From a072ab3fdb413ffeb05e0d5136400ad418d89623 Mon Sep 17 00:00:00 2001 From: Karl Berry Date: Sat, 26 Sep 2009 17:55:25 +0000 Subject: new script texdiff 0.4 (24sep09) git-svn-id: svn://tug.org/texlive/trunk@15464 c570f23f-e606-0410-a88d-b1316a301751 --- Master/texmf-dist/doc/support/texdiff/README | 84 ++++ Master/texmf-dist/scripts/texdiff/texdiff | 554 +++++++++++++++++++++++++++ 2 files changed, 638 insertions(+) create mode 100644 Master/texmf-dist/doc/support/texdiff/README create mode 100755 Master/texmf-dist/scripts/texdiff/texdiff (limited to 'Master/texmf-dist') diff --git a/Master/texmf-dist/doc/support/texdiff/README b/Master/texmf-dist/doc/support/texdiff/README new file mode 100644 index 00000000000..c0485dacbf0 --- /dev/null +++ b/Master/texmf-dist/doc/support/texdiff/README @@ -0,0 +1,84 @@ +NAME + texdiff - Compares two (La)TeX documents to create a merged version + showing changes, similar to that of 'Change Tracking' in some word + processors. + +SYNOPSIS + texdiff [--nofloats|-n] old.tex new.tex [diff.tex] + +DESCRIPTION + The first two files are compared and a new TeX file is output. When the + output file is processed with (La)TeX it marks new changes with blue and + old text with red with a strike-through line. Furthermore, passages with + changes are marked at the margin with grey bars by the LaTeX "changebar" + package (although only after running "dvips", in "xdvi" the bars are one + inch too far left to compensate the one inch offset inserted by TeX + printer drivers). + + "texdiff" uses and therefore requires the "wdiff" command to insert + these marks. If LaTeX macros are included in those striked out or + underlined passages, the latex run will most likely fail. The source of + most problems lies in "wdiff" tearing apart pairs of braces. Therefore + "texdiff" is a perl script that calls "wdiff" internally and fixes + "wdiff"'s output by re-pairing braces again and handling some commonly + used LaTeX macros. + + For texdiff to work, the following LaTeX code must be inserted in the + preamble of the LaTeX document: + + \usepackage{xcolor} \usepackage{ulem} \usepackage{changebar} + \newcommand\TLSins[1]{\cbstart{}\textcolor{ins}{\uline{#1}}\cbend{}} + \newcommand\TLSdel[1]{\cbdelete{}\textcolor{del}{\sout{#1}}} + + Define editing colors "ins" and "del" with existing "xcolor" color + names: \colorlet{ins}{blue} \colorlet{del}{red} + + Or from scratch: \definecolor{ins}{rgb}{0, 0.5, 0} + + These macros are NOT automatically interted at \begin{document} like the + original scripts did (see the ACKNOWLEDGMENTS section below), because + this will not work on files of a multi-file document (those included by + \include or \input). Furthermore, if you insert the macros manually, you + can tune them as you like, i. e. change text color or position of + changebars. Have a look at the documentation of the ulem and changebar + packages on how to do this! + +OPTIONS + --nofloats, -n + If given, skips contents of figure environments, but does show + insertions and deletions in the caption commands. This is done to + avoid complex messups in subfloat or complex figure environments. + + --debug, -d + Turns on debugging. Temporary files that contain numbered braces are + kept and a raw difference file is written as output file name + appended with '-wdiff.tex'. In addition, several regular expression + diagnostics are printed. Warning: produces a lot of output. + +AUTHOR + Mark Doll gmx.net>, Cengiz Gunay + users.sf.net>. + +ACKNOWLEDGMENTS + * Version 0.4, by CG, 2009/09/15. Added use of GetOpt to take an + optional parameter to skip float (e.g., figure or table) contents. + Added an optional debug mode. Embedded POD documentation. Fixed + problem with insertion and deletions appearing in the middle of + sectioning commands by throwing off LaTeX. + + * Version 0.3, modified by Cengiz Gunay , + 2009/06/30. Added some Latex commands that should be excluded and + fixed handling of '\%' characters. + + * Version 0.2, is a perl implementation of TeXdiff by Mark Doll + , 2006/08/09. It was derived from the original + bash and perl scripts created by Robert Maron , + available at http://www.robmar.net/TexDiff/. This version of texdiff + is available at http://mark.doll.name/texdiff/. + +COPYRIGHT AND LICENSE + Copyleft 2006-2009, Mark Doll and Cengiz Gunay + + This library is free software; you may redistribute it and/or modify it + under the same terms as Perl itself. + diff --git a/Master/texmf-dist/scripts/texdiff/texdiff b/Master/texmf-dist/scripts/texdiff/texdiff new file mode 100755 index 00000000000..4c27fab205a --- /dev/null +++ b/Master/texmf-dist/scripts/texdiff/texdiff @@ -0,0 +1,554 @@ +#! /usr/bin/env perl +# +# See POD documentation below or use: +# +# $ perldoc texdiff +# +# to see embedded documentation. Alternatively, you can create manual +# or HTML pages using pod2man and pod2html. Below the POD part you +# can add your own macros to be skipped by the edit marks. +# +# $Id: texdiff 766 2009-09-24 19:19:38Z cengiz $ + +=pod + +=head1 NAME + +B - Compares two (La)TeX documents to create a merged version showing changes, similar to that of 'Change Tracking' in some word processors. + +=head1 SYNOPSIS + +texdiff [--nofloats|-n] old.tex new.tex [diff.tex] + +=head1 DESCRIPTION + +The first two files are compared and a new TeX file is output. When +the output file is processed with (La)TeX it marks new changes with +blue and old text with red with a strike-through line. Furthermore, +passages with changes are marked at the margin with grey bars by the +LaTeX C package (although only after running C, in +C the bars are one inch too far left to compensate the one inch +offset inserted by TeX printer drivers). + +C uses and therefore requires the C command to insert +these marks. If LaTeX macros are included in those striked out or +underlined passages, the latex run will most likely fail. The source +of most problems lies in C tearing apart pairs of +braces. Therefore C is a perl script that calls C internally +and fixes C's output by re-pairing braces again and handling some +commonly used LaTeX macros. + +For texdiff to work, the following LaTeX code must be inserted in the +preamble of the LaTeX document: + + \usepackage{xcolor} \usepackage{ulem} \usepackage{changebar} + \newcommand\TLSins[1]{\cbstart{}\textcolor{ins}{\uline{#1}}\cbend{}} + \newcommand\TLSdel[1]{\cbdelete{}\textcolor{del}{\sout{#1}}} + +Define editing colors C and C with existing C color names: + \colorlet{ins}{blue} \colorlet{del}{red} + +Or from scratch: + \definecolor{ins}{rgb}{0, 0.5, 0} + +These macros are NOT automatically interted at \begin{document} like +the original scripts did (see the B section below), +because this will not work on files of a multi-file document (those +included by \include or \input). Furthermore, if you insert the +macros manually, you can tune them as you like, i. e. change text +color or position of changebars. Have a look at the documentation of +the ulem and changebar packages on how to do this! + +=head1 OPTIONS + +=over + +=item B<--nofloats>, B<-n> + +If given, skips contents of figure environments, but does show +insertions and deletions in the caption commands. This is done to +avoid complex messups in subfloat or complex figure environments. + +=item B<--debug>, B<-d> + +Turns on debugging. Temporary files that contain numbered braces are +kept and a raw difference file is written as output file name appended +with '-wdiff.tex'. In addition, several regular expression diagnostics +are printed. B + +=back + +=head1 AUTHOR + +Mark Doll gmx.net>, Cengiz Gunay users.sf.net>. + +=head1 ACKNOWLEDGMENTS + +=over + +=item B<*> + +Version 0.4, by CG, 2009/09/15. Added use of GetOpt to take an optional +parameter to skip float (e.g., figure or table) contents. Added an +optional debug mode. Embedded POD documentation. Fixed problem +with insertion and deletions appearing in the middle of sectioning +commands by throwing off LaTeX. + +=item B<*> + +Version 0.3, modified by Cengiz Gunay , +2009/06/30. Added some Latex commands that should be excluded and +fixed handling of '\%' characters. + +=item B<*> + +Version 0.2, is a perl implementation of TeXdiff by Mark Doll +, 2006/08/09. It was derived from the original bash +and perl scripts created by Robert Maron , +available at http://www.robmar.net/TexDiff/. This version of texdiff +is available at http://mark.doll.name/texdiff/. + +=back + +=head1 COPYRIGHT AND LICENSE + +Copyleft 2006-2009, Mark Doll and Cengiz Gunay + +This library is free software; you may redistribute it and/or modify +it under the same terms as Perl itself. + +=cut + +# TODO: +# +# - after removing excluded commands, unify deletions and insertions +# if there're no other differences +# - investigate causes of missing brace pairs +# - put option to add preamble automatically? +# - put option to flatten first? + +use Getopt::Long qw(GetOptions); +use Pod::Usage; +use strict; +use re 'eval'; # CG: for RE code for skipping float contents + +## Problematic commands, that will be removed from deletions and moved +## out of insertions. The array index equals the number of mandatory +## command parameters. + +## INSERT YOUR HOMEBREW MACROS HERE!!! +# Omit leading backslash and separate by '|', since this will be +# interpreted as a regular expression. +my @mycommands = ( + 'RCS\$[^\$]*\$', + 'cite\w*p|citet|svnid|citep', + 'if\w+', + '', + '', + 'Abb(?:ildung)?(?:p[psw])?', +); + +## standard LaTeX macros +#FIXME: add more macros here! +my @stdcommands = ( + '\\\\|item|\%', # will match one backslash inside regex + 'includegraphics|begin|end|input|fancyfoot|subfig|subfloat', +); + +# GetOpt::Long:: +my $nofloats = 0; +my $help; +my $DEBUG=0; + +my $result = GetOptions ("nofloats|n" => \$nofloats, + "debug|d" => \$DEBUG, + "help|?" => \$help ); + +#print "Figs= $nofloats\n"; + +pod2usage( -section => "SYNOPSIS") if ($#ARGV < 1 || $help); + +## allow nesting of braces (aka curly brackets) up to 10**$w-1 levels +my $w = 2; + +## wdiff marker +my $delstart = "TLSdel"; +my $delend = "TLEdel"; +my $insstart = "TLSins"; +my $insend = "TLEins"; + +## LaTeX macros used to mark deletions/insertions in the output +my $texdelstart = "\\protect\\TLSdel\{"; +my $texdelend = "\}"; +my $texinsstart = "\\protect\\TLSins\{"; +my $texinsend = "\}"; + + +## parse command line +my ($in1,$in2,$out) = @ARGV; + +## create temp files from input files with all comments removed and +## all braces numbered +write_temp($w,$in1,'tmp1.'.$$); +write_temp($w,$in2,'tmp2.'.$$); +sub write_temp { + my ($w,$file,$tmp) = @_; + + $_ = `cat $file`; + + ## mark pairs of braces with the same $w digits. + # i. e. convert "{ { {} } {}}" to "{01 {02 {03}03 }02 {02}02}01" ($w=2). + # + # FIXME: marking braces before comparison makes them appear + # as changes if depth of nesting changes. + + my $max=10**$w-1; + $::cnt = 0; + s/ + (?= $max); + sprintf("\{%0${w}d",++$::cnt) }) | + \} (?{ sprintf("\}%0${w}d",$::cnt--) }) + ) + /$^R/gx; + + print STDERR "Warning: ".abs($::cnt)." unmatched ".($::cnt > 0 ? "opening" : "closing") + ." brace".(abs($::cnt) == 1 ? "" : "s")."\n" if ($::cnt != 0); + + ## remove comments + # also remove trailing newline and all whitespace at the + # beginning of the following line like TeX does it + s/(?$tmp") or return("Can't open $tmp: $!\n"); + print O; + close O; +} + +my $wdiff_cmd = ( "wdiff --avoid-wraps" + . " --start-delete=\'$delstart\' --end-delete=\'$delend\'" + . " --start-insert=\'$insstart\' --end-insert=\'$insend\'" + . " tmp1.$$ tmp2.$$"); +$_ = `$wdiff_cmd`; + +if ($DEBUG) { + ## save wdiff output + if (open(O,">$out-wdiff.tex")) { + print O; + close O; + } else { + print STDERR "Failed to write wdiff output file $out-wdiff.tex.\n";; + } +} + +### +### Phase I: generic processing +### + +## remove paragraph boundaries (two newlines) that wdiff falsely inserted +# FIXME: Does wdiff always insert same spacing before deletion and the +# following insertion?! +s/( + ([ \t]*\n[ \t]*\n[ \t]*) + $delstart (?!$delend) + (?: . (?!$delend) )* . + $delend + ) + \2 + $insstart + /$1\n$insstart/gsx; + + +## Restore backslashed Spaces (prevents falsely escaped right braces) +# Escaped spaces '\ ' at the end of an insertion/deletion will result in the +# backslash before the end marker and the space after it. Therefore swap +# whitespace and end marker. (This might falsely make the insertion/deletion +# span the following paragraph boundary, but this will be fixed by the next +# step below). When the end markers finally get replaced by $tex(ins|del)end, +# which typically is a '}', without this swapping, this would result in escaped +# right braces and LaTeX complaining about a missing right brace. +s/(? (?: ## grab all paired braces (prevent backtracking): + (?: . (?! (? (?: + (?: . (?! (?$#stdcommands ? $#mycommands : $#stdcommands)) { + my $command = ""; + + $command .= $mycommands[$num] if ($mycommands[$num] ); + $command .= "|" if ($mycommands[$num] && $stdcommands[$num]); + $command .= $stdcommands[$num] if ( $stdcommands[$num]); + + next if(!$command); + + s/ + ( # $1: anything before the command + $delstart (?!$delend) + (?: . (?!$delend) )*? .? + ) + ( # $2: the command with all its parameters + \\(?:$command) # the command itself (preceded by a '\') + (?: \[ .*? (? tested with multiple figures +## => tested with multi-captions and no captions. should work with subfig and subfloat + +s/( \\begin\{[0-9]{$w}\s*((?:figure|table)\*?)(?{ $::floatname = $^N })\s*\}[0-9]{$w} ) # begin figure or table line + (?{ print "=====Found figure!\n" if $DEBUG; + $::figcontents=$^N }) # Save into variable + (?: # Groups BLOCK + caption command + (?: # BLOCK of insert and delete commands until caption + ( (?! \\caption) (?: (?! $insstart | $delstart | \\caption | \\end\{[0-9]{$w}\s*figure ) . )* ) + (?{ print "=====Found ins\|del start or caption '$^N'!\n" if $DEBUG; + $::figcontents .= $^N }) + (?: $delstart (?: (?!$delend) . )* $delend + (?{ print "=====Found del end!\n" if $DEBUG; }) )? # dont keep deletions + (?: $insstart ( (?: (?!$insend) . )* ) $insend + (?{ print "=====Found ins end '$^N'!\n" if $DEBUG; $::figcontents .= $^N }) )? # keep insertions + )*? # end BLOCK + (?: (\\caption (?: \[ (?: (?! (?$out")) { + print O; + close O; + } else { + print STDERR "Failed to write output file $out.\n"; + } +} else { + print ; +} + +## remove temp files +if (!$DEBUG) { + unlink('tmp1.'.$$); + unlink('tmp2.'.$$); +} -- cgit v1.2.3