreach $pkg ( @packagelist ) { map { $packages{$_}="" } split(/,/,$pkg) ; } - if ($showconfig || $showtext || $showsafe || $showpreamble) { + show_configuration(); exit 0; } + if ( @ARGV != 2 ) { print STDERR "2 and only 2 non-option arguments required. Write latexdiff -h to get help\n"; exit(2); @@ -1144,6 +1177,17 @@ push(@SAFECMDLIST, qr/^QLEFTBRACE$/, qr/^QRIGHTBRACE$/); # word: sequence of letters or accents followed by letter my $word_ja='\p{Han}+|\p{InHiragana}+|\p{InKatakana}+'; my $word='(?:' . $word_ja . '|(?:(?:[-\w\d*]|\\\\[\"\'\`~^][A-Za-z\*])(?!(?:' . $word_ja . ')))+)'; + + # for selected commands, the number of arguments is known, and we can therefore allow spaces between command and its argument + # Note that it is still expected that the arguments are blocks marked by parentheses rather than single characters, and that intervening comments will inhibit the association + my $predefinedcmdoptseq01='\\\\(?:url|BibitemShut)\s*\s*(?:\{'. $pat_n . '\}\s*){1}'; # Commands with one non-optional argument + my $predefinedcmdoptseq12='\\\\(?:href|bibfield|bibinfo)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\}\s*){2}'; # Commands with one optional and two non-optional arguments +# my $predefinedcmdoptseq11='\\\\(?:bibitem)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\}\s*){1}'; # Commands with one optional and one non-optional arguments +# \bibitem in revtex styles appears to be always followed by \BibItemOpen. We bind \BibItemOpen to the bibitem (if present) in order to prevent the comparison algorithm to interpret the \BibItemOpen as an identical part of the sequence; this interpretation can lead to added and removed entries to the reference list to become mixed. + my $predefinedbibitem='\\\\(?:bibitem)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\})(?:%?\s*\\\\BibitemOpen)?'; # Commands with one optional and one non-optional arguments + + my $predefinedcmdoptseq='(?:'.$predefinedcmdoptseq12.'|'.$predefinedcmdoptseq01.'|'.$predefinedbibitem.')'; + my $cmdleftright='\\\\(?:left|right|[Bb]igg?[lrm]?|middle)\s*(?:[<>()\[\]|\.]|\\\\(?:[|{}]|\w+))'; my $cmdoptseq='\\\\[\w\d@\*]+'.$extraspace.'(?:(?:<'.$abrat0.'>|\['.$brat_n.'\]|\{'. $pat_n . '\}|\(' . $coords .'\))'.$extraspace.')*'; my $defseq='\\\\def\\\\[\w\d@\*]+(?:#\d+|\[#\d+\])+(?:\{'. $pat_n . '\})?'; @@ -1152,7 +1196,7 @@ push(@SAFECMDLIST, qr/^QLEFTBRACE$/, qr/^QRIGHTBRACE$/); my $math='\$(?:[^$]|\\\$)*?\$|\\\\[(](?:.|\n)*?\\\\[)]'; ## the current maths command cannot cope with newline within the math expression my $comment='%[^\n]*\n'; - my $pat=qr/(?:\A\s*)?(?:${and}|${quotemarks}|${number}|${word}|$quotedunderscore|${defseq}|$cmdleftright|${cmdoptseq}|${math}|${backslashnl}|${oneletcmd}|${comment}|${punct}|${mathpunct}|\{|\})\s*/ ; + my $pat=qr/(?:\A\s*)?(?:${and}|${quotemarks}|${number}|${word}|$quotedunderscore|${defseq}|$cmdleftright|${predefinedcmdoptseq}|${cmdoptseq}|${math}|${backslashnl}|${oneletcmd}|${comment}|${punct}|${mathpunct}|\{|\})\s*/ ; @@ -1203,9 +1247,56 @@ if (lc($encoding) eq "utf8" ) { binmode(STDERR, ":utf8"); } +# filter($text) +# Runs $text through the script provided in $filterscript argument, if set +# If not set, just returns $text unchanged. +# If flatten was set, defer filtering to flatten. flatten will run the filter +# on all incoming text prior to its own processing. +# If flatten was not set, filter each of old and new once (see just below this def) +sub filter { + my ($text)=@_; + my ($textout,$pid); + if ($filterscript ne "") { + print STDERR "Passing " . length($text) . " chars to filter script " . $filterscript . "\n" if $verbose; + + if ($ignorefilterstderr) { + # If we need to capture and bury STDERR, use the Open3 version, and close CHLD_ERR below. + use IPC::Open3; + # We consume STDERR from the process, and hide it + $pid = open3(\*CHLD_IN, \*CHLD_OUT, \*CHLD_ERR, $filterscript) or die "open3() failed $!"; + } + else { + # Capture STDOUT and use as our new $text. Allow STDERR to go to console. + use IPC::Open2; + $pid = open2(\*CHLD_OUT, \*CHLD_IN, $filterscript) or die "open2() failed $!"; + } + # Send in $text + print CHLD_IN $text."\n"; # Adding a newline just to make sure there is one. + close CHLD_IN; + # Wait for output and gather it up + while () { + $textout = $textout.$_; + } + if ($ignorefilterstderr) { + close CHLD_ERR; # Enable only if Open3 used above + } + # On the off chance a very long running and/or frequently called script is used. + waitpid( $pid, 0 ); + $text = $textout; + print STDERR "Received " . length($text) . " chars after filtering\n" if $verbose; + print STDERR $text if $verbose; + } + return $text; +} + $old=read_file_with_encoding($oldfile,$encoding); $new=read_file_with_encoding($newfile,$encoding); +if (not defined($flatten)) { + $old=filter($old); + $new=filter($new); +} + @@ -1217,11 +1308,11 @@ exetime(1); if ($flatten) { - $oldbody=flatten($oldbody,$oldpreamble,$oldfile,$encoding); - $newbody=flatten($newbody,$newpreamble,$newfile,$encoding); + $oldbody=flatten($oldbody,$oldpreamble,File::Spec->rel2abs($oldfile),$encoding); + $newbody=flatten($newbody,$newpreamble,File::Spec->rel2abs($newfile),$encoding); # flatten preamble - $oldpreamble=flatten($oldpreamble,$oldpreamble,$oldfile,$encoding); - $newpreamble=flatten($newpreamble,$newpreamble,$newfile,$encoding); + $oldpreamble=flatten($oldpreamble,$oldpreamble,File::Spec->rel2abs($oldfile),$encoding); + $newpreamble=flatten($newpreamble,$newpreamble,File::Spec->rel2abs($newfile),$encoding); } @@ -1476,6 +1567,9 @@ my ( $citpat); if ( defined $packages{"apacite"} ) { print STDERR "apacite package detected.\n" if $verbose ; $citpat='(?:mask)?(?:full|short|no)?cite(?:A|author|year|meta)?(?:NP)?'; +} elsif (defined $packages{"biblatex"}) { + print STDERR "biblatex package detected.\n" if $verbose ; + $citpat='(?:[cC]ites?|(?:[pP]aren|foot|[Tt]ext|[sS]mart|super)cites?\*?|footnotecitetex)'; } else { # citation command pattern for all other citation schemes $citpat='(?:cite\w*|nocite)'; @@ -1798,11 +1892,13 @@ sub remove_endinput { # encoding is the encoding sub flatten { my ($text,$preamble,$filename,$encoding)=@_; - my ($includeonly,$dirname,$fname,$newpage,$fullfile,$filecontent,$replacement,$begline,$bblfile,$subfile,$command,$verbenv,$verboptions); + my ($includeonly,$dirname,$fname,$newpage,$fullfile,$filecontent,$replacement,$begline,$inputcmd,$bblfile,$subfile,$command,$verbenv,$verboptions,$ignore,$fileonly); my ($subpreamble,$subbody,$subpost); + my ($subdir,$subdirfull,$importfilepath); require File::Basename ; - require File::Spec ; - $dirname = File::Basename::dirname($filename); + ### require File::Spec ; # now this is needed even if flatten option not given + $filename = File::Spec->rel2abs( $filename ) ; + ($ignore, $dirname, $fileonly) = File::Spec->splitpath($filename) ; $bblfile = $filename; $bblfile=~s/\.tex$//; $bblfile.=".bbl"; @@ -1815,11 +1911,51 @@ sub flatten { print STDERR "DEBUG: includeonly $includeonly\n" if $debug; + + # Run through filter, to let filterscript have a pass if it was set + $text = filter($text); + + # Recursively replace \\import and \\subimport files + $text =~ s/(^(?:[^%\n]|\\%)*)(\\subimport\{(.*?)\}|\\import\{(.*?)\})(?:[\s]*)\{(.*?)\}/{ + # $1 is begline + # $3 is directory if subimport + # $4 is directory if import + # $5 is filename + $begline = (defined($1)? $1 : ""); + $subdir = $3 if defined($3); + $subdir = $4 if defined($4); + $fname = $5; + $fname .= ".tex" unless $fname =~ m|\.\w{3,4}$|; + print STDERR "DEBUG begline:", $begline, "\n" if $debug; + print STDERR "DEBUG", (defined($3)? "subimport_file:" : "import_file:"), $subdir, "\n" if $debug; + print STDERR "DEBUG file:", $fname, "\n" if $debug; + + # subimport appends $subdir to the current $dirname. import replaces it with an absolute path. + $subdirfull = (defined($3) ? File::Spec->catdir($dirname,$subdir) : $subdir); + + $importfilepath = File::Spec->catfile($subdirfull, $fname); + + print STDERR "importing importfilepath:", $importfilepath,"\n" if $verbose; + if ( -f $importfilepath ) { + # If file exists, replace input or include command with expanded input + $replacement=flatten(read_file_with_encoding($importfilepath, $encoding), $preamble,$importfilepath,$encoding) or die "Could not open file ",$fullfile,": $!"; + } else { + # if file does not exist, do not expand include or input command (do not warn if fname contains #[0-9] as it is then likely part of a command definition + # and is not meant to be expanded directly + print STDERR "WARNING: Could not find included file ",$importfilepath,". I will continue but not expand |$2|\n"; + $replacement=(defined($3)? "\\subimport" : "\\import"); + $replacement .= "{$subdir}{$fname} % Processed"; + } + "$begline$replacement"; + }/exgm; + # recursively replace \\input and \\include files $text =~ s/(^(?:[^%\n]|\\%)*)(\\input\{(.*?)\}|\\include\{(${includeonly}(?:\.tex)?)\})/{ $begline=(defined($1)? $1 : "") ; + $inputcmd=$2; $fname = $3 if defined($3) ; $fname = $4 if defined($4) ; + $newpage=(defined($4)? " \\newpage " : "") ; # # add tex extension unless there is a three or four letter extension already $fname .= ".tex" unless $fname =~ m|\.\w{3,4}$|; $fullfile = File::Spec->catfile($dirname,$fname); @@ -1832,12 +1968,11 @@ sub flatten { $replacement=flatten(read_file_with_encoding($fullfile, $encoding), $preamble,$filename,$encoding) or die "Could not open file ",$fullfile,": $!"; $replacement = remove_endinput($replacement); # \include always starts a new page; use explicit \newpage command to simulate this - $newpage=(defined($4)? " \\newpage " : "") ; } else { # if file does not exist, do not expand include or input command (do not warn if fname contains #[0-9] as it is then likely part of a command definition # and is not meant to be expanded directly - print STDERR "WARNING: Could not find included file ",$fullfile,". I will continue but not expand |$2|\n" unless $fname =~ m(#[0-9]) ; - $replacement = $2 ; # i.e. just the original command again -> make no change file does not exist + print STDERR "WARNING: Could not find included file ",$fullfile,". I will continue but not expand |$inputcmd|\n" unless $fname =~ m(#[0-9])n ; + $replacement = $inputcmd ; # i.e. just the original command again -> make no change file does not exist $newpage=""; } "$begline$newpage$replacement$newpage"; @@ -2116,6 +2251,7 @@ sub splitlatex { } } + unshift(@retval,$leadin) if (length($leadin)>0); return @retval; } @@ -2297,10 +2433,14 @@ sub extracttextblocks { my $wpat=qr/^(?:[a-zA-Z.,'`:;?()!]*)[\s~]*$/; #' my $retval=[]; + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace='\s*'; + for ($i=0;$i< scalar @$block;$i++) { ($token,$index)=@{ $block->[$i] }; # store pure text blocks - if ($token =~ /$wpat/ || ( $token =~/^\\((?:[`'^"~=.]|[\w\d@\*]+))((?:${extraspace}\[$brat_n\]${extraspace}|${extraspace}\{$pat_n\})*)/o + if ($token =~ /$wpat/ || ( $token =~/^\\((?:[`'^"~=.]|[\w\d@\*]+))((?:${extraspace}\[$brat_n\]${extraspace}|${extraspace}\{$pat_n\})*)/ && iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL) && !iscmd($1,\@TEXTCMDLIST,\@TEXTCMDEXCL))) { # we have text or a command which can be treated as text @@ -2345,6 +2485,10 @@ sub extractcommands { my ($i,$token,$index,$cmd,$open,$mid,$closing); my $retval=[]; + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace='\s*'; + for ($i=0;$i< scalar @$block;$i++) { ($token,$index)=@{ $block->[$i] }; # check if token is an alphanumeric command sequence with at least one non-optional argument @@ -2356,6 +2500,7 @@ sub extractcommands { # $4: } + trailing spaces if ( ( $token =~ m/^(\\([\w\d\*]+)(?:${extraspace}\[$brat_n\]|${extraspace}\{$pat_n\})*${extraspace}\{)($pat_n)(\}\s*)$/so ) && iscmd($2,\@TEXTCMDLIST,\@TEXTCMDEXCL) ) { + print STDERR "DEBUG EXTRACTCOMMANDS Match |$1|$2|$3|$4|$index \n" if $debug; # push(@$retval,[ $2,$index,$1,$3,$4 ]); ($cmd,$open,$mid,$closing) = ($2,$1,$3,$4) ; $closing =~ s/\}/\\RIGHTBRACE/ ; @@ -2371,20 +2516,20 @@ sub extractcommands { sub iscmd { my ($cmd,$regexar,$regexexcl)=@_; my ($ret)=0; - print STDERR "DEBUG: iscmd($cmd)=" if $debug; + ### print STDERR "DEBUG: iscmd($cmd)=" if $debug; foreach $pat ( @$regexar ) { if ( $cmd =~ m/^${pat}$/ ) { $ret=1 ; last; } } - print STDERR "0\n" if ($debug && !$ret) ; + ### print STDERR "0\n" if ($debug && !$ret) ; return 0 unless $ret; foreach $pat ( @$regexexcl ) { - print STDERR "0\n" if ( $debug && $cmd =~ m/^${pat}$/) ; + ### print STDERR "0\n" if ( $debug && $cmd =~ m/^${pat}$/) ; return 0 if ( $cmd =~ m/^${pat}$/ ); } - print STDERR "1\n" if $debug; + ### print STDERR "1\n" if $debug; return 1; } @@ -2443,7 +2588,6 @@ sub pass2 { print STDERR " $deltokcnt discarded tokens in $delblkcnt blocks.\n"; print STDERR " $addtokcnt appended tokens in $addblkcnt blocks.\n"; } - return(@$retval); } @@ -2466,13 +2610,20 @@ sub marktags { my $cmdcomment= ($opencmd =~ m/^%/); # Flag to indicate whether opencmd is a comment (i.e. if we intend to simply comment out changed commands) my ($command,$commandword,$closingbracket) ; # temporary variables needed below to remember sub-pattern matches -# split this block to flatten out sequences joined in pass1 + + +# split this block to split sequences joined in pass1 @$block=splitlatex(join "",@$block); ### print STDERR "DEBUG: marktags $openmark,$closemark,$open,$close,$opencmd,$closecmd,$comment\n" if $debug; ### print STDERR "DEBUG: marktags blocksplit ",join("|",@$block),"\n" if $debug; + + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace_mt='\s*'; + + foreach (@$block) { $word=$_; - ### print STDERR "DEBUG MARKTAGS: |$word|\n" if $debug; if ( $word =~ s/^%/%$comment/ ) { # a comment if ($cmd==1) { @@ -2497,9 +2648,7 @@ sub marktags { } # negative lookahead pattern (?!) in second clause is put in to avoid matching \( .. \) patterns # also note that second pattern will match \\ - print STDERR "DEBUG marktags: Considering word |$word|\n" if $debug; - if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\(?!\()(\\|[`'^"~=.]|[\w*@]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { - ###print STDERR "DEBUG MARKTAGS is a non-safe command ($1)\n" if $debug; + if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\(?!\()(\\|[`'^"~=.]|[\w*@]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { ### if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\([\w*@\\% ]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { # word is a command or other significant token (not in SAFECMDLIST) ## same conditions as in subroutine extractcommand: @@ -2511,13 +2660,13 @@ sub marktags { # $3: last argument # $4: } + trailing spaces ### pre-0.3 if ( ( $token =~ m/^(\\([\w\d\*]+)(?:\[$brat0\]|\{$pat_n\})*\{)($pat_n)(\}\s*)$/so ) - if ( ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace}\[$brat_n\]|${extraspace}\{$pat_n\})*${extraspace}\{)($pat_n)(\}\s*)$/so ) + if ( ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace_mt}\[$brat_n\]|${extraspace_mt}\{$pat_n\})*${extraspace_mt}\{)($pat_n)(\}\s*)$/so ) && (iscmd($2,\@TEXTCMDLIST,\@TEXTCMDEXCL)|| iscmd($2,\@MATHTEXTCMDLIST,\@MATHTEXTCMDEXCL)) && ( !$cmdcomment || !iscmd($2,\@CONTEXT2CMDLIST, \@CONTEXT2CMDEXCL) ) ) { # Condition 1: word is a command? - if yes, $1,$2,.. will be set as above # Condition 2: word is a text command - we mark up the interior of the word. There is a separate check for MATHTEXTCMDLIST # because for $mathmarkup=WHOLE, the commands should not be split in pass1 (ie. math mode commands are not in - # TEXTCMDLIST, but the interior of MATHTEXT commands should be highlighted in both deleted and added blocks + # TEXTCMDLIST, but the interior of MATHTEXT commnds should be highlighted in both deleted and added blocks # Condition 3: But if we are in a deleted block ($cmdcomment=1) and # $2 (the command) is in context2, just treat it as an ordinary command (i.e. comment it open with $opencmd) # Because we do not want to disable this command @@ -2553,6 +2702,13 @@ sub marktags { } push (@$retval,$AUXCMD,"\n") if $cmdcomment ; $cmd=-1 ; + } elsif ( $cmdcomment && + ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace_mt}\[$brat_n\]|${extraspace_mt}\{$pat_n\})*${extraspace_mt}\{)($pat_n)(\}\s*)/so ) + && iscmd($2,\@KEEPCMDLIST, \@KEEPCMDEXCL) ) { + # 'keepcmd' in a deleted environment: keep the command as is + push (@$retval,$close) if $cmd==0 ; + push (@$retval,$word); + $cmd=-1; # pretend we are at the beginning of a sequence because we do not want to add an additional $closecmd or $close before the next token, no matter what it is } else { # ordinary command push (@$retval,$opencmd) if $cmd==-1 ; @@ -2589,52 +2745,19 @@ sub marktags { } #used in preprocess -sub take_comments_and_enter_from_frac() { - #*************take the \n and % between frac and {}*********** - #notice all of the substitution are made none global - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\frac(([\s]*%[^\n]*?)*[\r\n|\r|\n])+\{(.*?)\\end\{\1}/s ) { - # if there isn't any % or \n in the pattern $2 then there should be an \\end{...} in $2 - ### print STDERR "Match the following in take_comments and_enter_from_frac(1):\n****$&****\n" if $debug; - if( $2 !~ m/\\end\{$1}/s ) { - # take out % and \n from the next match only (none global) - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\frac(([\s]*%[^\n]*?)*[\r\n|\r|\n])+\{(.*?)\\end\{\1}/\\begin{$1}$2\\frac{$5\\end{$1}/s; - } - else{ - #there are no more % and \n in $2, we want to find the next one so we clear the begin-end from the pattern - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\end\{\1}/MATHBLOCK$1\{$2\}MATHBLOCKEND/s; - } - } - ###cleaning up - while( s/MATHBLOCK($MATHENV|$MATHARRENV|SQUAREBRACKET)\{(.*?)\}MATHBLOCKEND/\\begin{$1}$2\\end{$1}/s ){} - ###*************take the \n and % between frac and {}*********** - - ###**********take the \n and % between {} and {} of the frac*************** - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\frac\{(.*?)\\end\{\1\}/s ) { - # if there isn't any more //frac before the first //end in the pattern $2 then there should be an \\end{...} in $2 - ###print STDERR "Match the following in take_comments and_enter_from_frac(2):\n****$&****\n" if $debug; - if( $2 !~ m/\\end\{$1\}/s ) { - # from now on CURRFRAC is the frac we are looking at - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\frac\{(.*?)\\end\{\1\}/\\begin\{$1\}$2CURRFRAC\{$3\\end\{$1\}/s; - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{(.*?)\\end\{\1\}/s ) { - if( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{($pat_n)\}([\s]*(%[^\n]*?)*[\r\n|\r|\n])+[\s]*\{(.*?)\\end\{\1}/s ) { - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{($pat_n)\}([\s]*(%[^\n]*?)*[\r\n|\r|\n])+[\s]*\{(.*?)\\end\{\1\}/\\begin\{$1\}$2CURRFRAC\{$3\}\{$6\\end\{$1\}/s; - } - else { # there is no comment or \n between the two brackets {}{} - # change CURRFRAC to FRACSTART so we can change them all back to //frac{ when we finish - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)CURRFRAC\{(.*?)\\end\{\1}/\\begin{$1}$2FRACSTART\{$3\\end{$1}/s; - } - } - } - else{ - ###there are no more frac in $2, we want to find the next one so we clear the begin-end from the pattern - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\end\{\1}/MATHBLOCK$1\{$2\}MATHBLOCKEND/s; - } - - } - ###cleaning up - while( s/MATHBLOCK($MATHENV|$MATHARRENV|SQUAREBRACKET)\{(.*?)\}MATHBLOCKEND/\\begin{$1}$2\\end{$1}/s ){} - s/FRACSTART/\\frac/g; - ###***************take the \n and % between {} and {} of the frac********************* +sub take_comments_and_newline_from_frac() { + # some special magic for common usage of frac, which does not conform to the latexdiff requirements but can be made to fit + # note that this is a rare exception to the general rule that the new tex can be reconstructed from the diff file + + # \frac12 -> \frac{1}{2} + s/\\frac(\d)(\w)/\\frac\{$1\}\{$2\}/g; + + # \frac1{2b} -> \frac{1}{2b} + s/\\frac(\d)/\\frac\{$1\}/g; + + # delete space and comment characters between \frac arguments +# s/\\frac(?:\s*?%[^\n]*?)*?(\{$pat_n\})\s*(\{$pat_n\})/\\frac$1$2/g; + s/\\frac(?:\s|%[^\n]*?)*(\{$pat_n\})(?:\s|%[^\n]*?)*(\{$pat_n\})/\\frac$1$2/g; } # preprocess($string, ..) @@ -2717,7 +2840,7 @@ sub preprocess { # resort to WHOLE or FINE, or NONE math mode processing. s/\\begin\{($ARRENV)}(.*?)\\end\{\1}/\\ARRAYBLOCK$1\{$2\\DIFANCHORARRB \}\\DIFANCHORARRE /sg; - take_comments_and_enter_from_frac(); + take_comments_and_newline_from_frac(); s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\end\{\1\}/\\MATHBLOCK$1\{$2\\DIFANCHORMATHB \}\\DIFANCHORMATHE /sg; } @@ -3335,8 +3458,7 @@ sub restore_item_commands { # pre-process preamble by looking for commands used in \maketitle (title, author, date etc commands) # the list of commands is defined in CONTEXT2CMD # if found then use a bodydiff to mark up content, and replace the corresponding commands - # in both preambles by marked up version to 'fool' the linediff (such that only body is marked - # up. + # in both preambles by marked up version to 'fool' the linediff (such that only body is marked up). # A special case are e.g. author commands being added (or removed) # 1. If commands are added, then the entire content is marked up as new, but also the lines are marked as new in the linediff # 2. If commands are removed, then the linediff will mark the line as deleted. The program returns @@ -3414,6 +3536,9 @@ EOF } $argnew=$newhash{$cmd}->[2]; $argdiff="{" . join("",bodydiff($argold,$argnew)) ."}"; + # Replace \RIGHTBRACE by } + $argdiff =~ s/\\RIGHTBRACE/}/g; + if ( length $optargnew ) { $optargdiff="[".join("",bodydiff($optargold,$optargnew))."]" ; $optargdiff =~ s/\\DIFaddbegin /\\DIFaddbeginFL /g; @@ -3759,6 +3884,7 @@ be real files (not pipes or similar) as they are opened twice. reason package specific parsing needs to be switched off, use --packages=none. The following packages trigger special behaviour: endfloat hyperref amsmath apacite siunitx cleveref glossaries mhchem chemformula/chemmacros + biblatex [ Default: scan the preamble for \\usepackage commands to determine loaded packages.] @@ -3840,7 +3966,17 @@ Miscelleneous options be located in the same directories as the old and new master files, respectively, making it possible to organise files into old and new directories. --flatten is applied recursively, so inputted files can contain further - \\input statements. + \\input statements. Also handles files included by the import package + (\\import and \\subimport), and \\subfile command. + +--filter-script=filterscript Run files through this filterscript (full path preferred) before processing. + The filterscript must take STDIN input and output to STDOUT. + When coupled with --flatten, each file will be run through the filter as it is brought in. + +--ignore-filter-stderr When running with --filter-script, STDERR from the script may cause readability issues. + Turn this flag on to ignore STDERR from the filter script. + + --help -h Show this help text. @@ -4095,7 +4231,7 @@ define new C<\DIFadd> and C<\DIFdel> commands, which provide a wrapper for these using them for the text but not for the link defining command (where any markup would cause errors). -=item C +=item C, C Redefine the commands recognised as citation commands. @@ -4401,12 +4537,26 @@ included sections are not expanded. The included files are assumed to be located in the same directories as the old and new master files, respectively, making it possible to organise files into old and new directories. --flatten is applied recursively, so inputted files can contain further -C<\input> statements. +C<\input> statements. Also handles files included by the import package +(C<\import> and C<\subimport>), and C<\subfile> command. Use of this option might result in prohibitive processing times for larger documents, and the resulting difference document no longer reflects the structure of the input documents. +=item B<--filter-script=filterscript> + +Run files through this filterscript (full path preferred) before processing. +The filterscript must take STDIN input and output to STDOUT. +When coupled with --flatten, each file will be run through the filter as it is brought in. + +=item B<--ignore-filter-stderr> + +When running with --filter-script, STDERR from the script may cause readability issues. +Turn this flag on to ignore STDERR from the filter script. + + + =item B<--help> or B<-h> @@ -4968,6 +5118,8 @@ min Pr sec sup +bibfield +bibinfo [Hclbkdruvt] [`'^"~=.] _ @@ -4994,6 +5146,7 @@ framebox footnote footnotetext framebox +href intertext part (sub){0,2}section\*? @@ -5007,6 +5160,7 @@ savebox sbox shortintertext shortstack +sidenote signature text.* value @@ -5151,7 +5305,7 @@ verbatim[*]? %DIF CULINECHBAR PREAMBLE \RequirePackage[normalem]{ulem} \RequirePackage[dvips]{changebar} -\RequirePackage{color} +\RequirePackage{color}\definecolor{RED}{rgb}{1,0,0}\definecolor{BLUE}{rgb}{0,0,1} \providecommand{\DIFadd}[1]{\protect\cbstart{\protect\color{blue}\uwave{#1}}\protect\cbend} \providecommand{\DIFdel}[1]{\protect\cbdelete{\protect\color{red}\sout{#1}}\protect\cbdelete} %DIF END CULINECHBAR PREAMBLE diff --git a/Build/source/texk/texlive/linked_scripts/latexdiff/latexrevise.pl b/Build/source/texk/texlive/linked_scripts/latexdiff/latexrevise.pl index 956d13e5061..f7f88f1b951 100755 --- a/Build/source/texk/texlive/linked_scripts/latexdiff/latexrevise.pl +++ b/Build/source/texk/texlive/linked_scripts/latexdiff/latexrevise.pl @@ -21,6 +21,7 @@ # # Detailed usage information at the end of the file # Note: version number now keeping up with latexdiff +# Version 1.3.1 (Bug fix) Remove some deprecation warnings due to uncommented left parentheses # Version 1.0.2 Option --version # Version 1.0.1 no changes to latexrevise # Version 0.3 Updated for compatibility with latexdiff 0.3 output (DIFAUXCMD removal) @@ -31,8 +32,8 @@ use strict; use warnings; my $versionstring=< or B<--comment=cmd> Remove C<\cmd{...}> sequences. C is supposed to mark some explicit -anotations which should be removed from the file before +annotations which should be removed from the file before release. =item B<-e envir> or B<--comment-environment=envir> @@ -508,8 +525,9 @@ which should have been removed already. =head1 BUGS The current version is a beta version which has not yet been -extensively tested, but worked fine locally. Please submit bug reports using the issue tracker of the github repository page I, -or send them to I.. Include the serial number of I +extensively tested. It has not been actively maintained so might not process output of newer versions of latexdiff entirely correctly. +Please submit bug reports using the issue tracker of the github repository page I, +or send them to I. Include the serial number of I (Option --version). If you come across latexdiff output which is not processed correctly by I please include the problem file as well as the old and new files on which it is based, diff --git a/Build/source/texk/texlive/linked_scripts/texlive/fmtutil.pl b/Build/source/texk/texlive/linked_scripts/texlive/fmtutil.pl index 814941ccff7..0cde1639919 100755 --- a/Build/source/texk/texlive/linked_scripts/texlive/fmtutil.pl +++ b/Build/source/texk/texlive/linked_scripts/texlive/fmtutil.pl @@ -1,5 +1,5 @@ #!/usr/bin/env perl -# $Id: fmtutil.pl 54303 2020-03-14 22:04:10Z karl $ +# $Id: fmtutil.pl 55294 2020-05-27 02:43:58Z preining $ # fmtutil - utility to maintain format files. # (Maintained in TeX Live:Master/texmf-dist/scripts/texlive.) # @@ -24,11 +24,11 @@ BEGIN { TeX::Update->import(); } -my $svnid = '$Id: fmtutil.pl 54303 2020-03-14 22:04:10Z karl $'; -my $lastchdate = '$Date: 2020-03-14 23:04:10 +0100 (Sat, 14 Mar 2020) $'; +my $svnid = '$Id: fmtutil.pl 55294 2020-05-27 02:43:58Z preining $'; +my $lastchdate = '$Date: 2020-05-27 04:43:58 +0200 (Wed, 27 May 2020) $'; $lastchdate =~ s/^\$Date:\s*//; $lastchdate =~ s/ \(.*$//; -my $svnrev = '$Revision: 54303 $'; +my $svnrev = '$Revision: 55294 $'; $svnrev =~ s/^\$Revision:\s*//; $svnrev =~ s/\s*\$$//; my $version = "r$svnrev ($lastchdate)"; @@ -440,7 +440,11 @@ sub callback_build_formats { # unless we tried to rebuild only missing formats. if ($what ne "missing") { if ($err + $suc == 0) { - print_info("did not find entry for $what=$whatarg, skipped\n"); + if ($what eq "all") { + print_warning("You seem to have no formats defined in your fmtutil.cnf files!\n"); + } else { + print_info("Did not find entry for $what=" . ($whatarg?$whatarg:"") . " skipped\n"); + } } } my $stdo = ($mktexfmtMode ? \*STDERR : \*STDOUT); diff --git a/Master/texmf-dist/doc/man/man1/latexdiff-vc.1 b/Master/texmf-dist/doc/man/man1/latexdiff-vc.1 index b4a00e943ed..c8340e71042 100644 --- a/Master/texmf-dist/doc/man/man1/latexdiff-vc.1 +++ b/Master/texmf-dist/doc/man/man1/latexdiff-vc.1 @@ -133,7 +133,7 @@ .\" ======================================================================== .\" .IX Title "LATEXDIFF-VC 1" -.TH LATEXDIFF-VC 1 "2018-10-07" "perl v5.22.1" " " +.TH LATEXDIFF-VC 1 "2020-05-27" "perl v5.22.1" " " .\" For nroff, turn off justification. Always turn off hyphenation; it makes .\" way too many mistakes in technical documents. .if n .ad l @@ -175,7 +175,7 @@ specified, and they result in different behaviour: .RS 4 .IP "\fBlatexdiff-vc\fR \-r \fIfile.tex\fR ..." 4 .IX Item "latexdiff-vc -r file.tex ..." -compares \fIfile.tex\fR with the most recent version checked into \s-1RCS.\s0 +compares \fIfile.tex\fR with the most recently checked-in version checked. .IP "\fBlatexdiff-vc\fR \-r \fIrev1\fR \fIfile.tex\fR ..." 4 .IX Item "latexdiff-vc -r rev1 file.tex ..." compares \fIfile.tex\fR with revision \fIrev1\fR. @@ -184,7 +184,7 @@ compares \fIfile.tex\fR with revision \fIrev1\fR. compares revisions \fIrev1\fR and \fIrev2\fR of \fIfile.tex\fR. .Sp Multiple files can be specified for all of the above options. All files must have the -extension \f(CW\*(C`.tex\*(C'\fR, though. +extensions \f(CW\*(C`.tex\*(C'\fR, \f(CW\*(C`.bbl\*(C'\fR, or \f(CW\*(C`.flt\*(C'\fR, though. .IP "\fBlatexdiff-vc\fR \fIold.tex\fR \fInew.tex\fR" 4 .IX Item "latexdiff-vc old.tex new.tex" compares two files. @@ -199,7 +199,7 @@ reported to stdout. Rather than appending the string \f(CW\*(C`diff\*(C'\fR and optionally the version numbers given to the output-file, this will prepend a directory name \f(CW\*(C`diff\*(C'\fR to the -original filename, creating the directory and subdirectories should they not exist already. This is particularly useful in order to clone a +original filename, creating the directory and sub-directories should they not exist already. This is particularly useful in order to clone a complete directory hierarchy. Optionally, a pathname \fIpath\fR can be specified, which is prepended instead of \f(CW\*(C`diff\*(C'\fR. .IP "\fB\-\-flatten,\-\-flatten=keep\-intermediate\fR" 4 .IX Item "--flatten,--flatten=keep-intermediate" @@ -209,12 +209,8 @@ Note that if additional files are needed which are not included in the flatten p The generic usage of this function is : \f(CW\*(C`latexdiff\-vc \-\-flatten \-r rev1 [\-r rev2] master.tex\*(C'\fR where master.tex is the project file containing the highest level of includes etc. .Sp With \f(CW\*(C`\-\-flatten=keep\-intermediate\*(C'\fR, the intermediate revision snapshots are kept in the current directory (Default is to store them in a temporary directory and delete them after generating the diff file.) -.IP "\fB\-\-config var1=val1,var2=val2,...\fR or \fB\-c var1=val1,..\fR" 4 -.IX Item "--config var1=val1,var2=val2,... or -c var1=val1,.." -.PD 0 .IP "\fB\-\-only\-changes\fR" 4 .IX Item "--only-changes" -.PD Post-process the output such that only pages with changes on them are displayed. This requires the use of subtype \s-1ZLABEL \s0 in latexdiff, which will be set automatically, but any manually set \-s option will be overruled (also requires zref package to be installed). (note that this option must be combined with \-\-ps or \-\-pdf to make sense) @@ -229,8 +225,8 @@ run latex command on diff file after generation of diff file. .IP "\fB\-\-dvi\fR" 4 .IX Item "--dvi" run latex and dvixxx commands after generation of diff file. -.IP "\fB\-c configfile\fR" 4 -.IX Item "-c configfile" +.IP "\fB\-c configfile\fR =item \fB\-\-config var1=val1,var2=val2,...\fR or \fB\-c var1=val1,..\fR" 4 +.IX Item "-c configfile =item --config var1=val1,var2=val2,... or -c var1=val1,.." Set configuration variables for latexdiff and latexdiff-vc. The option can be repeated to set different variables (as an alternative to the comma-separated list). Available variables for latexdiff-vc: @@ -252,7 +248,7 @@ Available variables for latexdiff-vc: .RS 4 .PD .Sp -All other config variables are passed to latexdiff. Explicity set configuration changes always override implicit +All other config variables are passed to latexdiff. Explicitly set configuration changes always override implicit changes by the following shortcut options \fB\-\-fast\fR, \fB\-\-so\fR, \fB\-\-ps\fR and \fB\-\-pdf\fR. .RE .IP "\fB\-\-fast\fR or \fB\-\-so\fR" 4 @@ -290,7 +286,7 @@ latexdiff .SH "PORTABILITY" .IX Header "PORTABILITY" \&\fIlatexdiff-vc\fR uses external commands and is therefore dependent on the system architecture; it has been -tested mainly on Unix-like systems. It also requires the a version control +tested mainly on Unix-like systems. It also requires a version control system and latex to be installed on the system to make use of all features. Modules from Perl 5.8 or higher are required. .SH "BUG REPORTING" diff --git a/Master/texmf-dist/doc/man/man1/latexdiff-vc.man1.pdf b/Master/texmf-dist/doc/man/man1/latexdiff-vc.man1.pdf index 65d92ea9bec..8e39d963772 100644 Binary files a/Master/texmf-dist/doc/man/man1/latexdiff-vc.man1.pdf and b/Master/texmf-dist/doc/man/man1/latexdiff-vc.man1.pdf differ diff --git a/Master/texmf-dist/doc/man/man1/latexdiff.1 b/Master/texmf-dist/doc/man/man1/latexdiff.1 index f48444e3074..f624ca652d2 100644 --- a/Master/texmf-dist/doc/man/man1/latexdiff.1 +++ b/Master/texmf-dist/doc/man/man1/latexdiff.1 @@ -133,7 +133,7 @@ .\" ======================================================================== .\" .IX Title "LATEXDIFF 1" -.TH LATEXDIFF 1 "2018-10-07" "perl v5.22.1" " " +.TH LATEXDIFF 1 "2020-05-26" "perl v5.22.1" " " .\" For nroff, turn off justification. Always turn off hyphenation; it makes .\" way too many mistakes in technical documents. .if n .ad l @@ -343,9 +343,9 @@ Change name of \f(CW\*(C`\eDIFadd\*(C'\fR and \f(CW\*(C`\eDIFdel\*(C'\fR command define new \f(CW\*(C`\eDIFadd\*(C'\fR and \f(CW\*(C`\eDIFdel\*(C'\fR commands, which provide a wrapper for these commands, using them for the text but not for the link defining command (where any markup would cause errors). -.ie n .IP """apacite""" 8 -.el .IP "\f(CWapacite\fR" 8 -.IX Item "apacite" +.ie n .IP """apacite"", ""biblatex""" 8 +.el .IP "\f(CWapacite\fR, \f(CWbiblatex\fR" 8 +.IX Item "apacite, biblatex" Redefine the commands recognised as citation commands. .ie n .IP """siunitx""" 8 .el .IP "\f(CWsiunitx\fR" 8 @@ -621,11 +621,21 @@ included sections are not expanded. The included files are assumed to be located in the same directories as the old and new master files, respectively, making it possible to organise files into old and new directories. \&\-\-flatten is applied recursively, so inputted files can contain further -\&\f(CW\*(C`\einput\*(C'\fR statements. +\&\f(CW\*(C`\einput\*(C'\fR statements. Also handles files included by the import package +(\f(CW\*(C`\eimport\*(C'\fR and \f(CW\*(C`\esubimport\*(C'\fR), and \f(CW\*(C`\esubfile\*(C'\fR command. .Sp Use of this option might result in prohibitive processing times for larger documents, and the resulting difference document no longer reflects the structure of the input documents. +.IP "\fB\-\-filter\-script=filterscript\fR" 4 +.IX Item "--filter-script=filterscript" +Run files through this filterscript (full path preferred) before processing. +The filterscript must take \s-1STDIN\s0 input and output to \s-1STDOUT.\s0 +When coupled with \-\-flatten, each file will be run through the filter as it is brought in. +.IP "\fB\-\-ignore\-filter\-stderr\fR" 4 +.IX Item "--ignore-filter-stderr" +When running with \-\-filter\-script, \s-1STDERR\s0 from the script may cause readability issues. +Turn this flag on to ignore \s-1STDERR\s0 from the filter script. .IP "\fB\-\-help\fR or \fB\-h\fR" 4 .IX Item "--help or -h" Show help text diff --git a/Master/texmf-dist/doc/man/man1/latexdiff.man1.pdf b/Master/texmf-dist/doc/man/man1/latexdiff.man1.pdf index 93c8dfd23e2..007cb1ddbff 100644 Binary files a/Master/texmf-dist/doc/man/man1/latexdiff.man1.pdf and b/Master/texmf-dist/doc/man/man1/latexdiff.man1.pdf differ diff --git a/Master/texmf-dist/doc/man/man1/latexrevise.1 b/Master/texmf-dist/doc/man/man1/latexrevise.1 index 49c8c35471b..234addcce27 100644 --- a/Master/texmf-dist/doc/man/man1/latexrevise.1 +++ b/Master/texmf-dist/doc/man/man1/latexrevise.1 @@ -133,7 +133,7 @@ .\" ======================================================================== .\" .IX Title "LATEXREVISE 1" -.TH LATEXREVISE 1 "2015-12-27" "perl v5.22.1" " " +.TH LATEXREVISE 1 "2020-05-27" "perl v5.22.1" " " .\" For nroff, turn off justification. Always turn off hyphenation; it makes .\" way too many mistakes in technical documents. .if n .ad l @@ -187,7 +187,7 @@ options. .IP "\fB\-c cmd\fR or \fB\-\-comment=cmd\fR" 4 .IX Item "-c cmd or --comment=cmd" Remove \f(CW\*(C`\ecmd{...}\*(C'\fR sequences. \f(CW\*(C`cmd\*(C'\fR is supposed to mark some explicit -anotations which should be removed from the file before +annotations which should be removed from the file before release. .IP "\fB\-e envir\fR or \fB\-\-comment\-environment=envir\fR" 4 .IX Item "-e envir or --comment-environment=envir" @@ -218,8 +218,9 @@ which should have been removed already. .SH "BUGS" .IX Header "BUGS" The current version is a beta version which has not yet been -extensively tested, but worked fine locally. Please submit bug reports using the issue tracker of the github repository page \fIhttps://github.com/ftilmann/latexdiff.git\fR, -or send them to \fItilmann \*(-- \s-1AT\s0 \*(-- gfz\-potsdam.de\fR.. Include the serial number of \fIlatexrevise\fR +extensively tested. It has not been actively maintained so might not process output of newer versions of latexdiff entirely correctly. +Please submit bug reports using the issue tracker of the github repository page \fIhttps://github.com/ftilmann/latexdiff.git\fR, +or send them to \fItilmann \*(-- \s-1AT\s0 \*(-- gfz\-potsdam.de\fR. Include the serial number of \fIlatexrevise\fR (Option \-\-version). If you come across latexdiff output which is not processed correctly by \fIlatexrevise\fR please include the problem file as well as the old and new files on which it is based, diff --git a/Master/texmf-dist/doc/man/man1/latexrevise.man1.pdf b/Master/texmf-dist/doc/man/man1/latexrevise.man1.pdf index 645084d0c82..5b002738906 100644 Binary files a/Master/texmf-dist/doc/man/man1/latexrevise.man1.pdf and b/Master/texmf-dist/doc/man/man1/latexrevise.man1.pdf differ diff --git a/Master/texmf-dist/doc/support/latexdiff/Makefile b/Master/texmf-dist/doc/support/latexdiff/Makefile index 143ddd213c2..cc5eb5ab797 100644 --- a/Master/texmf-dist/doc/support/latexdiff/Makefile +++ b/Master/texmf-dist/doc/support/latexdiff/Makefile @@ -12,7 +12,7 @@ default: @echo " of the PERL package Algorithm::Diff available from " @echo " the PERL archive www.cpan.org)" @echo " " - @echo "To install fast version (using UNIX diff) type: make install fast " + @echo "To install fast version (using UNIX diff) type: make install-fast " @echo " " @echo "To install the version which uses the system Algorithm::Diff package type: make install-ext" @echo " " @@ -62,3 +62,13 @@ test-fast: @echo "latexdiff example/example-draft.tex example/example-rev.tex (stand-alone version)" ./latexdiff-fast -V example/example-draft.tex example/example-rev.tex > example/example-diff.tex @echo "Difference file created: example/example-diff.tex" + +uninstall: +# -f is needed to ignore missing files + @rm -f $(INSTALLPATH)/bin/latexdiff{,-cvs,-git,-hg,-rcs,-so,-svn,-vc,-fast} $(INSTALLPATH)/bin/latexrevise + @rm -f $(INSTALLPATH)/man/man1/latex{diff{,-vc},revise}.1 +# maybe remove potentially empty directories + @rmdir --ignore-fail-on-non-empty $(INSTALLPATH)/bin + @rmdir --ignore-fail-on-non-empty $(INSTALLPATH)/man/man1 + @rmdir --ignore-fail-on-non-empty $(INSTALLPATH)/man + @rmdir --ignore-fail-on-non-empty $(INSTALLPATH) diff --git a/Master/texmf-dist/doc/support/latexdiff/README b/Master/texmf-dist/doc/support/latexdiff/README index b75fa841bc9..6f4d1a280a8 100644 --- a/Master/texmf-dist/doc/support/latexdiff/README +++ b/Master/texmf-dist/doc/support/latexdiff/README @@ -9,7 +9,7 @@ the latex source. Note that only files conforming to latex syntax will be processed correctly, not generic TeX files. Some further minor restrictions apply, see documentation. -A rudimentary revision facilility is provided by another Perl script, +A rudimentary revision facility is provided by another Perl script, latexrevise, which accepts or rejects all changes. Manual editing of the difference file can be used to override this default behaviour and accept or reject selected changes only. @@ -90,7 +90,7 @@ latexdiff-wrap (Author: V. Kuhlmann) An alternative wrapper script which can be latexdiff.spec (Author: T. Doerges) spec file for RPM generation -latexchanges (Author: Jan-Ake Larsson) Wrapper script for applying latexdiff with numbered documen version +latexchanges (Author: Jan-Ake Larsson) Wrapper script for applying latexdiff with numbered document version (see contrib/README.latexchanges for a more detailed description) Contributions by the following authors were incorporated into the latexdiff code, or inspired me to @@ -101,7 +101,7 @@ EXTERNAL LATEXDIFF SUPPORT PROGRAMS LATEXDIFFCITE (Author: Christer van der Meeren) is a wrapper around latexdiff to make citations diff properly. It works by expanding \cite type commands using the bbl or bib file, such that citations are treated just like normal text rather than as atomic in the plain latexdiff. https://latexdiffcite.readthedocs.org -GIT-LATEXDIFF (lead author: Matthieu Moy) is a wrapper (bash scipt) around latexdiff that allows using it to diff two revisions of a LaTeX file under git revision control (similar functionality is provided by latexdiff-vc --git with --flatten option included with this distribution but git-latexdiff allows more fine-grained control on (not to be confused with latexdiff-git, which is normally installed as a soft link to latexdiff-vc) +GIT-LATEXDIFF (lead author: Matthieu Moy) is a wrapper (bash script) around latexdiff that allows using it to diff two revisions of a LaTeX file under git revision control (similar functionality is provided by latexdiff-vc --git with --flatten option included with this distribution but git-latexdiff allows more fine-grained control on (not to be confused with latexdiff-git, which is normally installed as a soft link to latexdiff-vc) https://gitorious.org/git-latexdiff/ LICENSE (also see file COPYING) diff --git a/Master/texmf-dist/doc/support/latexdiff/contrib/latexdiff-wrap b/Master/texmf-dist/doc/support/latexdiff/contrib/latexdiff-wrap index 894b424de36..ed569d6095a 100755 --- a/Master/texmf-dist/doc/support/latexdiff/contrib/latexdiff-wrap +++ b/Master/texmf-dist/doc/support/latexdiff/contrib/latexdiff-wrap @@ -3,7 +3,7 @@ # latexdiff-wrap # # Wrapper for latexdiff, to -# * provide support for documents consiting of more than 1 latex file +# * provide support for documents consisting of more than 1 latex file # * provide my common arguments # # Copyright (C) by Volker Kuhlmann @@ -174,7 +174,7 @@ while [ $# -gt 0 ]; do ) else # Delete file, to make sure it's not clobbered by redirecting stdout - # in case it's a symlink to te original. + # in case it's a symlink to the original. test -f "$diffdir/$file" && (set -x rm "$diffdir/$file" ) diff --git a/Master/texmf-dist/doc/support/latexdiff/doc/latexdiff-man.pdf b/Master/texmf-dist/doc/support/latexdiff/doc/latexdiff-man.pdf index 32bfa06e2bd..d5a97b67cd3 100644 Binary files a/Master/texmf-dist/doc/support/latexdiff/doc/latexdiff-man.pdf and b/Master/texmf-dist/doc/support/latexdiff/doc/latexdiff-man.pdf differ diff --git a/Master/texmf-dist/doc/support/latexdiff/doc/latexdiff-man.tex b/Master/texmf-dist/doc/support/latexdiff/doc/latexdiff-man.tex index 1059eb25a3f..828f598b2b9 100644 --- a/Master/texmf-dist/doc/support/latexdiff/doc/latexdiff-man.tex +++ b/Master/texmf-dist/doc/support/latexdiff/doc/latexdiff-man.tex @@ -20,7 +20,7 @@ latex packages such as {\em color.sty}. Changes not directly affecting visible text, for example in formatting commands, are still marked in the latex source. -A rudimentary revision facilility is provided by another Perl script, +A rudimentary revision facility is provided by another Perl script, {\em latexrevise}, which accepts or rejects all changes. Manual editing of the difference file can be used to override this default behaviour and accept or reject selected changes only. @@ -201,7 +201,7 @@ To compare both revisions, type latexdiff -t UNDERLINE example-draft.tex example-rev.tex > example-diff.tex \end{verbatim} This results in the following difference file (a few newlines have been -added in this listing for legibility reasosn): +added in this listing for legibility reasons): {\scriptsize \begin{verbatim} \documentclass[12pt,a4paper]{article} @@ -359,7 +359,7 @@ The following is an incomplete list of wrappers written by others providing some \begin{description} \item[latexdiffcite] (Author: Christer van der Meeren) is a wrapper around latexdiff to make citations diff properly. It works by expanding \verb|\cite| type commands using the bbl or bib file, such that citations are treated just like normal text rather than as atomic in the plain latexdiff. \\ \verb|https://latexdiffcite.readthedocs.org| -\item[git-latexdiff] (lead author: Matthieu Moy) is a wrapper (bash scipt) around latexdiff that allows using it to diff two revisions of a LaTeX file under git revision control Similar functionality is provided by \verb|latexdiff-vc --git| with \verb|--flatten| option included with this distribution but git-latexdiff allows more fine-grained control on varous aspects. (Not to be confused with latexdiff-git, which is normally installed as a soft link to latexdiff-vc) \\ +\item[git-latexdiff] (lead author: Matthieu Moy) is a wrapper (bash script) around latexdiff that allows using it to diff two revisions of a \LaTeX file under git revision control Similar functionality is provided by \verb|latexdiff-vc --git| with \verb|--flatten| option included with this distribution but git-latexdiff allows more fine-grained control on various aspects. (Not to be confused with latexdiff-git, which is normally installed as a soft link to latexdiff-vc) \\ \verb|https://gitorious.org/git-latexdiff/| \end{description} diff --git a/Master/texmf-dist/doc/support/latexdiff/latexdiff b/Master/texmf-dist/doc/support/latexdiff/latexdiff index 227cc2c45e4..b8fde01b78a 100755 --- a/Master/texmf-dist/doc/support/latexdiff/latexdiff +++ b/Master/texmf-dist/doc/support/latexdiff/latexdiff @@ -23,6 +23,25 @@ # Detailed usage information at the end of the file # + +# Version 1.3.1 +# Bug fixes: +# - remove some uninitialised variable $2 warnings in string substitution in flatten function in case included file is not found +# - add minimal postprocessing to diff processing of preamble commands (replace \RIGHTBRACE by \} ) +# - pre-processing: replace (contributed) routine take_comments_and_enter_from_frac() with take_comments_and_newline_from_frac(), which does the same thing +# (remove whitespace characters and comments between the argument of \frac commands) in an easier and more robust way. In addition, it +# will replace commands like \frac12 with \frac{1}{2} as pre-processing step. Fixes issue #184 +# - add "intertext" to list of unsafe math commands @UNSAFEMATHCMD . Fixes issue #179 +# - provide citation command patterns for biblatex and protect them with mbox'es. Fixes issue #199 +# - hardcode number of parameters for \href and \url commands to allow spaces between commands and arguments even if --allow-spaces option is not used (this +# is needed because some bibliography styles add such in-command-sequence spaces) Fixes issues: #178 #198 +# - bibitem is now kept even in deleted blocks such that deleted references show up properly (this implies that the actual numbers in numerical referencing schemes will change) +# (this is implemented by introducing a new class of commands KEEPCMD , which are kept as is in deleted environments (no effect in added environments). Currently +# \bibitem is hardwired to be the only member of this class (fixes issue #194, #174) +# Features: +# - add some special processing for revtex bibliography commands, so that the spaces between bibliography commands \bibfield and \bibinfo and their arguments are ignored. +# (fixes issue #194, should fix #174) +# # Version 1.3.0 (7 October 2018): # - treat options to \documentclass as potential package names (some packages allow implicit loading of or imply selected packages # - improved pattern matching: now allows nested angular brackets, and is no longer confused by escaped curly braces @@ -34,6 +53,9 @@ # - added to textcmds: \intertext # - new config variable CUSTOMDIFCMD to allow defining special versions of commands in added or deleted blocks (Pull request by github user jprotze) # - added option -no-links (mostly for use by latexdiff-vc in only-changes modes) (Pull request by github user jprotze) +# - new option --filter-script to run both input through a pre-processing script (PR jasonmccsmith #167) +# new option --no-filter-stderr to hide stderr output from filter-script (potentially dangerous, as this might hide malfunctioning of filter scripts) +# - --flatten now can deal with imports made using the import package {PR jasonmccsmith #173) # Bug fixes: # - pattern matching of \verb and \lstinline commands had an error which meant they would trigger on commands beginning with \verb. # - In description environments, mark up item descriptions by effectively reating the insides of item commannds as text commands (fixes #161) @@ -173,11 +195,13 @@ use strict ; use warnings; use utf8 ; +use File::Spec ; + my ($algodiffversion)=split(/ /,$Algorithm::Diff::VERSION); my ($versionstring)=< undef, # COUNTERCMD textcmds which are associated with a counter # If any of these commands occur in a deleted block - # they will be succeeded by an \addtocounter{...}{-1} + # they will be followed by an \addtocounter{...}{-1} # for the associated counter such that the overall numbers # should be the same as in the new file LISTENV => undef , # list making environments - they will generally be kept @@ -247,7 +271,7 @@ my ($ARRENV, my $LABELCMD='(?:label)'; # matching commands are disabled within deleted blocks - mostly useful for maths mode, as otherwise it would be fine to just not add those to SAFECMDLIST -my @UNSAFEMATHCMD=('qedhere'); # Commands which are definitely unsafe for marking up in math mode (amsmath qedhere only tested to not work with UNDERLINE markup) (only affects WHOLE and COARSE math markup modes). Note that unlike text mode (or FINE math mode0 deleted unsafe commands are not deleted but simply taken outside \DIFdel +my @UNSAFEMATHCMD=('qedhere','intertext'); # Commands which are definitely unsafe for marking up in math mode (amsmath qedhere only tested to not work with UNDERLINE markup) (only affects WHOLE and COARSE math markup modes). Note that unlike text mode (or FINE math mode0 deleted unsafe commands are not deleted but simply taken outside \DIFdel my $MBOXINLINEMATH=0; # if set to 1 then surround marked-up inline maths expression with \mbox ( to get around compatibility # problems between some maths packages and ulem package @@ -302,6 +326,8 @@ our @SAFECMDEXCL=(); my @MBOXCMDLIST=(); # patterns for commands which are in principle safe but which need to be surrounded by an \mbox my @MBOXCMDEXCL=(); # all the patterns in MBOXCMDLIST will be appended to SAFECMDLIST +my @KEEPCMDLIST=( qr/^bibitem$/ ); # patterns for commands which should not be deleted in nominally delete text passages +my @KEEPCMDEXCL=(); my ($i,$j,$l); my ($old,$new); @@ -315,6 +341,7 @@ my ($oldpost, $newpost); my ($diffall); # Option names my ($type,$subtype,$floattype,$config,$preamblefile,$encoding,$nolabel,$visiblelabel, + $filterscript,$ignorefilterstderr, $showpreamble,$showsafe,$showtext,$showconfig,$showall, $replacesafe,$appendsafe,$excludesafe, $replacetext,$appendtext,$excludetext, @@ -416,6 +443,8 @@ GetOptions('type|t=s' => \$type, 'ignore-warnings' => \$ignorewarnings, 'driver=s'=> \$driver, 'flatten' => \$flatten, + 'filter-script=s' => \$filterscript, + 'ignore-filter-stderr' => \$ignorefilterstderr, 'no-links' => \$nolinks, 'version' => \$version, 'help|h' => \$help, @@ -460,7 +489,8 @@ if (defined($mathmarkup)) { # else use numerical value } - +# Give filterscript a default empty string +$filterscript="" unless defined($filterscript); # setting extra preamble commands if (defined($preamblefile)) { @@ -618,17 +648,20 @@ if ( $mathmarkup == COARSE || $mathmarkup == WHOLE ) { + + foreach $pkg ( @packagelist ) { map { $packages{$_}="" } split(/,/,$pkg) ; } - if ($showconfig || $showtext || $showsafe || $showpreamble) { + show_configuration(); exit 0; } + if ( @ARGV != 2 ) { print STDERR "2 and only 2 non-option arguments required. Write latexdiff -h to get help\n"; exit(2); @@ -687,6 +720,17 @@ push(@SAFECMDLIST, qr/^QLEFTBRACE$/, qr/^QRIGHTBRACE$/); # word: sequence of letters or accents followed by letter my $word_ja='\p{Han}+|\p{InHiragana}+|\p{InKatakana}+'; my $word='(?:' . $word_ja . '|(?:(?:[-\w\d*]|\\\\[\"\'\`~^][A-Za-z\*])(?!(?:' . $word_ja . ')))+)'; + + # for selected commands, the number of arguments is known, and we can therefore allow spaces between command and its argument + # Note that it is still expected that the arguments are blocks marked by parentheses rather than single characters, and that intervening comments will inhibit the association + my $predefinedcmdoptseq01='\\\\(?:url|BibitemShut)\s*\s*(?:\{'. $pat_n . '\}\s*){1}'; # Commands with one non-optional argument + my $predefinedcmdoptseq12='\\\\(?:href|bibfield|bibinfo)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\}\s*){2}'; # Commands with one optional and two non-optional arguments +# my $predefinedcmdoptseq11='\\\\(?:bibitem)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\}\s*){1}'; # Commands with one optional and one non-optional arguments +# \bibitem in revtex styles appears to be always followed by \BibItemOpen. We bind \BibItemOpen to the bibitem (if present) in order to prevent the comparison algorithm to interpret the \BibItemOpen as an identical part of the sequence; this interpretation can lead to added and removed entries to the reference list to become mixed. + my $predefinedbibitem='\\\\(?:bibitem)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\})(?:%?\s*\\\\BibitemOpen)?'; # Commands with one optional and one non-optional arguments + + my $predefinedcmdoptseq='(?:'.$predefinedcmdoptseq12.'|'.$predefinedcmdoptseq01.'|'.$predefinedbibitem.')'; + my $cmdleftright='\\\\(?:left|right|[Bb]igg?[lrm]?|middle)\s*(?:[<>()\[\]|\.]|\\\\(?:[|{}]|\w+))'; my $cmdoptseq='\\\\[\w\d@\*]+'.$extraspace.'(?:(?:<'.$abrat0.'>|\['.$brat_n.'\]|\{'. $pat_n . '\}|\(' . $coords .'\))'.$extraspace.')*'; my $defseq='\\\\def\\\\[\w\d@\*]+(?:#\d+|\[#\d+\])+(?:\{'. $pat_n . '\})?'; @@ -695,7 +739,7 @@ push(@SAFECMDLIST, qr/^QLEFTBRACE$/, qr/^QRIGHTBRACE$/); my $math='\$(?:[^$]|\\\$)*?\$|\\\\[(](?:.|\n)*?\\\\[)]'; ## the current maths command cannot cope with newline within the math expression my $comment='%[^\n]*\n'; - my $pat=qr/(?:\A\s*)?(?:${and}|${quotemarks}|${number}|${word}|$quotedunderscore|${defseq}|$cmdleftright|${cmdoptseq}|${math}|${backslashnl}|${oneletcmd}|${comment}|${punct}|${mathpunct}|\{|\})\s*/ ; + my $pat=qr/(?:\A\s*)?(?:${and}|${quotemarks}|${number}|${word}|$quotedunderscore|${defseq}|$cmdleftright|${predefinedcmdoptseq}|${cmdoptseq}|${math}|${backslashnl}|${oneletcmd}|${comment}|${punct}|${mathpunct}|\{|\})\s*/ ; @@ -746,9 +790,56 @@ if (lc($encoding) eq "utf8" ) { binmode(STDERR, ":utf8"); } +# filter($text) +# Runs $text through the script provided in $filterscript argument, if set +# If not set, just returns $text unchanged. +# If flatten was set, defer filtering to flatten. flatten will run the filter +# on all incoming text prior to its own processing. +# If flatten was not set, filter each of old and new once (see just below this def) +sub filter { + my ($text)=@_; + my ($textout,$pid); + if ($filterscript ne "") { + print STDERR "Passing " . length($text) . " chars to filter script " . $filterscript . "\n" if $verbose; + + if ($ignorefilterstderr) { + # If we need to capture and bury STDERR, use the Open3 version, and close CHLD_ERR below. + use IPC::Open3; + # We consume STDERR from the process, and hide it + $pid = open3(\*CHLD_IN, \*CHLD_OUT, \*CHLD_ERR, $filterscript) or die "open3() failed $!"; + } + else { + # Capture STDOUT and use as our new $text. Allow STDERR to go to console. + use IPC::Open2; + $pid = open2(\*CHLD_OUT, \*CHLD_IN, $filterscript) or die "open2() failed $!"; + } + # Send in $text + print CHLD_IN $text."\n"; # Adding a newline just to make sure there is one. + close CHLD_IN; + # Wait for output and gather it up + while () { + $textout = $textout.$_; + } + if ($ignorefilterstderr) { + close CHLD_ERR; # Enable only if Open3 used above + } + # On the off chance a very long running and/or frequently called script is used. + waitpid( $pid, 0 ); + $text = $textout; + print STDERR "Received " . length($text) . " chars after filtering\n" if $verbose; + print STDERR $text if $verbose; + } + return $text; +} + $old=read_file_with_encoding($oldfile,$encoding); $new=read_file_with_encoding($newfile,$encoding); +if (not defined($flatten)) { + $old=filter($old); + $new=filter($new); +} + @@ -760,11 +851,11 @@ exetime(1); if ($flatten) { - $oldbody=flatten($oldbody,$oldpreamble,$oldfile,$encoding); - $newbody=flatten($newbody,$newpreamble,$newfile,$encoding); + $oldbody=flatten($oldbody,$oldpreamble,File::Spec->rel2abs($oldfile),$encoding); + $newbody=flatten($newbody,$newpreamble,File::Spec->rel2abs($newfile),$encoding); # flatten preamble - $oldpreamble=flatten($oldpreamble,$oldpreamble,$oldfile,$encoding); - $newpreamble=flatten($newpreamble,$newpreamble,$newfile,$encoding); + $oldpreamble=flatten($oldpreamble,$oldpreamble,File::Spec->rel2abs($oldfile),$encoding); + $newpreamble=flatten($newpreamble,$newpreamble,File::Spec->rel2abs($newfile),$encoding); } @@ -1019,6 +1110,9 @@ my ( $citpat); if ( defined $packages{"apacite"} ) { print STDERR "apacite package detected.\n" if $verbose ; $citpat='(?:mask)?(?:full|short|no)?cite(?:A|author|year|meta)?(?:NP)?'; +} elsif (defined $packages{"biblatex"}) { + print STDERR "biblatex package detected.\n" if $verbose ; + $citpat='(?:[cC]ites?|(?:[pP]aren|foot|[Tt]ext|[sS]mart|super)cites?\*?|footnotecitetex)'; } else { # citation command pattern for all other citation schemes $citpat='(?:cite\w*|nocite)'; @@ -1341,11 +1435,13 @@ sub remove_endinput { # encoding is the encoding sub flatten { my ($text,$preamble,$filename,$encoding)=@_; - my ($includeonly,$dirname,$fname,$newpage,$fullfile,$filecontent,$replacement,$begline,$bblfile,$subfile,$command,$verbenv,$verboptions); + my ($includeonly,$dirname,$fname,$newpage,$fullfile,$filecontent,$replacement,$begline,$inputcmd,$bblfile,$subfile,$command,$verbenv,$verboptions,$ignore,$fileonly); my ($subpreamble,$subbody,$subpost); + my ($subdir,$subdirfull,$importfilepath); require File::Basename ; - require File::Spec ; - $dirname = File::Basename::dirname($filename); + ### require File::Spec ; # now this is needed even if flatten option not given + $filename = File::Spec->rel2abs( $filename ) ; + ($ignore, $dirname, $fileonly) = File::Spec->splitpath($filename) ; $bblfile = $filename; $bblfile=~s/\.tex$//; $bblfile.=".bbl"; @@ -1358,11 +1454,51 @@ sub flatten { print STDERR "DEBUG: includeonly $includeonly\n" if $debug; + + # Run through filter, to let filterscript have a pass if it was set + $text = filter($text); + + # Recursively replace \\import and \\subimport files + $text =~ s/(^(?:[^%\n]|\\%)*)(\\subimport\{(.*?)\}|\\import\{(.*?)\})(?:[\s]*)\{(.*?)\}/{ + # $1 is begline + # $3 is directory if subimport + # $4 is directory if import + # $5 is filename + $begline = (defined($1)? $1 : ""); + $subdir = $3 if defined($3); + $subdir = $4 if defined($4); + $fname = $5; + $fname .= ".tex" unless $fname =~ m|\.\w{3,4}$|; + print STDERR "DEBUG begline:", $begline, "\n" if $debug; + print STDERR "DEBUG", (defined($3)? "subimport_file:" : "import_file:"), $subdir, "\n" if $debug; + print STDERR "DEBUG file:", $fname, "\n" if $debug; + + # subimport appends $subdir to the current $dirname. import replaces it with an absolute path. + $subdirfull = (defined($3) ? File::Spec->catdir($dirname,$subdir) : $subdir); + + $importfilepath = File::Spec->catfile($subdirfull, $fname); + + print STDERR "importing importfilepath:", $importfilepath,"\n" if $verbose; + if ( -f $importfilepath ) { + # If file exists, replace input or include command with expanded input + $replacement=flatten(read_file_with_encoding($importfilepath, $encoding), $preamble,$importfilepath,$encoding) or die "Could not open file ",$fullfile,": $!"; + } else { + # if file does not exist, do not expand include or input command (do not warn if fname contains #[0-9] as it is then likely part of a command definition + # and is not meant to be expanded directly + print STDERR "WARNING: Could not find included file ",$importfilepath,". I will continue but not expand |$2|\n"; + $replacement=(defined($3)? "\\subimport" : "\\import"); + $replacement .= "{$subdir}{$fname} % Processed"; + } + "$begline$replacement"; + }/exgm; + # recursively replace \\input and \\include files $text =~ s/(^(?:[^%\n]|\\%)*)(\\input\{(.*?)\}|\\include\{(${includeonly}(?:\.tex)?)\})/{ $begline=(defined($1)? $1 : "") ; + $inputcmd=$2; $fname = $3 if defined($3) ; $fname = $4 if defined($4) ; + $newpage=(defined($4)? " \\newpage " : "") ; # # add tex extension unless there is a three or four letter extension already $fname .= ".tex" unless $fname =~ m|\.\w{3,4}$|; $fullfile = File::Spec->catfile($dirname,$fname); @@ -1375,12 +1511,11 @@ sub flatten { $replacement=flatten(read_file_with_encoding($fullfile, $encoding), $preamble,$filename,$encoding) or die "Could not open file ",$fullfile,": $!"; $replacement = remove_endinput($replacement); # \include always starts a new page; use explicit \newpage command to simulate this - $newpage=(defined($4)? " \\newpage " : "") ; } else { # if file does not exist, do not expand include or input command (do not warn if fname contains #[0-9] as it is then likely part of a command definition # and is not meant to be expanded directly - print STDERR "WARNING: Could not find included file ",$fullfile,". I will continue but not expand |$2|\n" unless $fname =~ m(#[0-9]) ; - $replacement = $2 ; # i.e. just the original command again -> make no change file does not exist + print STDERR "WARNING: Could not find included file ",$fullfile,". I will continue but not expand |$inputcmd|\n" unless $fname =~ m(#[0-9])n ; + $replacement = $inputcmd ; # i.e. just the original command again -> make no change file does not exist $newpage=""; } "$begline$newpage$replacement$newpage"; @@ -1659,6 +1794,7 @@ sub splitlatex { } } + unshift(@retval,$leadin) if (length($leadin)>0); return @retval; } @@ -1840,10 +1976,14 @@ sub extracttextblocks { my $wpat=qr/^(?:[a-zA-Z.,'`:;?()!]*)[\s~]*$/; #' my $retval=[]; + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace='\s*'; + for ($i=0;$i< scalar @$block;$i++) { ($token,$index)=@{ $block->[$i] }; # store pure text blocks - if ($token =~ /$wpat/ || ( $token =~/^\\((?:[`'^"~=.]|[\w\d@\*]+))((?:${extraspace}\[$brat_n\]${extraspace}|${extraspace}\{$pat_n\})*)/o + if ($token =~ /$wpat/ || ( $token =~/^\\((?:[`'^"~=.]|[\w\d@\*]+))((?:${extraspace}\[$brat_n\]${extraspace}|${extraspace}\{$pat_n\})*)/ && iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL) && !iscmd($1,\@TEXTCMDLIST,\@TEXTCMDEXCL))) { # we have text or a command which can be treated as text @@ -1888,6 +2028,10 @@ sub extractcommands { my ($i,$token,$index,$cmd,$open,$mid,$closing); my $retval=[]; + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace='\s*'; + for ($i=0;$i< scalar @$block;$i++) { ($token,$index)=@{ $block->[$i] }; # check if token is an alphanumeric command sequence with at least one non-optional argument @@ -1899,6 +2043,7 @@ sub extractcommands { # $4: } + trailing spaces if ( ( $token =~ m/^(\\([\w\d\*]+)(?:${extraspace}\[$brat_n\]|${extraspace}\{$pat_n\})*${extraspace}\{)($pat_n)(\}\s*)$/so ) && iscmd($2,\@TEXTCMDLIST,\@TEXTCMDEXCL) ) { + print STDERR "DEBUG EXTRACTCOMMANDS Match |$1|$2|$3|$4|$index \n" if $debug; # push(@$retval,[ $2,$index,$1,$3,$4 ]); ($cmd,$open,$mid,$closing) = ($2,$1,$3,$4) ; $closing =~ s/\}/\\RIGHTBRACE/ ; @@ -1914,20 +2059,20 @@ sub extractcommands { sub iscmd { my ($cmd,$regexar,$regexexcl)=@_; my ($ret)=0; - print STDERR "DEBUG: iscmd($cmd)=" if $debug; + ### print STDERR "DEBUG: iscmd($cmd)=" if $debug; foreach $pat ( @$regexar ) { if ( $cmd =~ m/^${pat}$/ ) { $ret=1 ; last; } } - print STDERR "0\n" if ($debug && !$ret) ; + ### print STDERR "0\n" if ($debug && !$ret) ; return 0 unless $ret; foreach $pat ( @$regexexcl ) { - print STDERR "0\n" if ( $debug && $cmd =~ m/^${pat}$/) ; + ### print STDERR "0\n" if ( $debug && $cmd =~ m/^${pat}$/) ; return 0 if ( $cmd =~ m/^${pat}$/ ); } - print STDERR "1\n" if $debug; + ### print STDERR "1\n" if $debug; return 1; } @@ -1986,7 +2131,6 @@ sub pass2 { print STDERR " $deltokcnt discarded tokens in $delblkcnt blocks.\n"; print STDERR " $addtokcnt appended tokens in $addblkcnt blocks.\n"; } - return(@$retval); } @@ -2009,13 +2153,20 @@ sub marktags { my $cmdcomment= ($opencmd =~ m/^%/); # Flag to indicate whether opencmd is a comment (i.e. if we intend to simply comment out changed commands) my ($command,$commandword,$closingbracket) ; # temporary variables needed below to remember sub-pattern matches -# split this block to flatten out sequences joined in pass1 + + +# split this block to split sequences joined in pass1 @$block=splitlatex(join "",@$block); ### print STDERR "DEBUG: marktags $openmark,$closemark,$open,$close,$opencmd,$closecmd,$comment\n" if $debug; ### print STDERR "DEBUG: marktags blocksplit ",join("|",@$block),"\n" if $debug; + + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace_mt='\s*'; + + foreach (@$block) { $word=$_; - ### print STDERR "DEBUG MARKTAGS: |$word|\n" if $debug; if ( $word =~ s/^%/%$comment/ ) { # a comment if ($cmd==1) { @@ -2040,9 +2191,7 @@ sub marktags { } # negative lookahead pattern (?!) in second clause is put in to avoid matching \( .. \) patterns # also note that second pattern will match \\ - print STDERR "DEBUG marktags: Considering word |$word|\n" if $debug; - if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\(?!\()(\\|[`'^"~=.]|[\w*@]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { - ###print STDERR "DEBUG MARKTAGS is a non-safe command ($1)\n" if $debug; + if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\(?!\()(\\|[`'^"~=.]|[\w*@]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { ### if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\([\w*@\\% ]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { # word is a command or other significant token (not in SAFECMDLIST) ## same conditions as in subroutine extractcommand: @@ -2054,13 +2203,13 @@ sub marktags { # $3: last argument # $4: } + trailing spaces ### pre-0.3 if ( ( $token =~ m/^(\\([\w\d\*]+)(?:\[$brat0\]|\{$pat_n\})*\{)($pat_n)(\}\s*)$/so ) - if ( ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace}\[$brat_n\]|${extraspace}\{$pat_n\})*${extraspace}\{)($pat_n)(\}\s*)$/so ) + if ( ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace_mt}\[$brat_n\]|${extraspace_mt}\{$pat_n\})*${extraspace_mt}\{)($pat_n)(\}\s*)$/so ) && (iscmd($2,\@TEXTCMDLIST,\@TEXTCMDEXCL)|| iscmd($2,\@MATHTEXTCMDLIST,\@MATHTEXTCMDEXCL)) && ( !$cmdcomment || !iscmd($2,\@CONTEXT2CMDLIST, \@CONTEXT2CMDEXCL) ) ) { # Condition 1: word is a command? - if yes, $1,$2,.. will be set as above # Condition 2: word is a text command - we mark up the interior of the word. There is a separate check for MATHTEXTCMDLIST # because for $mathmarkup=WHOLE, the commands should not be split in pass1 (ie. math mode commands are not in - # TEXTCMDLIST, but the interior of MATHTEXT commands should be highlighted in both deleted and added blocks + # TEXTCMDLIST, but the interior of MATHTEXT commnds should be highlighted in both deleted and added blocks # Condition 3: But if we are in a deleted block ($cmdcomment=1) and # $2 (the command) is in context2, just treat it as an ordinary command (i.e. comment it open with $opencmd) # Because we do not want to disable this command @@ -2096,6 +2245,13 @@ sub marktags { } push (@$retval,$AUXCMD,"\n") if $cmdcomment ; $cmd=-1 ; + } elsif ( $cmdcomment && + ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace_mt}\[$brat_n\]|${extraspace_mt}\{$pat_n\})*${extraspace_mt}\{)($pat_n)(\}\s*)/so ) + && iscmd($2,\@KEEPCMDLIST, \@KEEPCMDEXCL) ) { + # 'keepcmd' in a deleted environment: keep the command as is + push (@$retval,$close) if $cmd==0 ; + push (@$retval,$word); + $cmd=-1; # pretend we are at the beginning of a sequence because we do not want to add an additional $closecmd or $close before the next token, no matter what it is } else { # ordinary command push (@$retval,$opencmd) if $cmd==-1 ; @@ -2132,52 +2288,19 @@ sub marktags { } #used in preprocess -sub take_comments_and_enter_from_frac() { - #*************take the \n and % between frac and {}*********** - #notice all of the substitution are made none global - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\frac(([\s]*%[^\n]*?)*[\r\n|\r|\n])+\{(.*?)\\end\{\1}/s ) { - # if there isn't any % or \n in the pattern $2 then there should be an \\end{...} in $2 - ### print STDERR "Match the following in take_comments and_enter_from_frac(1):\n****$&****\n" if $debug; - if( $2 !~ m/\\end\{$1}/s ) { - # take out % and \n from the next match only (none global) - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\frac(([\s]*%[^\n]*?)*[\r\n|\r|\n])+\{(.*?)\\end\{\1}/\\begin{$1}$2\\frac{$5\\end{$1}/s; - } - else{ - #there are no more % and \n in $2, we want to find the next one so we clear the begin-end from the pattern - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\end\{\1}/MATHBLOCK$1\{$2\}MATHBLOCKEND/s; - } - } - ###cleaning up - while( s/MATHBLOCK($MATHENV|$MATHARRENV|SQUAREBRACKET)\{(.*?)\}MATHBLOCKEND/\\begin{$1}$2\\end{$1}/s ){} - ###*************take the \n and % between frac and {}*********** - - ###**********take the \n and % between {} and {} of the frac*************** - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\frac\{(.*?)\\end\{\1\}/s ) { - # if there isn't any more //frac before the first //end in the pattern $2 then there should be an \\end{...} in $2 - ###print STDERR "Match the following in take_comments and_enter_from_frac(2):\n****$&****\n" if $debug; - if( $2 !~ m/\\end\{$1\}/s ) { - # from now on CURRFRAC is the frac we are looking at - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\frac\{(.*?)\\end\{\1\}/\\begin\{$1\}$2CURRFRAC\{$3\\end\{$1\}/s; - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{(.*?)\\end\{\1\}/s ) { - if( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{($pat_n)\}([\s]*(%[^\n]*?)*[\r\n|\r|\n])+[\s]*\{(.*?)\\end\{\1}/s ) { - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{($pat_n)\}([\s]*(%[^\n]*?)*[\r\n|\r|\n])+[\s]*\{(.*?)\\end\{\1\}/\\begin\{$1\}$2CURRFRAC\{$3\}\{$6\\end\{$1\}/s; - } - else { # there is no comment or \n between the two brackets {}{} - # change CURRFRAC to FRACSTART so we can change them all back to //frac{ when we finish - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)CURRFRAC\{(.*?)\\end\{\1}/\\begin{$1}$2FRACSTART\{$3\\end{$1}/s; - } - } - } - else{ - ###there are no more frac in $2, we want to find the next one so we clear the begin-end from the pattern - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\end\{\1}/MATHBLOCK$1\{$2\}MATHBLOCKEND/s; - } - - } - ###cleaning up - while( s/MATHBLOCK($MATHENV|$MATHARRENV|SQUAREBRACKET)\{(.*?)\}MATHBLOCKEND/\\begin{$1}$2\\end{$1}/s ){} - s/FRACSTART/\\frac/g; - ###***************take the \n and % between {} and {} of the frac********************* +sub take_comments_and_newline_from_frac() { + # some special magic for common usage of frac, which does not conform to the latexdiff requirements but can be made to fit + # note that this is a rare exception to the general rule that the new tex can be reconstructed from the diff file + + # \frac12 -> \frac{1}{2} + s/\\frac(\d)(\w)/\\frac\{$1\}\{$2\}/g; + + # \frac1{2b} -> \frac{1}{2b} + s/\\frac(\d)/\\frac\{$1\}/g; + + # delete space and comment characters between \frac arguments +# s/\\frac(?:\s*?%[^\n]*?)*?(\{$pat_n\})\s*(\{$pat_n\})/\\frac$1$2/g; + s/\\frac(?:\s|%[^\n]*?)*(\{$pat_n\})(?:\s|%[^\n]*?)*(\{$pat_n\})/\\frac$1$2/g; } # preprocess($string, ..) @@ -2260,7 +2383,7 @@ sub preprocess { # resort to WHOLE or FINE, or NONE math mode processing. s/\\begin\{($ARRENV)}(.*?)\\end\{\1}/\\ARRAYBLOCK$1\{$2\\DIFANCHORARRB \}\\DIFANCHORARRE /sg; - take_comments_and_enter_from_frac(); + take_comments_and_newline_from_frac(); s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\end\{\1\}/\\MATHBLOCK$1\{$2\\DIFANCHORMATHB \}\\DIFANCHORMATHE /sg; } @@ -2878,8 +3001,7 @@ sub restore_item_commands { # pre-process preamble by looking for commands used in \maketitle (title, author, date etc commands) # the list of commands is defined in CONTEXT2CMD # if found then use a bodydiff to mark up content, and replace the corresponding commands - # in both preambles by marked up version to 'fool' the linediff (such that only body is marked - # up. + # in both preambles by marked up version to 'fool' the linediff (such that only body is marked up). # A special case are e.g. author commands being added (or removed) # 1. If commands are added, then the entire content is marked up as new, but also the lines are marked as new in the linediff # 2. If commands are removed, then the linediff will mark the line as deleted. The program returns @@ -2957,6 +3079,9 @@ EOF } $argnew=$newhash{$cmd}->[2]; $argdiff="{" . join("",bodydiff($argold,$argnew)) ."}"; + # Replace \RIGHTBRACE by } + $argdiff =~ s/\\RIGHTBRACE/}/g; + if ( length $optargnew ) { $optargdiff="[".join("",bodydiff($optargold,$optargnew))."]" ; $optargdiff =~ s/\\DIFaddbegin /\\DIFaddbeginFL /g; @@ -3302,6 +3427,7 @@ be real files (not pipes or similar) as they are opened twice. reason package specific parsing needs to be switched off, use --packages=none. The following packages trigger special behaviour: endfloat hyperref amsmath apacite siunitx cleveref glossaries mhchem chemformula/chemmacros + biblatex [ Default: scan the preamble for \\usepackage commands to determine loaded packages.] @@ -3383,7 +3509,17 @@ Miscelleneous options be located in the same directories as the old and new master files, respectively, making it possible to organise files into old and new directories. --flatten is applied recursively, so inputted files can contain further - \\input statements. + \\input statements. Also handles files included by the import package + (\\import and \\subimport), and \\subfile command. + +--filter-script=filterscript Run files through this filterscript (full path preferred) before processing. + The filterscript must take STDIN input and output to STDOUT. + When coupled with --flatten, each file will be run through the filter as it is brought in. + +--ignore-filter-stderr When running with --filter-script, STDERR from the script may cause readability issues. + Turn this flag on to ignore STDERR from the filter script. + + --help -h Show this help text. @@ -3638,7 +3774,7 @@ define new C<\DIFadd> and C<\DIFdel> commands, which provide a wrapper for these using them for the text but not for the link defining command (where any markup would cause errors). -=item C +=item C, C Redefine the commands recognised as citation commands. @@ -3944,12 +4080,26 @@ included sections are not expanded. The included files are assumed to be located in the same directories as the old and new master files, respectively, making it possible to organise files into old and new directories. --flatten is applied recursively, so inputted files can contain further -C<\input> statements. +C<\input> statements. Also handles files included by the import package +(C<\import> and C<\subimport>), and C<\subfile> command. Use of this option might result in prohibitive processing times for larger documents, and the resulting difference document no longer reflects the structure of the input documents. +=item B<--filter-script=filterscript> + +Run files through this filterscript (full path preferred) before processing. +The filterscript must take STDIN input and output to STDOUT. +When coupled with --flatten, each file will be run through the filter as it is brought in. + +=item B<--ignore-filter-stderr> + +When running with --filter-script, STDERR from the script may cause readability issues. +Turn this flag on to ignore STDERR from the filter script. + + + =item B<--help> or B<-h> @@ -4511,6 +4661,8 @@ min Pr sec sup +bibfield +bibinfo [Hclbkdruvt] [`'^"~=.] _ @@ -4537,6 +4689,7 @@ framebox footnote footnotetext framebox +href intertext part (sub){0,2}section\*? @@ -4550,6 +4703,7 @@ savebox sbox shortintertext shortstack +sidenote signature text.* value @@ -4694,7 +4848,7 @@ verbatim[*]? %DIF CULINECHBAR PREAMBLE \RequirePackage[normalem]{ulem} \RequirePackage[dvips]{changebar} -\RequirePackage{color} +\RequirePackage{color}\definecolor{RED}{rgb}{1,0,0}\definecolor{BLUE}{rgb}{0,0,1} \providecommand{\DIFadd}[1]{\protect\cbstart{\protect\color{blue}\uwave{#1}}\protect\cbend} \providecommand{\DIFdel}[1]{\protect\cbdelete{\protect\color{red}\sout{#1}}\protect\cbdelete} %DIF END CULINECHBAR PREAMBLE diff --git a/Master/texmf-dist/doc/support/latexdiff/latexdiff-fast b/Master/texmf-dist/doc/support/latexdiff/latexdiff-fast index 80ae57c7430..a6a7d9164d5 100755 --- a/Master/texmf-dist/doc/support/latexdiff/latexdiff-fast +++ b/Master/texmf-dist/doc/support/latexdiff/latexdiff-fast @@ -23,6 +23,25 @@ # Detailed usage information at the end of the file # + +# Version 1.3.1 +# Bug fixes: +# - remove some uninitialised variable $2 warnings in string substitution in flatten function in case included file is not found +# - add minimal postprocessing to diff processing of preamble commands (replace \RIGHTBRACE by \} ) +# - pre-processing: replace (contributed) routine take_comments_and_enter_from_frac() with take_comments_and_newline_from_frac(), which does the same thing +# (remove whitespace characters and comments between the argument of \frac commands) in an easier and more robust way. In addition, it +# will replace commands like \frac12 with \frac{1}{2} as pre-processing step. Fixes issue #184 +# - add "intertext" to list of unsafe math commands @UNSAFEMATHCMD . Fixes issue #179 +# - provide citation command patterns for biblatex and protect them with mbox'es. Fixes issue #199 +# - hardcode number of parameters for \href and \url commands to allow spaces between commands and arguments even if --allow-spaces option is not used (this +# is needed because some bibliography styles add such in-command-sequence spaces) Fixes issues: #178 #198 +# - bibitem is now kept even in deleted blocks such that deleted references show up properly (this implies that the actual numbers in numerical referencing schemes will change) +# (this is implemented by introducing a new class of commands KEEPCMD , which are kept as is in deleted environments (no effect in added environments). Currently +# \bibitem is hardwired to be the only member of this class (fixes issue #194, #174) +# Features: +# - add some special processing for revtex bibliography commands, so that the spaces between bibliography commands \bibfield and \bibinfo and their arguments are ignored. +# (fixes issue #194, should fix #174) +# # Version 1.3.0 (7 October 2018): # - treat options to \documentclass as potential package names (some packages allow implicit loading of or imply selected packages # - improved pattern matching: now allows nested angular brackets, and is no longer confused by escaped curly braces @@ -34,6 +53,9 @@ # - added to textcmds: \intertext # - new config variable CUSTOMDIFCMD to allow defining special versions of commands in added or deleted blocks (Pull request by github user jprotze) # - added option -no-links (mostly for use by latexdiff-vc in only-changes modes) (Pull request by github user jprotze) +# - new option --filter-script to run both input through a pre-processing script (PR jasonmccsmith #167) +# new option --no-filter-stderr to hide stderr output from filter-script (potentially dangerous, as this might hide malfunctioning of filter scripts) +# - --flatten now can deal with imports made using the import package {PR jasonmccsmith #173) # Bug fixes: # - pattern matching of \verb and \lstinline commands had an error which meant they would trigger on commands beginning with \verb. # - In description environments, mark up item descriptions by effectively reating the insides of item commannds as text commands (fixes #161) @@ -180,7 +202,7 @@ # (the distribution on which this modification is based is available # from http://search.cpan.org/~nedkonz/Algorithm-Diff-1.15 # the most recent version can be found via http://search.cpan.org/search?module=Algorithm::Diff ) -# Please note that the LICENCSE for Algorithm::Diff : +# Please note the LICENCE for Algorithm::Diff : # "Copyright (c) 2000-2002 Ned Konz. All rights reserved. # This program is free software; # you can redistribute it and/or modify it under the same terms @@ -734,11 +756,13 @@ use strict ; use warnings; use utf8 ; +use File::Spec ; + my ($algodiffversion)=split(/ /,$Algorithm::Diff::VERSION); my ($versionstring)=< undef, # COUNTERCMD textcmds which are associated with a counter # If any of these commands occur in a deleted block - # they will be succeeded by an \addtocounter{...}{-1} + # they will be followed by an \addtocounter{...}{-1} # for the associated counter such that the overall numbers # should be the same as in the new file LISTENV => undef , # list making environments - they will generally be kept @@ -808,7 +832,7 @@ my ($ARRENV, my $LABELCMD='(?:label)'; # matching commands are disabled within deleted blocks - mostly useful for maths mode, as otherwise it would be fine to just not add those to SAFECMDLIST -my @UNSAFEMATHCMD=('qedhere'); # Commands which are definitely unsafe for marking up in math mode (amsmath qedhere only tested to not work with UNDERLINE markup) (only affects WHOLE and COARSE math markup modes). Note that unlike text mode (or FINE math mode0 deleted unsafe commands are not deleted but simply taken outside \DIFdel +my @UNSAFEMATHCMD=('qedhere','intertext'); # Commands which are definitely unsafe for marking up in math mode (amsmath qedhere only tested to not work with UNDERLINE markup) (only affects WHOLE and COARSE math markup modes). Note that unlike text mode (or FINE math mode0 deleted unsafe commands are not deleted but simply taken outside \DIFdel my $MBOXINLINEMATH=0; # if set to 1 then surround marked-up inline maths expression with \mbox ( to get around compatibility # problems between some maths packages and ulem package @@ -863,6 +887,8 @@ our @SAFECMDEXCL=(); my @MBOXCMDLIST=(); # patterns for commands which are in principle safe but which need to be surrounded by an \mbox my @MBOXCMDEXCL=(); # all the patterns in MBOXCMDLIST will be appended to SAFECMDLIST +my @KEEPCMDLIST=( qr/^bibitem$/ ); # patterns for commands which should not be deleted in nominally delete text passages +my @KEEPCMDEXCL=(); my ($i,$j,$l); my ($old,$new); @@ -876,6 +902,7 @@ my ($oldpost, $newpost); my ($diffall); # Option names my ($type,$subtype,$floattype,$config,$preamblefile,$encoding,$nolabel,$visiblelabel, + $filterscript,$ignorefilterstderr, $showpreamble,$showsafe,$showtext,$showconfig,$showall, $replacesafe,$appendsafe,$excludesafe, $replacetext,$appendtext,$excludetext, @@ -977,6 +1004,8 @@ GetOptions('type|t=s' => \$type, 'ignore-warnings' => \$ignorewarnings, 'driver=s'=> \$driver, 'flatten' => \$flatten, + 'filter-script=s' => \$filterscript, + 'ignore-filter-stderr' => \$ignorefilterstderr, 'no-links' => \$nolinks, 'version' => \$version, 'help|h' => \$help, @@ -1021,7 +1050,8 @@ if (defined($mathmarkup)) { # else use numerical value } - +# Give filterscript a default empty string +$filterscript="" unless defined($filterscript); # setting extra preamble commands if (defined($preamblefile)) { @@ -1179,17 +1209,20 @@ if ( $mathmarkup == COARSE || $mathmarkup == WHOLE ) { + + foreach $pkg ( @packagelist ) { map { $packages{$_}="" } split(/,/,$pkg) ; } - if ($showconfig || $showtext || $showsafe || $showpreamble) { + show_configuration(); exit 0; } + if ( @ARGV != 2 ) { print STDERR "2 and only 2 non-option arguments required. Write latexdiff -h to get help\n"; exit(2); @@ -1248,6 +1281,17 @@ push(@SAFECMDLIST, qr/^QLEFTBRACE$/, qr/^QRIGHTBRACE$/); # word: sequence of letters or accents followed by letter my $word_ja='\p{Han}+|\p{InHiragana}+|\p{InKatakana}+'; my $word='(?:' . $word_ja . '|(?:(?:[-\w\d*]|\\\\[\"\'\`~^][A-Za-z\*])(?!(?:' . $word_ja . ')))+)'; + + # for selected commands, the number of arguments is known, and we can therefore allow spaces between command and its argument + # Note that it is still expected that the arguments are blocks marked by parentheses rather than single characters, and that intervening comments will inhibit the association + my $predefinedcmdoptseq01='\\\\(?:url|BibitemShut)\s*\s*(?:\{'. $pat_n . '\}\s*){1}'; # Commands with one non-optional argument + my $predefinedcmdoptseq12='\\\\(?:href|bibfield|bibinfo)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\}\s*){2}'; # Commands with one optional and two non-optional arguments +# my $predefinedcmdoptseq11='\\\\(?:bibitem)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\}\s*){1}'; # Commands with one optional and one non-optional arguments +# \bibitem in revtex styles appears to be always followed by \BibItemOpen. We bind \BibItemOpen to the bibitem (if present) in order to prevent the comparison algorithm to interpret the \BibItemOpen as an identical part of the sequence; this interpretation can lead to added and removed entries to the reference list to become mixed. + my $predefinedbibitem='\\\\(?:bibitem)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\})(?:%?\s*\\\\BibitemOpen)?'; # Commands with one optional and one non-optional arguments + + my $predefinedcmdoptseq='(?:'.$predefinedcmdoptseq12.'|'.$predefinedcmdoptseq01.'|'.$predefinedbibitem.')'; + my $cmdleftright='\\\\(?:left|right|[Bb]igg?[lrm]?|middle)\s*(?:[<>()\[\]|\.]|\\\\(?:[|{}]|\w+))'; my $cmdoptseq='\\\\[\w\d@\*]+'.$extraspace.'(?:(?:<'.$abrat0.'>|\['.$brat_n.'\]|\{'. $pat_n . '\}|\(' . $coords .'\))'.$extraspace.')*'; my $defseq='\\\\def\\\\[\w\d@\*]+(?:#\d+|\[#\d+\])+(?:\{'. $pat_n . '\})?'; @@ -1256,7 +1300,7 @@ push(@SAFECMDLIST, qr/^QLEFTBRACE$/, qr/^QRIGHTBRACE$/); my $math='\$(?:[^$]|\\\$)*?\$|\\\\[(](?:.|\n)*?\\\\[)]'; ## the current maths command cannot cope with newline within the math expression my $comment='%[^\n]*\n'; - my $pat=qr/(?:\A\s*)?(?:${and}|${quotemarks}|${number}|${word}|$quotedunderscore|${defseq}|$cmdleftright|${cmdoptseq}|${math}|${backslashnl}|${oneletcmd}|${comment}|${punct}|${mathpunct}|\{|\})\s*/ ; + my $pat=qr/(?:\A\s*)?(?:${and}|${quotemarks}|${number}|${word}|$quotedunderscore|${defseq}|$cmdleftright|${predefinedcmdoptseq}|${cmdoptseq}|${math}|${backslashnl}|${oneletcmd}|${comment}|${punct}|${mathpunct}|\{|\})\s*/ ; @@ -1307,9 +1351,56 @@ if (lc($encoding) eq "utf8" ) { binmode(STDERR, ":utf8"); } +# filter($text) +# Runs $text through the script provided in $filterscript argument, if set +# If not set, just returns $text unchanged. +# If flatten was set, defer filtering to flatten. flatten will run the filter +# on all incoming text prior to its own processing. +# If flatten was not set, filter each of old and new once (see just below this def) +sub filter { + my ($text)=@_; + my ($textout,$pid); + if ($filterscript ne "") { + print STDERR "Passing " . length($text) . " chars to filter script " . $filterscript . "\n" if $verbose; + + if ($ignorefilterstderr) { + # If we need to capture and bury STDERR, use the Open3 version, and close CHLD_ERR below. + use IPC::Open3; + # We consume STDERR from the process, and hide it + $pid = open3(\*CHLD_IN, \*CHLD_OUT, \*CHLD_ERR, $filterscript) or die "open3() failed $!"; + } + else { + # Capture STDOUT and use as our new $text. Allow STDERR to go to console. + use IPC::Open2; + $pid = open2(\*CHLD_OUT, \*CHLD_IN, $filterscript) or die "open2() failed $!"; + } + # Send in $text + print CHLD_IN $text."\n"; # Adding a newline just to make sure there is one. + close CHLD_IN; + # Wait for output and gather it up + while () { + $textout = $textout.$_; + } + if ($ignorefilterstderr) { + close CHLD_ERR; # Enable only if Open3 used above + } + # On the off chance a very long running and/or frequently called script is used. + waitpid( $pid, 0 ); + $text = $textout; + print STDERR "Received " . length($text) . " chars after filtering\n" if $verbose; + print STDERR $text if $verbose; + } + return $text; +} + $old=read_file_with_encoding($oldfile,$encoding); $new=read_file_with_encoding($newfile,$encoding); +if (not defined($flatten)) { + $old=filter($old); + $new=filter($new); +} + @@ -1321,11 +1412,11 @@ exetime(1); if ($flatten) { - $oldbody=flatten($oldbody,$oldpreamble,$oldfile,$encoding); - $newbody=flatten($newbody,$newpreamble,$newfile,$encoding); + $oldbody=flatten($oldbody,$oldpreamble,File::Spec->rel2abs($oldfile),$encoding); + $newbody=flatten($newbody,$newpreamble,File::Spec->rel2abs($newfile),$encoding); # flatten preamble - $oldpreamble=flatten($oldpreamble,$oldpreamble,$oldfile,$encoding); - $newpreamble=flatten($newpreamble,$newpreamble,$newfile,$encoding); + $oldpreamble=flatten($oldpreamble,$oldpreamble,File::Spec->rel2abs($oldfile),$encoding); + $newpreamble=flatten($newpreamble,$newpreamble,File::Spec->rel2abs($newfile),$encoding); } @@ -1580,6 +1671,9 @@ my ( $citpat); if ( defined $packages{"apacite"} ) { print STDERR "apacite package detected.\n" if $verbose ; $citpat='(?:mask)?(?:full|short|no)?cite(?:A|author|year|meta)?(?:NP)?'; +} elsif (defined $packages{"biblatex"}) { + print STDERR "biblatex package detected.\n" if $verbose ; + $citpat='(?:[cC]ites?|(?:[pP]aren|foot|[Tt]ext|[sS]mart|super)cites?\*?|footnotecitetex)'; } else { # citation command pattern for all other citation schemes $citpat='(?:cite\w*|nocite)'; @@ -1902,11 +1996,13 @@ sub remove_endinput { # encoding is the encoding sub flatten { my ($text,$preamble,$filename,$encoding)=@_; - my ($includeonly,$dirname,$fname,$newpage,$fullfile,$filecontent,$replacement,$begline,$bblfile,$subfile,$command,$verbenv,$verboptions); + my ($includeonly,$dirname,$fname,$newpage,$fullfile,$filecontent,$replacement,$begline,$inputcmd,$bblfile,$subfile,$command,$verbenv,$verboptions,$ignore,$fileonly); my ($subpreamble,$subbody,$subpost); + my ($subdir,$subdirfull,$importfilepath); require File::Basename ; - require File::Spec ; - $dirname = File::Basename::dirname($filename); + ### require File::Spec ; # now this is needed even if flatten option not given + $filename = File::Spec->rel2abs( $filename ) ; + ($ignore, $dirname, $fileonly) = File::Spec->splitpath($filename) ; $bblfile = $filename; $bblfile=~s/\.tex$//; $bblfile.=".bbl"; @@ -1919,11 +2015,51 @@ sub flatten { print STDERR "DEBUG: includeonly $includeonly\n" if $debug; + + # Run through filter, to let filterscript have a pass if it was set + $text = filter($text); + + # Recursively replace \\import and \\subimport files + $text =~ s/(^(?:[^%\n]|\\%)*)(\\subimport\{(.*?)\}|\\import\{(.*?)\})(?:[\s]*)\{(.*?)\}/{ + # $1 is begline + # $3 is directory if subimport + # $4 is directory if import + # $5 is filename + $begline = (defined($1)? $1 : ""); + $subdir = $3 if defined($3); + $subdir = $4 if defined($4); + $fname = $5; + $fname .= ".tex" unless $fname =~ m|\.\w{3,4}$|; + print STDERR "DEBUG begline:", $begline, "\n" if $debug; + print STDERR "DEBUG", (defined($3)? "subimport_file:" : "import_file:"), $subdir, "\n" if $debug; + print STDERR "DEBUG file:", $fname, "\n" if $debug; + + # subimport appends $subdir to the current $dirname. import replaces it with an absolute path. + $subdirfull = (defined($3) ? File::Spec->catdir($dirname,$subdir) : $subdir); + + $importfilepath = File::Spec->catfile($subdirfull, $fname); + + print STDERR "importing importfilepath:", $importfilepath,"\n" if $verbose; + if ( -f $importfilepath ) { + # If file exists, replace input or include command with expanded input + $replacement=flatten(read_file_with_encoding($importfilepath, $encoding), $preamble,$importfilepath,$encoding) or die "Could not open file ",$fullfile,": $!"; + } else { + # if file does not exist, do not expand include or input command (do not warn if fname contains #[0-9] as it is then likely part of a command definition + # and is not meant to be expanded directly + print STDERR "WARNING: Could not find included file ",$importfilepath,". I will continue but not expand |$2|\n"; + $replacement=(defined($3)? "\\subimport" : "\\import"); + $replacement .= "{$subdir}{$fname} % Processed"; + } + "$begline$replacement"; + }/exgm; + # recursively replace \\input and \\include files $text =~ s/(^(?:[^%\n]|\\%)*)(\\input\{(.*?)\}|\\include\{(${includeonly}(?:\.tex)?)\})/{ $begline=(defined($1)? $1 : "") ; + $inputcmd=$2; $fname = $3 if defined($3) ; $fname = $4 if defined($4) ; + $newpage=(defined($4)? " \\newpage " : "") ; # # add tex extension unless there is a three or four letter extension already $fname .= ".tex" unless $fname =~ m|\.\w{3,4}$|; $fullfile = File::Spec->catfile($dirname,$fname); @@ -1936,12 +2072,11 @@ sub flatten { $replacement=flatten(read_file_with_encoding($fullfile, $encoding), $preamble,$filename,$encoding) or die "Could not open file ",$fullfile,": $!"; $replacement = remove_endinput($replacement); # \include always starts a new page; use explicit \newpage command to simulate this - $newpage=(defined($4)? " \\newpage " : "") ; } else { # if file does not exist, do not expand include or input command (do not warn if fname contains #[0-9] as it is then likely part of a command definition # and is not meant to be expanded directly - print STDERR "WARNING: Could not find included file ",$fullfile,". I will continue but not expand |$2|\n" unless $fname =~ m(#[0-9]) ; - $replacement = $2 ; # i.e. just the original command again -> make no change file does not exist + print STDERR "WARNING: Could not find included file ",$fullfile,". I will continue but not expand |$inputcmd|\n" unless $fname =~ m(#[0-9])n ; + $replacement = $inputcmd ; # i.e. just the original command again -> make no change file does not exist $newpage=""; } "$begline$newpage$replacement$newpage"; @@ -2220,6 +2355,7 @@ sub splitlatex { } } + unshift(@retval,$leadin) if (length($leadin)>0); return @retval; } @@ -2401,10 +2537,14 @@ sub extracttextblocks { my $wpat=qr/^(?:[a-zA-Z.,'`:;?()!]*)[\s~]*$/; #' my $retval=[]; + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace='\s*'; + for ($i=0;$i< scalar @$block;$i++) { ($token,$index)=@{ $block->[$i] }; # store pure text blocks - if ($token =~ /$wpat/ || ( $token =~/^\\((?:[`'^"~=.]|[\w\d@\*]+))((?:${extraspace}\[$brat_n\]${extraspace}|${extraspace}\{$pat_n\})*)/o + if ($token =~ /$wpat/ || ( $token =~/^\\((?:[`'^"~=.]|[\w\d@\*]+))((?:${extraspace}\[$brat_n\]${extraspace}|${extraspace}\{$pat_n\})*)/ && iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL) && !iscmd($1,\@TEXTCMDLIST,\@TEXTCMDEXCL))) { # we have text or a command which can be treated as text @@ -2449,6 +2589,10 @@ sub extractcommands { my ($i,$token,$index,$cmd,$open,$mid,$closing); my $retval=[]; + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace='\s*'; + for ($i=0;$i< scalar @$block;$i++) { ($token,$index)=@{ $block->[$i] }; # check if token is an alphanumeric command sequence with at least one non-optional argument @@ -2460,6 +2604,7 @@ sub extractcommands { # $4: } + trailing spaces if ( ( $token =~ m/^(\\([\w\d\*]+)(?:${extraspace}\[$brat_n\]|${extraspace}\{$pat_n\})*${extraspace}\{)($pat_n)(\}\s*)$/so ) && iscmd($2,\@TEXTCMDLIST,\@TEXTCMDEXCL) ) { + print STDERR "DEBUG EXTRACTCOMMANDS Match |$1|$2|$3|$4|$index \n" if $debug; # push(@$retval,[ $2,$index,$1,$3,$4 ]); ($cmd,$open,$mid,$closing) = ($2,$1,$3,$4) ; $closing =~ s/\}/\\RIGHTBRACE/ ; @@ -2475,20 +2620,20 @@ sub extractcommands { sub iscmd { my ($cmd,$regexar,$regexexcl)=@_; my ($ret)=0; - print STDERR "DEBUG: iscmd($cmd)=" if $debug; + ### print STDERR "DEBUG: iscmd($cmd)=" if $debug; foreach $pat ( @$regexar ) { if ( $cmd =~ m/^${pat}$/ ) { $ret=1 ; last; } } - print STDERR "0\n" if ($debug && !$ret) ; + ### print STDERR "0\n" if ($debug && !$ret) ; return 0 unless $ret; foreach $pat ( @$regexexcl ) { - print STDERR "0\n" if ( $debug && $cmd =~ m/^${pat}$/) ; + ### print STDERR "0\n" if ( $debug && $cmd =~ m/^${pat}$/) ; return 0 if ( $cmd =~ m/^${pat}$/ ); } - print STDERR "1\n" if $debug; + ### print STDERR "1\n" if $debug; return 1; } @@ -2547,7 +2692,6 @@ sub pass2 { print STDERR " $deltokcnt discarded tokens in $delblkcnt blocks.\n"; print STDERR " $addtokcnt appended tokens in $addblkcnt blocks.\n"; } - return(@$retval); } @@ -2570,13 +2714,20 @@ sub marktags { my $cmdcomment= ($opencmd =~ m/^%/); # Flag to indicate whether opencmd is a comment (i.e. if we intend to simply comment out changed commands) my ($command,$commandword,$closingbracket) ; # temporary variables needed below to remember sub-pattern matches -# split this block to flatten out sequences joined in pass1 + + +# split this block to split sequences joined in pass1 @$block=splitlatex(join "",@$block); ### print STDERR "DEBUG: marktags $openmark,$closemark,$open,$close,$opencmd,$closecmd,$comment\n" if $debug; ### print STDERR "DEBUG: marktags blocksplit ",join("|",@$block),"\n" if $debug; + + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace_mt='\s*'; + + foreach (@$block) { $word=$_; - ### print STDERR "DEBUG MARKTAGS: |$word|\n" if $debug; if ( $word =~ s/^%/%$comment/ ) { # a comment if ($cmd==1) { @@ -2601,9 +2752,7 @@ sub marktags { } # negative lookahead pattern (?!) in second clause is put in to avoid matching \( .. \) patterns # also note that second pattern will match \\ - print STDERR "DEBUG marktags: Considering word |$word|\n" if $debug; - if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\(?!\()(\\|[`'^"~=.]|[\w*@]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { - ###print STDERR "DEBUG MARKTAGS is a non-safe command ($1)\n" if $debug; + if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\(?!\()(\\|[`'^"~=.]|[\w*@]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { ### if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\([\w*@\\% ]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { # word is a command or other significant token (not in SAFECMDLIST) ## same conditions as in subroutine extractcommand: @@ -2615,13 +2764,13 @@ sub marktags { # $3: last argument # $4: } + trailing spaces ### pre-0.3 if ( ( $token =~ m/^(\\([\w\d\*]+)(?:\[$brat0\]|\{$pat_n\})*\{)($pat_n)(\}\s*)$/so ) - if ( ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace}\[$brat_n\]|${extraspace}\{$pat_n\})*${extraspace}\{)($pat_n)(\}\s*)$/so ) + if ( ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace_mt}\[$brat_n\]|${extraspace_mt}\{$pat_n\})*${extraspace_mt}\{)($pat_n)(\}\s*)$/so ) && (iscmd($2,\@TEXTCMDLIST,\@TEXTCMDEXCL)|| iscmd($2,\@MATHTEXTCMDLIST,\@MATHTEXTCMDEXCL)) && ( !$cmdcomment || !iscmd($2,\@CONTEXT2CMDLIST, \@CONTEXT2CMDEXCL) ) ) { # Condition 1: word is a command? - if yes, $1,$2,.. will be set as above # Condition 2: word is a text command - we mark up the interior of the word. There is a separate check for MATHTEXTCMDLIST # because for $mathmarkup=WHOLE, the commands should not be split in pass1 (ie. math mode commands are not in - # TEXTCMDLIST, but the interior of MATHTEXT commands should be highlighted in both deleted and added blocks + # TEXTCMDLIST, but the interior of MATHTEXT commnds should be highlighted in both deleted and added blocks # Condition 3: But if we are in a deleted block ($cmdcomment=1) and # $2 (the command) is in context2, just treat it as an ordinary command (i.e. comment it open with $opencmd) # Because we do not want to disable this command @@ -2657,6 +2806,13 @@ sub marktags { } push (@$retval,$AUXCMD,"\n") if $cmdcomment ; $cmd=-1 ; + } elsif ( $cmdcomment && + ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace_mt}\[$brat_n\]|${extraspace_mt}\{$pat_n\})*${extraspace_mt}\{)($pat_n)(\}\s*)/so ) + && iscmd($2,\@KEEPCMDLIST, \@KEEPCMDEXCL) ) { + # 'keepcmd' in a deleted environment: keep the command as is + push (@$retval,$close) if $cmd==0 ; + push (@$retval,$word); + $cmd=-1; # pretend we are at the beginning of a sequence because we do not want to add an additional $closecmd or $close before the next token, no matter what it is } else { # ordinary command push (@$retval,$opencmd) if $cmd==-1 ; @@ -2693,52 +2849,19 @@ sub marktags { } #used in preprocess -sub take_comments_and_enter_from_frac() { - #*************take the \n and % between frac and {}*********** - #notice all of the substitution are made none global - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\frac(([\s]*%[^\n]*?)*[\r\n|\r|\n])+\{(.*?)\\end\{\1}/s ) { - # if there isn't any % or \n in the pattern $2 then there should be an \\end{...} in $2 - ### print STDERR "Match the following in take_comments and_enter_from_frac(1):\n****$&****\n" if $debug; - if( $2 !~ m/\\end\{$1}/s ) { - # take out % and \n from the next match only (none global) - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\frac(([\s]*%[^\n]*?)*[\r\n|\r|\n])+\{(.*?)\\end\{\1}/\\begin{$1}$2\\frac{$5\\end{$1}/s; - } - else{ - #there are no more % and \n in $2, we want to find the next one so we clear the begin-end from the pattern - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\end\{\1}/MATHBLOCK$1\{$2\}MATHBLOCKEND/s; - } - } - ###cleaning up - while( s/MATHBLOCK($MATHENV|$MATHARRENV|SQUAREBRACKET)\{(.*?)\}MATHBLOCKEND/\\begin{$1}$2\\end{$1}/s ){} - ###*************take the \n and % between frac and {}*********** - - ###**********take the \n and % between {} and {} of the frac*************** - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\frac\{(.*?)\\end\{\1\}/s ) { - # if there isn't any more //frac before the first //end in the pattern $2 then there should be an \\end{...} in $2 - ###print STDERR "Match the following in take_comments and_enter_from_frac(2):\n****$&****\n" if $debug; - if( $2 !~ m/\\end\{$1\}/s ) { - # from now on CURRFRAC is the frac we are looking at - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\frac\{(.*?)\\end\{\1\}/\\begin\{$1\}$2CURRFRAC\{$3\\end\{$1\}/s; - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{(.*?)\\end\{\1\}/s ) { - if( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{($pat_n)\}([\s]*(%[^\n]*?)*[\r\n|\r|\n])+[\s]*\{(.*?)\\end\{\1}/s ) { - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{($pat_n)\}([\s]*(%[^\n]*?)*[\r\n|\r|\n])+[\s]*\{(.*?)\\end\{\1\}/\\begin\{$1\}$2CURRFRAC\{$3\}\{$6\\end\{$1\}/s; - } - else { # there is no comment or \n between the two brackets {}{} - # change CURRFRAC to FRACSTART so we can change them all back to //frac{ when we finish - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)CURRFRAC\{(.*?)\\end\{\1}/\\begin{$1}$2FRACSTART\{$3\\end{$1}/s; - } - } - } - else{ - ###there are no more frac in $2, we want to find the next one so we clear the begin-end from the pattern - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\end\{\1}/MATHBLOCK$1\{$2\}MATHBLOCKEND/s; - } - - } - ###cleaning up - while( s/MATHBLOCK($MATHENV|$MATHARRENV|SQUAREBRACKET)\{(.*?)\}MATHBLOCKEND/\\begin{$1}$2\\end{$1}/s ){} - s/FRACSTART/\\frac/g; - ###***************take the \n and % between {} and {} of the frac********************* +sub take_comments_and_newline_from_frac() { + # some special magic for common usage of frac, which does not conform to the latexdiff requirements but can be made to fit + # note that this is a rare exception to the general rule that the new tex can be reconstructed from the diff file + + # \frac12 -> \frac{1}{2} + s/\\frac(\d)(\w)/\\frac\{$1\}\{$2\}/g; + + # \frac1{2b} -> \frac{1}{2b} + s/\\frac(\d)/\\frac\{$1\}/g; + + # delete space and comment characters between \frac arguments +# s/\\frac(?:\s*?%[^\n]*?)*?(\{$pat_n\})\s*(\{$pat_n\})/\\frac$1$2/g; + s/\\frac(?:\s|%[^\n]*?)*(\{$pat_n\})(?:\s|%[^\n]*?)*(\{$pat_n\})/\\frac$1$2/g; } # preprocess($string, ..) @@ -2821,7 +2944,7 @@ sub preprocess { # resort to WHOLE or FINE, or NONE math mode processing. s/\\begin\{($ARRENV)}(.*?)\\end\{\1}/\\ARRAYBLOCK$1\{$2\\DIFANCHORARRB \}\\DIFANCHORARRE /sg; - take_comments_and_enter_from_frac(); + take_comments_and_newline_from_frac(); s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\end\{\1\}/\\MATHBLOCK$1\{$2\\DIFANCHORMATHB \}\\DIFANCHORMATHE /sg; } @@ -3439,8 +3562,7 @@ sub restore_item_commands { # pre-process preamble by looking for commands used in \maketitle (title, author, date etc commands) # the list of commands is defined in CONTEXT2CMD # if found then use a bodydiff to mark up content, and replace the corresponding commands - # in both preambles by marked up version to 'fool' the linediff (such that only body is marked - # up. + # in both preambles by marked up version to 'fool' the linediff (such that only body is marked up). # A special case are e.g. author commands being added (or removed) # 1. If commands are added, then the entire content is marked up as new, but also the lines are marked as new in the linediff # 2. If commands are removed, then the linediff will mark the line as deleted. The program returns @@ -3518,6 +3640,9 @@ EOF } $argnew=$newhash{$cmd}->[2]; $argdiff="{" . join("",bodydiff($argold,$argnew)) ."}"; + # Replace \RIGHTBRACE by } + $argdiff =~ s/\\RIGHTBRACE/}/g; + if ( length $optargnew ) { $optargdiff="[".join("",bodydiff($optargold,$optargnew))."]" ; $optargdiff =~ s/\\DIFaddbegin /\\DIFaddbeginFL /g; @@ -3863,6 +3988,7 @@ be real files (not pipes or similar) as they are opened twice. reason package specific parsing needs to be switched off, use --packages=none. The following packages trigger special behaviour: endfloat hyperref amsmath apacite siunitx cleveref glossaries mhchem chemformula/chemmacros + biblatex [ Default: scan the preamble for \\usepackage commands to determine loaded packages.] @@ -3944,7 +4070,17 @@ Miscelleneous options be located in the same directories as the old and new master files, respectively, making it possible to organise files into old and new directories. --flatten is applied recursively, so inputted files can contain further - \\input statements. + \\input statements. Also handles files included by the import package + (\\import and \\subimport), and \\subfile command. + +--filter-script=filterscript Run files through this filterscript (full path preferred) before processing. + The filterscript must take STDIN input and output to STDOUT. + When coupled with --flatten, each file will be run through the filter as it is brought in. + +--ignore-filter-stderr When running with --filter-script, STDERR from the script may cause readability issues. + Turn this flag on to ignore STDERR from the filter script. + + --help -h Show this help text. @@ -4199,7 +4335,7 @@ define new C<\DIFadd> and C<\DIFdel> commands, which provide a wrapper for these using them for the text but not for the link defining command (where any markup would cause errors). -=item C +=item C, C Redefine the commands recognised as citation commands. @@ -4505,12 +4641,26 @@ included sections are not expanded. The included files are assumed to be located in the same directories as the old and new master files, respectively, making it possible to organise files into old and new directories. --flatten is applied recursively, so inputted files can contain further -C<\input> statements. +C<\input> statements. Also handles files included by the import package +(C<\import> and C<\subimport>), and C<\subfile> command. Use of this option might result in prohibitive processing times for larger documents, and the resulting difference document no longer reflects the structure of the input documents. +=item B<--filter-script=filterscript> + +Run files through this filterscript (full path preferred) before processing. +The filterscript must take STDIN input and output to STDOUT. +When coupled with --flatten, each file will be run through the filter as it is brought in. + +=item B<--ignore-filter-stderr> + +When running with --filter-script, STDERR from the script may cause readability issues. +Turn this flag on to ignore STDERR from the filter script. + + + =item B<--help> or B<-h> @@ -5072,6 +5222,8 @@ min Pr sec sup +bibfield +bibinfo [Hclbkdruvt] [`'^"~=.] _ @@ -5098,6 +5250,7 @@ framebox footnote footnotetext framebox +href intertext part (sub){0,2}section\*? @@ -5111,6 +5264,7 @@ savebox sbox shortintertext shortstack +sidenote signature text.* value @@ -5255,7 +5409,7 @@ verbatim[*]? %DIF CULINECHBAR PREAMBLE \RequirePackage[normalem]{ulem} \RequirePackage[dvips]{changebar} -\RequirePackage{color} +\RequirePackage{color}\definecolor{RED}{rgb}{1,0,0}\definecolor{BLUE}{rgb}{0,0,1} \providecommand{\DIFadd}[1]{\protect\cbstart{\protect\color{blue}\uwave{#1}}\protect\cbend} \providecommand{\DIFdel}[1]{\protect\cbdelete{\protect\color{red}\sout{#1}}\protect\cbdelete} %DIF END CULINECHBAR PREAMBLE diff --git a/Master/texmf-dist/scripts/latexdiff/latexdiff-vc.pl b/Master/texmf-dist/scripts/latexdiff/latexdiff-vc.pl index 21bf1f3adf3..9843238c457 100755 --- a/Master/texmf-dist/scripts/latexdiff/latexdiff-vc.pl +++ b/Master/texmf-dist/scripts/latexdiff/latexdiff-vc.pl @@ -26,18 +26,17 @@ # Detailed usage information at the end of the file # # TODO/IDEAS: - option to call external pre-processing codes -# - choose type of latex processor / bibtex (luatex, xelatex etc) # version 1.3.0 ((7 October 2018) # - option --only-changes with hyperref will suppress hyperrefs (pull request jprotze)_ # - option --only-changes now moves (rather than copies) file with only changes # version 1.2.1 (22 June 2017): -# - update: use qpdf (instead of pdftk) to select particular pages of output (pull request #102 submited by Tom Scogland via github). This was necessary because pdftk is being deprecated. pdftk is still used as a fall-back +# - update: use qpdf (instead of pdftk) to select particular pages of output (pull request #102 submitted by Tom Scogland via github). This was necessary because pdftk is being deprecated. pdftk is still used as a fall-back # - bug fix: --hg option was not recognised (partially fixes github issue #93 ) # - wrap passed-through options to latexdiff in quotation marks (fix github issue #58 ) # - program names for latexdiff, latex, dvips, bibtex configurable (fixes issue #40) # # version 1.2.0: -# - depracation fix: left brace in RegEx now needs to be escaped +# - deprecation fix: left brace in RegEx now needs to be escaped # # version 1.1.1: # - better detection of RCS system @@ -72,8 +71,8 @@ use strict ; use warnings ; my $versionstring=< -r F ... -compares F with the most recent version checked into RCS. +compares F with the most recently checked-in version checked. =item B -r F F ... @@ -685,7 +684,7 @@ compares F with revision F. compares revisions F and F of F. Multiple files can be specified for all of the above options. All files must have the -extension C<.tex>, though. +extensions C<.tex>, C<.bbl>, or C<.flt>, though. =item B F F @@ -701,7 +700,7 @@ reported to stdout. Rather than appending the string C and optionally the version numbers given to the output-file, this will prepend a directory name C to the -original filename, creating the directory and subdirectories should they not exist already. This is particularly useful in order to clone a +original filename, creating the directory and sub-directories should they not exist already. This is particularly useful in order to clone a complete directory hierarchy. Optionally, a pathname F can be specified, which is prepended instead of C. =item B<--flatten,--flatten=keep-intermediate> @@ -713,9 +712,6 @@ The generic usage of this function is : C, the intermediate revision snapshots are kept in the current directory (Default is to store them in a temporary directory and delete them after generating the diff file.) -=item B<--config var1=val1,var2=val2,...> or B<-c var1=val1,..> - - =item B<--only-changes> @@ -738,6 +734,7 @@ run latex command on diff file after generation of diff file. run latex and dvixxx commands after generation of diff file. =item B<-c configfile> +=item B<--config var1=val1,var2=val2,...> or B<-c var1=val1,..> Set configuration variables for latexdiff and latexdiff-vc. The option can be repeated to set different variables (as an alternative to the comma-separated list). @@ -755,7 +752,7 @@ Available variables for latexdiff-vc: =back -All other config variables are passed to latexdiff. Explicity set configuration changes always override implicit +All other config variables are passed to latexdiff. Explicitly set configuration changes always override implicit changes by the following shortcut options B<--fast>, B<--so>, B<--ps> and B<--pdf>. =item B<--fast> or B<--so> @@ -803,7 +800,7 @@ L =head1 PORTABILITY I uses external commands and is therefore dependent on the system architecture; it has been -tested mainly on Unix-like systems. It also requires the a version control +tested mainly on Unix-like systems. It also requires a version control system and latex to be installed on the system to make use of all features. Modules from Perl 5.8 or higher are required. diff --git a/Master/texmf-dist/scripts/latexdiff/latexdiff.pl b/Master/texmf-dist/scripts/latexdiff/latexdiff.pl index 6d07c9a88f8..81387135c58 100755 --- a/Master/texmf-dist/scripts/latexdiff/latexdiff.pl +++ b/Master/texmf-dist/scripts/latexdiff/latexdiff.pl @@ -23,6 +23,25 @@ # Detailed usage information at the end of the file # + +# Version 1.3.1 +# Bug fixes: +# - remove some uninitialised variable $2 warnings in string substitution in flatten function in case included file is not found +# - add minimal postprocessing to diff processing of preamble commands (replace \RIGHTBRACE by \} ) +# - pre-processing: replace (contributed) routine take_comments_and_enter_from_frac() with take_comments_and_newline_from_frac(), which does the same thing +# (remove whitespace characters and comments between the argument of \frac commands) in an easier and more robust way. In addition, it +# will replace commands like \frac12 with \frac{1}{2} as pre-processing step. Fixes issue #184 +# - add "intertext" to list of unsafe math commands @UNSAFEMATHCMD . Fixes issue #179 +# - provide citation command patterns for biblatex and protect them with mbox'es. Fixes issue #199 +# - hardcode number of parameters for \href and \url commands to allow spaces between commands and arguments even if --allow-spaces option is not used (this +# is needed because some bibliography styles add such in-command-sequence spaces) Fixes issues: #178 #198 +# - bibitem is now kept even in deleted blocks such that deleted references show up properly (this implies that the actual numbers in numerical referencing schemes will change) +# (this is implemented by introducing a new class of commands KEEPCMD , which are kept as is in deleted environments (no effect in added environments). Currently +# \bibitem is hardwired to be the only member of this class (fixes issue #194, #174) +# Features: +# - add some special processing for revtex bibliography commands, so that the spaces between bibliography commands \bibfield and \bibinfo and their arguments are ignored. +# (fixes issue #194, should fix #174) +# # Version 1.3.0 (7 October 2018): # - treat options to \documentclass as potential package names (some packages allow implicit loading of or imply selected packages # - improved pattern matching: now allows nested angular brackets, and is no longer confused by escaped curly braces @@ -34,6 +53,9 @@ # - added to textcmds: \intertext # - new config variable CUSTOMDIFCMD to allow defining special versions of commands in added or deleted blocks (Pull request by github user jprotze) # - added option -no-links (mostly for use by latexdiff-vc in only-changes modes) (Pull request by github user jprotze) +# - new option --filter-script to run both input through a pre-processing script (PR jasonmccsmith #167) +# new option --no-filter-stderr to hide stderr output from filter-script (potentially dangerous, as this might hide malfunctioning of filter scripts) +# - --flatten now can deal with imports made using the import package {PR jasonmccsmith #173) # Bug fixes: # - pattern matching of \verb and \lstinline commands had an error which meant they would trigger on commands beginning with \verb. # - In description environments, mark up item descriptions by effectively reating the insides of item commannds as text commands (fixes #161) @@ -176,7 +198,7 @@ # package and use the standard latexdiff version instead # (current distribution available from http://search.cpan.org/~nedkonz/Algorithm-Diff-1.15 # the most recent version can be found via http://search.cpan.org/search?module=Algorithm::Diff ) -# Please note that the LICENCSE for Algorithm::Diff : +# Please note the LICENCE for Algorithm::Diff : # "Copyright (c) 2000-2002 Ned Konz. All rights reserved. # This program is free software; # you can redistribute it and/or modify it under the same terms @@ -630,11 +652,13 @@ use strict ; use warnings; use utf8 ; +use File::Spec ; + my ($algodiffversion)=split(/ /,$Algorithm::Diff::VERSION); my ($versionstring)=< undef, # COUNTERCMD textcmds which are associated with a counter # If any of these commands occur in a deleted block - # they will be succeeded by an \addtocounter{...}{-1} + # they will be followed by an \addtocounter{...}{-1} # for the associated counter such that the overall numbers # should be the same as in the new file LISTENV => undef , # list making environments - they will generally be kept @@ -704,7 +728,7 @@ my ($ARRENV, my $LABELCMD='(?:label)'; # matching commands are disabled within deleted blocks - mostly useful for maths mode, as otherwise it would be fine to just not add those to SAFECMDLIST -my @UNSAFEMATHCMD=('qedhere'); # Commands which are definitely unsafe for marking up in math mode (amsmath qedhere only tested to not work with UNDERLINE markup) (only affects WHOLE and COARSE math markup modes). Note that unlike text mode (or FINE math mode0 deleted unsafe commands are not deleted but simply taken outside \DIFdel +my @UNSAFEMATHCMD=('qedhere','intertext'); # Commands which are definitely unsafe for marking up in math mode (amsmath qedhere only tested to not work with UNDERLINE markup) (only affects WHOLE and COARSE math markup modes). Note that unlike text mode (or FINE math mode0 deleted unsafe commands are not deleted but simply taken outside \DIFdel my $MBOXINLINEMATH=0; # if set to 1 then surround marked-up inline maths expression with \mbox ( to get around compatibility # problems between some maths packages and ulem package @@ -759,6 +783,8 @@ our @SAFECMDEXCL=(); my @MBOXCMDLIST=(); # patterns for commands which are in principle safe but which need to be surrounded by an \mbox my @MBOXCMDEXCL=(); # all the patterns in MBOXCMDLIST will be appended to SAFECMDLIST +my @KEEPCMDLIST=( qr/^bibitem$/ ); # patterns for commands which should not be deleted in nominally delete text passages +my @KEEPCMDEXCL=(); my ($i,$j,$l); my ($old,$new); @@ -772,6 +798,7 @@ my ($oldpost, $newpost); my ($diffall); # Option names my ($type,$subtype,$floattype,$config,$preamblefile,$encoding,$nolabel,$visiblelabel, + $filterscript,$ignorefilterstderr, $showpreamble,$showsafe,$showtext,$showconfig,$showall, $replacesafe,$appendsafe,$excludesafe, $replacetext,$appendtext,$excludetext, @@ -873,6 +900,8 @@ GetOptions('type|t=s' => \$type, 'ignore-warnings' => \$ignorewarnings, 'driver=s'=> \$driver, 'flatten' => \$flatten, + 'filter-script=s' => \$filterscript, + 'ignore-filter-stderr' => \$ignorefilterstderr, 'no-links' => \$nolinks, 'version' => \$version, 'help|h' => \$help, @@ -917,7 +946,8 @@ if (defined($mathmarkup)) { # else use numerical value } - +# Give filterscript a default empty string +$filterscript="" unless defined($filterscript); # setting extra preamble commands if (defined($preamblefile)) { @@ -1075,17 +1105,20 @@ if ( $mathmarkup == COARSE || $mathmarkup == WHOLE ) { + + foreach $pkg ( @packagelist ) { map { $packages{$_}="" } split(/,/,$pkg) ; } - if ($showconfig || $showtext || $showsafe || $showpreamble) { + show_configuration(); exit 0; } + if ( @ARGV != 2 ) { print STDERR "2 and only 2 non-option arguments required. Write latexdiff -h to get help\n"; exit(2); @@ -1144,6 +1177,17 @@ push(@SAFECMDLIST, qr/^QLEFTBRACE$/, qr/^QRIGHTBRACE$/); # word: sequence of letters or accents followed by letter my $word_ja='\p{Han}+|\p{InHiragana}+|\p{InKatakana}+'; my $word='(?:' . $word_ja . '|(?:(?:[-\w\d*]|\\\\[\"\'\`~^][A-Za-z\*])(?!(?:' . $word_ja . ')))+)'; + + # for selected commands, the number of arguments is known, and we can therefore allow spaces between command and its argument + # Note that it is still expected that the arguments are blocks marked by parentheses rather than single characters, and that intervening comments will inhibit the association + my $predefinedcmdoptseq01='\\\\(?:url|BibitemShut)\s*\s*(?:\{'. $pat_n . '\}\s*){1}'; # Commands with one non-optional argument + my $predefinedcmdoptseq12='\\\\(?:href|bibfield|bibinfo)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\}\s*){2}'; # Commands with one optional and two non-optional arguments +# my $predefinedcmdoptseq11='\\\\(?:bibitem)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\}\s*){1}'; # Commands with one optional and one non-optional arguments +# \bibitem in revtex styles appears to be always followed by \BibItemOpen. We bind \BibItemOpen to the bibitem (if present) in order to prevent the comparison algorithm to interpret the \BibItemOpen as an identical part of the sequence; this interpretation can lead to added and removed entries to the reference list to become mixed. + my $predefinedbibitem='\\\\(?:bibitem)\s*(?:\['.$brat_n.'\])?\s*(?:\{'. $pat_n . '\})(?:%?\s*\\\\BibitemOpen)?'; # Commands with one optional and one non-optional arguments + + my $predefinedcmdoptseq='(?:'.$predefinedcmdoptseq12.'|'.$predefinedcmdoptseq01.'|'.$predefinedbibitem.')'; + my $cmdleftright='\\\\(?:left|right|[Bb]igg?[lrm]?|middle)\s*(?:[<>()\[\]|\.]|\\\\(?:[|{}]|\w+))'; my $cmdoptseq='\\\\[\w\d@\*]+'.$extraspace.'(?:(?:<'.$abrat0.'>|\['.$brat_n.'\]|\{'. $pat_n . '\}|\(' . $coords .'\))'.$extraspace.')*'; my $defseq='\\\\def\\\\[\w\d@\*]+(?:#\d+|\[#\d+\])+(?:\{'. $pat_n . '\})?'; @@ -1152,7 +1196,7 @@ push(@SAFECMDLIST, qr/^QLEFTBRACE$/, qr/^QRIGHTBRACE$/); my $math='\$(?:[^$]|\\\$)*?\$|\\\\[(](?:.|\n)*?\\\\[)]'; ## the current maths command cannot cope with newline within the math expression my $comment='%[^\n]*\n'; - my $pat=qr/(?:\A\s*)?(?:${and}|${quotemarks}|${number}|${word}|$quotedunderscore|${defseq}|$cmdleftright|${cmdoptseq}|${math}|${backslashnl}|${oneletcmd}|${comment}|${punct}|${mathpunct}|\{|\})\s*/ ; + my $pat=qr/(?:\A\s*)?(?:${and}|${quotemarks}|${number}|${word}|$quotedunderscore|${defseq}|$cmdleftright|${predefinedcmdoptseq}|${cmdoptseq}|${math}|${backslashnl}|${oneletcmd}|${comment}|${punct}|${mathpunct}|\{|\})\s*/ ; @@ -1203,9 +1247,56 @@ if (lc($encoding) eq "utf8" ) { binmode(STDERR, ":utf8"); } +# filter($text) +# Runs $text through the script provided in $filterscript argument, if set +# If not set, just returns $text unchanged. +# If flatten was set, defer filtering to flatten. flatten will run the filter +# on all incoming text prior to its own processing. +# If flatten was not set, filter each of old and new once (see just below this def) +sub filter { + my ($text)=@_; + my ($textout,$pid); + if ($filterscript ne "") { + print STDERR "Passing " . length($text) . " chars to filter script " . $filterscript . "\n" if $verbose; + + if ($ignorefilterstderr) { + # If we need to capture and bury STDERR, use the Open3 version, and close CHLD_ERR below. + use IPC::Open3; + # We consume STDERR from the process, and hide it + $pid = open3(\*CHLD_IN, \*CHLD_OUT, \*CHLD_ERR, $filterscript) or die "open3() failed $!"; + } + else { + # Capture STDOUT and use as our new $text. Allow STDERR to go to console. + use IPC::Open2; + $pid = open2(\*CHLD_OUT, \*CHLD_IN, $filterscript) or die "open2() failed $!"; + } + # Send in $text + print CHLD_IN $text."\n"; # Adding a newline just to make sure there is one. + close CHLD_IN; + # Wait for output and gather it up + while () { + $textout = $textout.$_; + } + if ($ignorefilterstderr) { + close CHLD_ERR; # Enable only if Open3 used above + } + # On the off chance a very long running and/or frequently called script is used. + waitpid( $pid, 0 ); + $text = $textout; + print STDERR "Received " . length($text) . " chars after filtering\n" if $verbose; + print STDERR $text if $verbose; + } + return $text; +} + $old=read_file_with_encoding($oldfile,$encoding); $new=read_file_with_encoding($newfile,$encoding); +if (not defined($flatten)) { + $old=filter($old); + $new=filter($new); +} + @@ -1217,11 +1308,11 @@ exetime(1); if ($flatten) { - $oldbody=flatten($oldbody,$oldpreamble,$oldfile,$encoding); - $newbody=flatten($newbody,$newpreamble,$newfile,$encoding); + $oldbody=flatten($oldbody,$oldpreamble,File::Spec->rel2abs($oldfile),$encoding); + $newbody=flatten($newbody,$newpreamble,File::Spec->rel2abs($newfile),$encoding); # flatten preamble - $oldpreamble=flatten($oldpreamble,$oldpreamble,$oldfile,$encoding); - $newpreamble=flatten($newpreamble,$newpreamble,$newfile,$encoding); + $oldpreamble=flatten($oldpreamble,$oldpreamble,File::Spec->rel2abs($oldfile),$encoding); + $newpreamble=flatten($newpreamble,$newpreamble,File::Spec->rel2abs($newfile),$encoding); } @@ -1476,6 +1567,9 @@ my ( $citpat); if ( defined $packages{"apacite"} ) { print STDERR "apacite package detected.\n" if $verbose ; $citpat='(?:mask)?(?:full|short|no)?cite(?:A|author|year|meta)?(?:NP)?'; +} elsif (defined $packages{"biblatex"}) { + print STDERR "biblatex package detected.\n" if $verbose ; + $citpat='(?:[cC]ites?|(?:[pP]aren|foot|[Tt]ext|[sS]mart|super)cites?\*?|footnotecitetex)'; } else { # citation command pattern for all other citation schemes $citpat='(?:cite\w*|nocite)'; @@ -1798,11 +1892,13 @@ sub remove_endinput { # encoding is the encoding sub flatten { my ($text,$preamble,$filename,$encoding)=@_; - my ($includeonly,$dirname,$fname,$newpage,$fullfile,$filecontent,$replacement,$begline,$bblfile,$subfile,$command,$verbenv,$verboptions); + my ($includeonly,$dirname,$fname,$newpage,$fullfile,$filecontent,$replacement,$begline,$inputcmd,$bblfile,$subfile,$command,$verbenv,$verboptions,$ignore,$fileonly); my ($subpreamble,$subbody,$subpost); + my ($subdir,$subdirfull,$importfilepath); require File::Basename ; - require File::Spec ; - $dirname = File::Basename::dirname($filename); + ### require File::Spec ; # now this is needed even if flatten option not given + $filename = File::Spec->rel2abs( $filename ) ; + ($ignore, $dirname, $fileonly) = File::Spec->splitpath($filename) ; $bblfile = $filename; $bblfile=~s/\.tex$//; $bblfile.=".bbl"; @@ -1815,11 +1911,51 @@ sub flatten { print STDERR "DEBUG: includeonly $includeonly\n" if $debug; + + # Run through filter, to let filterscript have a pass if it was set + $text = filter($text); + + # Recursively replace \\import and \\subimport files + $text =~ s/(^(?:[^%\n]|\\%)*)(\\subimport\{(.*?)\}|\\import\{(.*?)\})(?:[\s]*)\{(.*?)\}/{ + # $1 is begline + # $3 is directory if subimport + # $4 is directory if import + # $5 is filename + $begline = (defined($1)? $1 : ""); + $subdir = $3 if defined($3); + $subdir = $4 if defined($4); + $fname = $5; + $fname .= ".tex" unless $fname =~ m|\.\w{3,4}$|; + print STDERR "DEBUG begline:", $begline, "\n" if $debug; + print STDERR "DEBUG", (defined($3)? "subimport_file:" : "import_file:"), $subdir, "\n" if $debug; + print STDERR "DEBUG file:", $fname, "\n" if $debug; + + # subimport appends $subdir to the current $dirname. import replaces it with an absolute path. + $subdirfull = (defined($3) ? File::Spec->catdir($dirname,$subdir) : $subdir); + + $importfilepath = File::Spec->catfile($subdirfull, $fname); + + print STDERR "importing importfilepath:", $importfilepath,"\n" if $verbose; + if ( -f $importfilepath ) { + # If file exists, replace input or include command with expanded input + $replacement=flatten(read_file_with_encoding($importfilepath, $encoding), $preamble,$importfilepath,$encoding) or die "Could not open file ",$fullfile,": $!"; + } else { + # if file does not exist, do not expand include or input command (do not warn if fname contains #[0-9] as it is then likely part of a command definition + # and is not meant to be expanded directly + print STDERR "WARNING: Could not find included file ",$importfilepath,". I will continue but not expand |$2|\n"; + $replacement=(defined($3)? "\\subimport" : "\\import"); + $replacement .= "{$subdir}{$fname} % Processed"; + } + "$begline$replacement"; + }/exgm; + # recursively replace \\input and \\include files $text =~ s/(^(?:[^%\n]|\\%)*)(\\input\{(.*?)\}|\\include\{(${includeonly}(?:\.tex)?)\})/{ $begline=(defined($1)? $1 : "") ; + $inputcmd=$2; $fname = $3 if defined($3) ; $fname = $4 if defined($4) ; + $newpage=(defined($4)? " \\newpage " : "") ; # # add tex extension unless there is a three or four letter extension already $fname .= ".tex" unless $fname =~ m|\.\w{3,4}$|; $fullfile = File::Spec->catfile($dirname,$fname); @@ -1832,12 +1968,11 @@ sub flatten { $replacement=flatten(read_file_with_encoding($fullfile, $encoding), $preamble,$filename,$encoding) or die "Could not open file ",$fullfile,": $!"; $replacement = remove_endinput($replacement); # \include always starts a new page; use explicit \newpage command to simulate this - $newpage=(defined($4)? " \\newpage " : "") ; } else { # if file does not exist, do not expand include or input command (do not warn if fname contains #[0-9] as it is then likely part of a command definition # and is not meant to be expanded directly - print STDERR "WARNING: Could not find included file ",$fullfile,". I will continue but not expand |$2|\n" unless $fname =~ m(#[0-9]) ; - $replacement = $2 ; # i.e. just the original command again -> make no change file does not exist + print STDERR "WARNING: Could not find included file ",$fullfile,". I will continue but not expand |$inputcmd|\n" unless $fname =~ m(#[0-9])n ; + $replacement = $inputcmd ; # i.e. just the original command again -> make no change file does not exist $newpage=""; } "$begline$newpage$replacement$newpage"; @@ -2116,6 +2251,7 @@ sub splitlatex { } } + unshift(@retval,$leadin) if (length($leadin)>0); return @retval; } @@ -2297,10 +2433,14 @@ sub extracttextblocks { my $wpat=qr/^(?:[a-zA-Z.,'`:;?()!]*)[\s~]*$/; #' my $retval=[]; + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace='\s*'; + for ($i=0;$i< scalar @$block;$i++) { ($token,$index)=@{ $block->[$i] }; # store pure text blocks - if ($token =~ /$wpat/ || ( $token =~/^\\((?:[`'^"~=.]|[\w\d@\*]+))((?:${extraspace}\[$brat_n\]${extraspace}|${extraspace}\{$pat_n\})*)/o + if ($token =~ /$wpat/ || ( $token =~/^\\((?:[`'^"~=.]|[\w\d@\*]+))((?:${extraspace}\[$brat_n\]${extraspace}|${extraspace}\{$pat_n\})*)/ && iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL) && !iscmd($1,\@TEXTCMDLIST,\@TEXTCMDEXCL))) { # we have text or a command which can be treated as text @@ -2345,6 +2485,10 @@ sub extractcommands { my ($i,$token,$index,$cmd,$open,$mid,$closing); my $retval=[]; + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace='\s*'; + for ($i=0;$i< scalar @$block;$i++) { ($token,$index)=@{ $block->[$i] }; # check if token is an alphanumeric command sequence with at least one non-optional argument @@ -2356,6 +2500,7 @@ sub extractcommands { # $4: } + trailing spaces if ( ( $token =~ m/^(\\([\w\d\*]+)(?:${extraspace}\[$brat_n\]|${extraspace}\{$pat_n\})*${extraspace}\{)($pat_n)(\}\s*)$/so ) && iscmd($2,\@TEXTCMDLIST,\@TEXTCMDEXCL) ) { + print STDERR "DEBUG EXTRACTCOMMANDS Match |$1|$2|$3|$4|$index \n" if $debug; # push(@$retval,[ $2,$index,$1,$3,$4 ]); ($cmd,$open,$mid,$closing) = ($2,$1,$3,$4) ; $closing =~ s/\}/\\RIGHTBRACE/ ; @@ -2371,20 +2516,20 @@ sub extractcommands { sub iscmd { my ($cmd,$regexar,$regexexcl)=@_; my ($ret)=0; - print STDERR "DEBUG: iscmd($cmd)=" if $debug; + ### print STDERR "DEBUG: iscmd($cmd)=" if $debug; foreach $pat ( @$regexar ) { if ( $cmd =~ m/^${pat}$/ ) { $ret=1 ; last; } } - print STDERR "0\n" if ($debug && !$ret) ; + ### print STDERR "0\n" if ($debug && !$ret) ; return 0 unless $ret; foreach $pat ( @$regexexcl ) { - print STDERR "0\n" if ( $debug && $cmd =~ m/^${pat}$/) ; + ### print STDERR "0\n" if ( $debug && $cmd =~ m/^${pat}$/) ; return 0 if ( $cmd =~ m/^${pat}$/ ); } - print STDERR "1\n" if $debug; + ### print STDERR "1\n" if $debug; return 1; } @@ -2443,7 +2588,6 @@ sub pass2 { print STDERR " $deltokcnt discarded tokens in $delblkcnt blocks.\n"; print STDERR " $addtokcnt appended tokens in $addblkcnt blocks.\n"; } - return(@$retval); } @@ -2466,13 +2610,20 @@ sub marktags { my $cmdcomment= ($opencmd =~ m/^%/); # Flag to indicate whether opencmd is a comment (i.e. if we intend to simply comment out changed commands) my ($command,$commandword,$closingbracket) ; # temporary variables needed below to remember sub-pattern matches -# split this block to flatten out sequences joined in pass1 + + +# split this block to split sequences joined in pass1 @$block=splitlatex(join "",@$block); ### print STDERR "DEBUG: marktags $openmark,$closemark,$open,$close,$opencmd,$closecmd,$comment\n" if $debug; ### print STDERR "DEBUG: marktags blocksplit ",join("|",@$block),"\n" if $debug; + + # we redefine locally $extraspace (shadowing the global definition) to capture command sequences with intervening spaces no matter what the global setting + # this is done so we can capture those commands with a predefined number of arguments without having to introduce them again explicitly here + my $extraspace_mt='\s*'; + + foreach (@$block) { $word=$_; - ### print STDERR "DEBUG MARKTAGS: |$word|\n" if $debug; if ( $word =~ s/^%/%$comment/ ) { # a comment if ($cmd==1) { @@ -2497,9 +2648,7 @@ sub marktags { } # negative lookahead pattern (?!) in second clause is put in to avoid matching \( .. \) patterns # also note that second pattern will match \\ - print STDERR "DEBUG marktags: Considering word |$word|\n" if $debug; - if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\(?!\()(\\|[`'^"~=.]|[\w*@]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { - ###print STDERR "DEBUG MARKTAGS is a non-safe command ($1)\n" if $debug; + if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\(?!\()(\\|[`'^"~=.]|[\w*@]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { ### if ( $word =~ /^[&{}\[\]]/ || ( $word =~ /^\\([\w*@\\% ]+)/ && !iscmd($1,\@SAFECMDLIST,\@SAFECMDEXCL)) ) { # word is a command or other significant token (not in SAFECMDLIST) ## same conditions as in subroutine extractcommand: @@ -2511,13 +2660,13 @@ sub marktags { # $3: last argument # $4: } + trailing spaces ### pre-0.3 if ( ( $token =~ m/^(\\([\w\d\*]+)(?:\[$brat0\]|\{$pat_n\})*\{)($pat_n)(\}\s*)$/so ) - if ( ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace}\[$brat_n\]|${extraspace}\{$pat_n\})*${extraspace}\{)($pat_n)(\}\s*)$/so ) + if ( ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace_mt}\[$brat_n\]|${extraspace_mt}\{$pat_n\})*${extraspace_mt}\{)($pat_n)(\}\s*)$/so ) && (iscmd($2,\@TEXTCMDLIST,\@TEXTCMDEXCL)|| iscmd($2,\@MATHTEXTCMDLIST,\@MATHTEXTCMDEXCL)) && ( !$cmdcomment || !iscmd($2,\@CONTEXT2CMDLIST, \@CONTEXT2CMDEXCL) ) ) { # Condition 1: word is a command? - if yes, $1,$2,.. will be set as above # Condition 2: word is a text command - we mark up the interior of the word. There is a separate check for MATHTEXTCMDLIST # because for $mathmarkup=WHOLE, the commands should not be split in pass1 (ie. math mode commands are not in - # TEXTCMDLIST, but the interior of MATHTEXT commands should be highlighted in both deleted and added blocks + # TEXTCMDLIST, but the interior of MATHTEXT commnds should be highlighted in both deleted and added blocks # Condition 3: But if we are in a deleted block ($cmdcomment=1) and # $2 (the command) is in context2, just treat it as an ordinary command (i.e. comment it open with $opencmd) # Because we do not want to disable this command @@ -2553,6 +2702,13 @@ sub marktags { } push (@$retval,$AUXCMD,"\n") if $cmdcomment ; $cmd=-1 ; + } elsif ( $cmdcomment && + ( $word =~ m/^(\\([\w\d\*]+)(?:${extraspace_mt}\[$brat_n\]|${extraspace_mt}\{$pat_n\})*${extraspace_mt}\{)($pat_n)(\}\s*)/so ) + && iscmd($2,\@KEEPCMDLIST, \@KEEPCMDEXCL) ) { + # 'keepcmd' in a deleted environment: keep the command as is + push (@$retval,$close) if $cmd==0 ; + push (@$retval,$word); + $cmd=-1; # pretend we are at the beginning of a sequence because we do not want to add an additional $closecmd or $close before the next token, no matter what it is } else { # ordinary command push (@$retval,$opencmd) if $cmd==-1 ; @@ -2589,52 +2745,19 @@ sub marktags { } #used in preprocess -sub take_comments_and_enter_from_frac() { - #*************take the \n and % between frac and {}*********** - #notice all of the substitution are made none global - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\frac(([\s]*%[^\n]*?)*[\r\n|\r|\n])+\{(.*?)\\end\{\1}/s ) { - # if there isn't any % or \n in the pattern $2 then there should be an \\end{...} in $2 - ### print STDERR "Match the following in take_comments and_enter_from_frac(1):\n****$&****\n" if $debug; - if( $2 !~ m/\\end\{$1}/s ) { - # take out % and \n from the next match only (none global) - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\frac(([\s]*%[^\n]*?)*[\r\n|\r|\n])+\{(.*?)\\end\{\1}/\\begin{$1}$2\\frac{$5\\end{$1}/s; - } - else{ - #there are no more % and \n in $2, we want to find the next one so we clear the begin-end from the pattern - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\end\{\1}/MATHBLOCK$1\{$2\}MATHBLOCKEND/s; - } - } - ###cleaning up - while( s/MATHBLOCK($MATHENV|$MATHARRENV|SQUAREBRACKET)\{(.*?)\}MATHBLOCKEND/\\begin{$1}$2\\end{$1}/s ){} - ###*************take the \n and % between frac and {}*********** - - ###**********take the \n and % between {} and {} of the frac*************** - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\frac\{(.*?)\\end\{\1\}/s ) { - # if there isn't any more //frac before the first //end in the pattern $2 then there should be an \\end{...} in $2 - ###print STDERR "Match the following in take_comments and_enter_from_frac(2):\n****$&****\n" if $debug; - if( $2 !~ m/\\end\{$1\}/s ) { - # from now on CURRFRAC is the frac we are looking at - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\frac\{(.*?)\\end\{\1\}/\\begin\{$1\}$2CURRFRAC\{$3\\end\{$1\}/s; - while( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{(.*?)\\end\{\1\}/s ) { - if( m/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{($pat_n)\}([\s]*(%[^\n]*?)*[\r\n|\r|\n])+[\s]*\{(.*?)\\end\{\1}/s ) { - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)CURRFRAC\{($pat_n)\}([\s]*(%[^\n]*?)*[\r\n|\r|\n])+[\s]*\{(.*?)\\end\{\1\}/\\begin\{$1\}$2CURRFRAC\{$3\}\{$6\\end\{$1\}/s; - } - else { # there is no comment or \n between the two brackets {}{} - # change CURRFRAC to FRACSTART so we can change them all back to //frac{ when we finish - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)CURRFRAC\{(.*?)\\end\{\1}/\\begin{$1}$2FRACSTART\{$3\\end{$1}/s; - } - } - } - else{ - ###there are no more frac in $2, we want to find the next one so we clear the begin-end from the pattern - s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)}(.*?)\\end\{\1}/MATHBLOCK$1\{$2\}MATHBLOCKEND/s; - } - - } - ###cleaning up - while( s/MATHBLOCK($MATHENV|$MATHARRENV|SQUAREBRACKET)\{(.*?)\}MATHBLOCKEND/\\begin{$1}$2\\end{$1}/s ){} - s/FRACSTART/\\frac/g; - ###***************take the \n and % between {} and {} of the frac********************* +sub take_comments_and_newline_from_frac() { + # some special magic for common usage of frac, which does not conform to the latexdiff requirements but can be made to fit + # note that this is a rare exception to the general rule that the new tex can be reconstructed from the diff file + + # \frac12 -> \frac{1}{2} + s/\\frac(\d)(\w)/\\frac\{$1\}\{$2\}/g; + + # \frac1{2b} -> \frac{1}{2b} + s/\\frac(\d)/\\frac\{$1\}/g; + + # delete space and comment characters between \frac arguments +# s/\\frac(?:\s*?%[^\n]*?)*?(\{$pat_n\})\s*(\{$pat_n\})/\\frac$1$2/g; + s/\\frac(?:\s|%[^\n]*?)*(\{$pat_n\})(?:\s|%[^\n]*?)*(\{$pat_n\})/\\frac$1$2/g; } # preprocess($string, ..) @@ -2717,7 +2840,7 @@ sub preprocess { # resort to WHOLE or FINE, or NONE math mode processing. s/\\begin\{($ARRENV)}(.*?)\\end\{\1}/\\ARRAYBLOCK$1\{$2\\DIFANCHORARRB \}\\DIFANCHORARRE /sg; - take_comments_and_enter_from_frac(); + take_comments_and_newline_from_frac(); s/\\begin\{($MATHENV|$MATHARRENV|SQUAREBRACKET)\}(.*?)\\end\{\1\}/\\MATHBLOCK$1\{$2\\DIFANCHORMATHB \}\\DIFANCHORMATHE /sg; } @@ -3335,8 +3458,7 @@ sub restore_item_commands { # pre-process preamble by looking for commands used in \maketitle (title, author, date etc commands) # the list of commands is defined in CONTEXT2CMD # if found then use a bodydiff to mark up content, and replace the corresponding commands - # in both preambles by marked up version to 'fool' the linediff (such that only body is marked - # up. + # in both preambles by marked up version to 'fool' the linediff (such that only body is marked up). # A special case are e.g. author commands being added (or removed) # 1. If commands are added, then the entire content is marked up as new, but also the lines are marked as new in the linediff # 2. If commands are removed, then the linediff will mark the line as deleted. The program returns @@ -3414,6 +3536,9 @@ EOF } $argnew=$newhash{$cmd}->[2]; $argdiff="{" . join("",bodydiff($argold,$argnew)) ."}"; + # Replace \RIGHTBRACE by } + $argdiff =~ s/\\RIGHTBRACE/}/g; + if ( length $optargnew ) { $optargdiff="[".join("",bodydiff($optargold,$optargnew))."]" ; $optargdiff =~ s/\\DIFaddbegin /\\DIFaddbeginFL /g; @@ -3759,6 +3884,7 @@ be real files (not pipes or similar) as they are opened twice. reason package specific parsing needs to be switched off, use --packages=none. The following packages trigger special behaviour: endfloat hyperref amsmath apacite siunitx cleveref glossaries mhchem chemformula/chemmacros + biblatex [ Default: scan the preamble for \\usepackage commands to determine loaded packages.] @@ -3840,7 +3966,17 @@ Miscelleneous options be located in the same directories as the old and new master files, respectively, making it possible to organise files into old and new directories. --flatten is applied recursively, so inputted files can contain further - \\input statements. + \\input statements. Also handles files included by the import package + (\\import and \\subimport), and \\subfile command. + +--filter-script=filterscript Run files through this filterscript (full path preferred) before processing. + The filterscript must take STDIN input and output to STDOUT. + When coupled with --flatten, each file will be run through the filter as it is brought in. + +--ignore-filter-stderr When running with --filter-script, STDERR from the script may cause readability issues. + Turn this flag on to ignore STDERR from the filter script. + + --help -h Show this help text. @@ -4095,7 +4231,7 @@ define new C<\DIFadd> and C<\DIFdel> commands, which provide a wrapper for these using them for the text but not for the link defining command (where any markup would cause errors). -=item C +=item C, C Redefine the commands recognised as citation commands. @@ -4401,12 +4537,26 @@ included sections are not expanded. The included files are assumed to be located in the same directories as the old and new master files, respectively, making it possible to organise files into old and new directories. --flatten is applied recursively, so inputted files can contain further -C<\input> statements. +C<\input> statements. Also handles files included by the import package +(C<\import> and C<\subimport>), and C<\subfile> command. Use of this option might result in prohibitive processing times for larger documents, and the resulting difference document no longer reflects the structure of the input documents. +=item B<--filter-script=filterscript> + +Run files through this filterscript (full path preferred) before processing. +The filterscript must take STDIN input and output to STDOUT. +When coupled with --flatten, each file will be run through the filter as it is brought in. + +=item B<--ignore-filter-stderr> + +When running with --filter-script, STDERR from the script may cause readability issues. +Turn this flag on to ignore STDERR from the filter script. + + + =item B<--help> or B<-h> @@ -4968,6 +5118,8 @@ min Pr sec sup +bibfield +bibinfo [Hclbkdruvt] [`'^"~=.] _ @@ -4994,6 +5146,7 @@ framebox footnote footnotetext framebox +href intertext part (sub){0,2}section\*? @@ -5007,6 +5160,7 @@ savebox sbox shortintertext shortstack +sidenote signature text.* value @@ -5151,7 +5305,7 @@ verbatim[*]? %DIF CULINECHBAR PREAMBLE \RequirePackage[normalem]{ulem} \RequirePackage[dvips]{changebar} -\RequirePackage{color} +\RequirePackage{color}\definecolor{RED}{rgb}{1,0,0}\definecolor{BLUE}{rgb}{0,0,1} \providecommand{\DIFadd}[1]{\protect\cbstart{\protect\color{blue}\uwave{#1}}\protect\cbend} \providecommand{\DIFdel}[1]{\protect\cbdelete{\protect\color{red}\sout{#1}}\protect\cbdelete} %DIF END CULINECHBAR PREAMBLE diff --git a/Master/texmf-dist/scripts/latexdiff/latexrevise.pl b/Master/texmf-dist/scripts/latexdiff/latexrevise.pl index 956d13e5061..f7f88f1b951 100755 --- a/Master/texmf-dist/scripts/latexdiff/latexrevise.pl +++ b/Master/texmf-dist/scripts/latexdiff/latexrevise.pl @@ -21,6 +21,7 @@ # # Detailed usage information at the end of the file # Note: version number now keeping up with latexdiff +# Version 1.3.1 (Bug fix) Remove some deprecation warnings due to uncommented left parentheses # Version 1.0.2 Option --version # Version 1.0.1 no changes to latexrevise # Version 0.3 Updated for compatibility with latexdiff 0.3 output (DIFAUXCMD removal) @@ -31,8 +32,8 @@ use strict; use warnings; my $versionstring=< or B<--comment=cmd> Remove C<\cmd{...}> sequences. C is supposed to mark some explicit -anotations which should be removed from the file before +annotations which should be removed from the file before release. =item B<-e envir> or B<--comment-environment=envir> @@ -508,8 +525,9 @@ which should have been removed already. =head1 BUGS The current version is a beta version which has not yet been -extensively tested, but worked fine locally. Please submit bug reports using the issue tracker of the github repository page I, -or send them to I.. Include the serial number of I +extensively tested. It has not been actively maintained so might not process output of newer versions of latexdiff entirely correctly. +Please submit bug reports using the issue tracker of the github repository page I, +or send them to I. Include the serial number of I (Option --version). If you come across latexdiff output which is not processed correctly by I please include the problem file as well as the old and new files on which it is based, -- cgit v1.2.3