binutils-gdb/readline/doc/texi2dvi
Patrick Palka 4a11f20659 Sync readline/ to version 7.0 alpha
This patch syncs our upstream copy of readline from version 6.2 to the
latest version, 7.0 alpha (released July 10 2015).

I essentially copied what was done the last time readline was synced,
when Jan updated to readline 6.2 in 2011:
http://sourceware.org/ml/gdb-patches/2011-05/msg00003.html

Procedure:

1. I extracted the readline-7.0-alpha tarball on top of readline/.
2. I deleted all the new files under doc/ that were deliberately omitted
   before.
3. I regenerated readline/configure and readline/examples/rlfe/configure
   using autoconf 2.64.  No other configure files need regenerating.
4. I updated the function gdb_printable_part in completer.c with a
   trivial change made to the readline function it is based off of,
   printable_part in readline/complete.c.  There is more work to be done in
   completer.c to sync it with readline/complete.c, but it is non-trivial
   and should probably be done separately anyway.

Local patches that had to be reapplied:

    None.  readline 7.0 alpha contains all of our local readline
    patches.

New files in readline/:

    colors.{c,h}
    examples/{hist_erasedups,hist_purgecmd,rl-callbacktest,rlbasic}.c
    parse-colors.{c,h}
    readline.pc.in
    configure.ac

Deleted files in readline/:

    configure.in

Regressions:

After the sync there is one testsuite regression, the test
"signal SIGINT" in gdb.gdb/selftest.exp which now FAILs.  Previously,
the readline 6.2 SIGINT handler would temporarily reinstall the
underlying application's SIGINT handler and immediately re-raise SIGINT
so that the orginal handler gets invoked.  But now (since readline 6.3)
its SIGINT handler does not re-raise SIGINT or directly invoke the
original handler; it now sets a flag marking that SIGINT was raised, and
waits until readline explicitly has control to call the application's
SIGINT handler.  Anyway, because SIGINT is no longer re-raised from
within readline's SIGINT handler, doing "signal SIGINT" with a stopped
inferior gdb process will no longer resume and then immediately stop the
process (since there is no 2nd SIGINT to immediately catch).  Instead,
the inferior gdb process will now just print "Quit" and continue to run.
So with this commit, this particular test case is adjusted to reflect
this change in behavior (we now have to send a 2nd SIGINT manually to
stop it).

Aside from this one testsuite regression, I personally noticed no
regression in user-visible behavior.  Though I only tested on x86_64
and on i686 Debian Stretch.

Getting this kind of change in at the start of the GDB 7.11 development
cycle will allow us to get a lot of passive testing from developers and
from bleeding-edge users.

readline/ChangeLog.gdb:

	Import readline 7.0 alpha
	* configure: Regenerate.
	* examples/rlfe/configure: Regenerate.

gdb/ChangeLog:

	* completer.c (gdb_printable_part): Sync with readline function
	it is based off of.

gdb/testsuite/ChangeLog:

	* gdb.gdb/selftest.exp (test_with_self): Update test to now
	expect the GDB inferior to no longer immediately stop after
	being resumed with "signal SIGINT".
2015-07-25 09:53:01 -04:00

1945 lines
56 KiB
Bash
Executable File
Raw Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

#! /bin/sh
# texi2dvi --- produce DVI (or PDF) files from Texinfo (or (La)TeX) sources.
# $Id: texi2dvi 5704 2014-07-07 17:45:16Z karl $
#
# Copyright 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2001, 2002,
# 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014
# Free Software Foundation, Inc.
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License,
# or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
# Originally written by Noah Friedman.
#
# Please send bug reports, etc. to bug-texinfo@gnu.org.
# If possible, please send a copy of the output of the script called with
# the `--debug' option when making a bug report.
test -f /bin/ksh && test -z "$RUNNING_KSH" \
&& { UNAMES=`uname -s`; test "x$UNAMES" = xULTRIX; } 2>/dev/null \
&& { RUNNING_KSH=true; export RUNNING_KSH; exec /bin/ksh $0 ${1+"$@"}; }
unset RUNNING_KSH
# No failure shall remain unpunished.
set -e
# In case the default sed doesn't suffice.
: ${SED=sed}
# This string is expanded automatically when this file is checked out.
rcs_revision='$Revision: 5704 $'
rcs_version=`set - $rcs_revision; echo $2`
program=`echo $0 | $SED -e 's!.*/!!'`
build_mode=${TEXI2DVI_BUILD_MODE:-local}
build_dir=${TEXI2DVI_BUILD_DIRECTORY:-.}
# Initialize variables for option overriding and otherwise.
# Don't use `unset' since old bourne shells don't have this command.
# Instead, assign them an empty value.
action=compile
batch=false # interact normally
catcode_special=maybe
debug=false
escape="\\"
expand=false # true for expansion via makeinfo
includes=
line_error=true # pass --file-line-error to TeX
max_iters=7 # when to quit
oname= # --output
out_lang=dvi
quiet=false # let the tools' message be displayed
set_language=
src_specials=
shell_escape=
latex2html=hevea # or set to tex4ht
textra= # Extra TeX commands to insert in the input file.
txiprereq=19990129 # minimum texinfo.tex version with macro expansion
verb=false # true for verbose mode
translate_file= # name of charset translation file
orig_pwd=`pwd`
# We have to initialize IFS to space tab newline since we save and
# restore IFS and apparently POSIX allows stupid/broken behavior with
# empty-but-set IFS.
# http://lists.gnu.org/archive/html/automake-patches/2006-05/msg00008.html
# We need space, tab and new line, in precisely that order. And don't leave
# trailing blanks.
space=' '
tab=' '
newline='
'
IFS="$space$tab$newline"
# In case someone pedantic insists on using grep -E.
: ${EGREP=egrep}
# Systems which define $COMSPEC or $ComSpec use semicolons to separate
# directories in TEXINPUTS -- except for Cygwin et al., where COMSPEC
# might be inherited, but : is used.
if test -n "$COMSPEC$ComSpec" \
&& uname | $EGREP -iv 'cygwin|mingw|djgpp' >/dev/null; then
path_sep=";"
else
path_sep=":"
fi
# Pacify verbose cds.
CDPATH=${ZSH_VERSION+.}$path_sep
# If $TEX is set to a directory, don't use it.
test -n "$TEX" && test -d "$TEX" && unset TEX
#
## --------------------- ##
## Auxiliary functions. ##
## --------------------- ##
# In case `local' is not supported by the shell, provide a function
# that simulates it by simply performing the assignments. This means
# that we must not expect `local' to work, i.e., we must not (i) rely
# on it during recursion, and (ii) have two local declarations of the
# same variable. (ii) is easy to check statically, and our test suite
# does make sure there is never twice a static local declaration of a
# variable. (i) cannot be checked easily, so just be careful.
#
# Note that since we might use a function simulating `local', we can
# no longer rely on the fact that no IFS-splitting is performed. So,
# while
#
# foo=$bar
#
# is fine (no IFS-splitting), never write
#
# local foo=$bar
#
# but rather
#
# local foo="$bar"
(
foo=bar
test_local () {
local foo=foo
}
test_local >/dev/null 2>&1
test $foo = bar
) || eval '
local () {
case $1 in
*=*) eval "$1";;
esac
}
'
# cd_orig
# -------
# Return to the original directory.
cd_orig ()
{
# In case $orig_pwd is on a different drive (for DOS).
cd /
# Return to the original directory so that
# - the next file is processed in correct conditions
# - the temporary file can be removed
cd "$orig_pwd" || exit 1
}
# func_dirname FILE
# -----------------
# Return the directory part of FILE.
func_dirname ()
{
dirname "$1" 2>/dev/null \
|| { echo "$1" | $SED 's!/[^/]*$!!;s!^$!.!'; }
}
# noexit FILE
# -----------
# Return FILE with one extension remove. foo.bar.baz -> foo.bar.
noext ()
{
echo "$1" | $SED -e 's/\.[^/.][^/.]*$//'
}
# absolute NAME -> ABS-NAME
# -------------------------
# Return an absolute path to NAME.
absolute ()
{
case $1 in
[\\/]* | ?:[\\/]*)
# Absolute paths don't need to be expanded.
echo "$1"
;;
*) local slashes
slashes=`echo "$1" | $SED -n 's,.*[^/]\(/*\)$,\1,p'`
local rel
rel=$orig_pwd/`func_dirname "$1"`
if test -d "$rel"; then
(cd "$rel" 2>/dev/null \
&& local n
n=`pwd`/`basename "$1"`"$slashes"
echo "$n")
else
error 1 "not a directory: $rel"
fi
;;
esac
}
# ensure_dir DIR1 DIR2...
# -----------------------
# Make sure the directories exist.
ensure_dir ()
{
for dir
do
# Beware that in parallel builds we may have several concurrent
# attempts to create the directory. So fail only if "mkdir"
# failed *and* the directory still does not exist.
test -d "$dir" \
|| mkdir "$dir" \
|| test -d "$dir" \
|| error 1 "cannot create directory: $dir"
done
}
# error EXIT_STATUS LINE1 LINE2...
# --------------------------------
# Report an error and exit with failure if EXIT_STATUS is non-null.
error ()
{
local s="$1"
shift
report "$@"
if test "$s" != 0; then
exit $s
fi
}
# findprog PROG
# -------------
# Return true if PROG is somewhere in PATH, else false.
findprog ()
{
local saveIFS="$IFS"
IFS=$path_sep # break path components at the path separator
for dir in $PATH; do
IFS=$saveIFS
# The basic test for an executable is `test -f $f && test -x $f'.
# (`test -x' is not enough, because it can also be true for directories.)
# We have to try this both for $1 and $1.exe.
#
# Note: On Cygwin and DJGPP, `test -x' also looks for .exe. On Cygwin,
# also `test -f' has this enhancement, but not on DJGPP. (Both are
# design decisions, so there is little chance to make them consistent.)
# Thusly, it seems to be difficult to make use of these enhancements.
#
if { test -f "$dir/$1" && test -x "$dir/$1"; } \
|| { test -f "$dir/$1.exe" && test -x "$dir/$1.exe"; }; then
return 0
fi
done
return 1
}
# report LINE1 LINE2...
# ---------------------
# Report some information on stderr.
report ()
{
for i in "$@"
do
echo >&2 "$0: $i"
done
}
# run COMMAND-LINE
# ----------------
# Run the COMMAND-LINE verbosely, and catching errors as failures.
run ()
{
verbose "Running $@"
"$@" 2>&5 1>&2 \
|| error 1 "$1 failed"
}
# usage
# -----
# Display usage and exit successfully.
usage ()
{
# We used to simply have `echo "$usage"', but coping with the
# changing behavior of `echo' is much harder than simply using a
# here-doc.
#
# echo '\noto' echo '\\noto' echo -e '\\noto'
# bash 3.1 \noto \\noto \noto
# bash 3.2 %oto \noto -e \noto
#
# where % denotes the eol character.
cat <<EOF
Usage: $program [OPTION]... FILE...
or: texi2pdf [OPTION]... FILE...
or: pdftexi2dvi [OPTION]... FILE...
Run each Texinfo or (La)TeX FILE through TeX in turn until all
cross-references are resolved, building all indices. The directory
containing each FILE is searched for included files. The suffix of FILE
is used to determine its language ((La)TeX or Texinfo). To process
(e)plain TeX files, set the environment variable LATEX=tex.
In order to make texi2dvi a drop-in replacement of TeX/LaTeX in AUC-TeX,
the FILE may also be composed of the following simple TeX commands.
\`\\input{FILE}' the actual file to compile
\`\\nonstopmode' same as --batch
When invoked as \`texi2pdf' or \`pdftexi2dvi', or given the option --pdf
or --dvipdf, generate PDF output. Otherwise, generate DVI.
General options:
-b, --batch no interaction
-D, --debug turn on shell debugging (set -x)
-h, --help display this help and exit successfully
-o, --output=OFILE leave output in OFILE; only one input FILE is allowed
-q, --quiet no output unless errors (implies --batch)
-s, --silent same as --quiet
-v, --version display version information and exit successfully
-V, --verbose report on what is done
Output format:
--dvi output a DVI file [default]
--dvipdf output a PDF file via DVI (using a dvi-to-pdf program)
--html output an HTML file from LaTeX, using HeVeA
--info output an Info file from LaTeX, using HeVeA
-p, --pdf use pdftex or pdflatex for processing
--ps output a PostScript file via DVI (using dvips)
--text output a plain text file from LaTeX, using HeVeA
TeX tuning:
-@ use @input instead of \input for preloaded Texinfo
-e, -E, --expand force macro expansion using makeinfo
-I DIR search DIR for Texinfo files
-l, --language=LANG specify LANG for FILE, either latex or texinfo
--no-line-error do not pass --file-line-error to TeX
--shell-escape pass --shell-escape to TeX
--src-specials pass --src-specials to TeX
-t, --command=CMD insert CMD in copy of input file
or --texinfo=CMD multiple values accumulate
--translate-file=FILE use given charset translation file for TeX
Build modes:
--build=MODE specify the treatment of auxiliary files [$build_mode]
--tidy same as --build=tidy
-c, --clean same as --build=clean
--build-dir=DIR specify where the tidy compilation is performed;
implies --tidy;
defaults to TEXI2DVI_BUILD_DIRECTORY [$build_dir]
--mostly-clean remove the auxiliary files and directories
but not the output
--max-iterations=N don't process files more than N times [$max_iters]
The MODE specifies where the TeX compilation takes place, and, as a
consequence, how auxiliary files are treated. The build mode
can also be set using the environment variable TEXI2DVI_BUILD_MODE.
Valid MODEs are:
\`local' compile in the current directory, leaving all the auxiliary
files around. This is the traditional TeX use.
\`tidy' compile in a local *.t2d directory, where the auxiliary files
are left. Output files are copied back to the original file.
\`clean' same as \`tidy', but remove the auxiliary directory afterwards.
Every compilation therefore requires the full cycle.
Using the \`tidy' mode brings several advantages:
- the current directory is not cluttered with plethora of temporary files.
- clutter can be even further reduced using --build-dir=dir: all the *.t2d
directories are stored there.
- clutter can be reduced to zero using, e.g., --build-dir=/tmp/\$USER.t2d
or --build-dir=\$HOME/.t2d.
- the output file is updated after every successful TeX run, for
sake of concurrent visualization of the output. In a \`local' build
the viewer stops during the whole TeX run.
- if the compilation fails, the previous state of the output file
is preserved.
- PDF and DVI compilation are kept in separate subdirectories
preventing any possibility of auxiliary file incompatibility.
On the other hand, because \`tidy' compilation takes place in another
directory, occasionally TeX won't be able to find some files (e.g., when
using \\graphicspath): in that case, use -I to specify the additional
directories to consider.
The values of the BIBER, BIBTEX, DVIPDF, DVIPS, HEVEA, LATEX, MAKEINDEX,
MAKEINFO, PDFLATEX, PDFTEX, SED, T4HT, TEX, TEX4HT, TEXINDEX, and THUMBPDF_CMD
environment variables are used to run those commands, if they are set.
Regarding --dvipdf, if DVIPDF is not set in the environment, the
following programs are looked for (in this order): dvipdfmx dvipdfm
dvipdf dvi2pdf dvitopdf.
Any CMD strings are added after @setfilename for Texinfo input, or in
the first line for LaTeX input.
Report bugs to bug-texinfo@gnu.org,
general questions and discussion to help-texinfo@gnu.org.
GNU Texinfo home page: <http://www.gnu.org/software/texinfo/>
General help using GNU software: <http://www.gnu.org/gethelp/>
EOF
exit 0
}
# verbose WORD1 WORD2
# -------------------
# Report some verbose information.
verbose ()
{
if $verb; then
echo >&2 "$0: $@"
fi
}
# version
# -------
# Display version info and exit successfully.
version ()
{
cat <<EOF
texi2dvi (GNU Texinfo 5.2) $rcs_version
Copyright (C) 2014 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
EOF
exit 0
}
## ---------------- ##
## Handling lists. ##
## ---------------- ##
# list_append LIST-NAME ELEM
# --------------------------
# Set LIST-NAME to its former contents, with ELEM appended.
list_append ()
{
local la_l="$1"
shift
eval set X \$$la_l "$@"
shift
eval $la_l=\""$@"\"
}
# list_concat_dirs LIST-NAME DIR-LIST
# -----------------------------------
# Append to LIST-NAME all the components (included empty) from
# the $path_sep separated list DIR-LIST. Make the paths absolute.
list_concat_dirs ()
{
local lcd_list="$1"
# Empty path components are meaningful to tex. We rewrite them as
# `EMPTY' so they don't get lost when we split on $path_sep.
# Hopefully no one will have an actual directory named EMPTY.
local replace_EMPTY="-e 's/^$path_sep/EMPTY$path_sep/g' \
-e 's/$path_sep\$/${path_sep}EMPTY/g' \
-e 's/$path_sep$path_sep/${path_sep}EMPTY:/g'"
save_IFS=$IFS
IFS=$path_sep
set x `echo "$2" | eval $SED $replace_EMPTY`; shift
IFS=$save_IFS
local dir
for dir
do
case $dir in
EMPTY)
list_append $lcd_list ""
;;
*)
if test -d $dir; then
dir=`absolute "$dir"`
list_append $lcd_list "$dir"
fi
;;
esac
done
}
# list_prefix LIST-NAME SEP -> STRING
# -----------------------------------
# Return a string that is composed of the LIST-NAME with each item
# preceded by SEP.
list_prefix ()
{
local lp_p="$2"
eval set X \$$1
shift
local lp_res
for i
do
lp_res="$lp_res \"$lp_p\" \"$i\""
done
echo "$lp_res"
}
# list_infix LIST-NAME SEP -> STRING
# ----------------------------------
# Same as list_prefix, but a separator.
list_infix ()
{
eval set X \$$1
shift
local la_IFS="$IFS"
IFS=$path_sep
echo "$*"
IFS=$la_IFS
}
# list_dir_to_abs LIST-NAME
# -------------------------
# Convert the list to using only absolute dir names.
# Currently unused, but should replace absolute_filenames some day.
list_dir_to_abs ()
{
local ld_l="$1"
eval set X \$$ld_l
shift
local ld_res
for dir
do
dir=`absolute "$dir"`
test -d "$dir" || continue
ld_res="$ld_res \"$dir\""
done
set X $ld_res; shift
eval $ld_l=\"$@\"
}
## ------------------------------ ##
## Language auxiliary functions. ##
## ------------------------------ ##
# out_lang_set LANG
# -----------------
out_lang_set ()
{
case $1 in
dvi|dvipdf|html|info|pdf|ps|text) out_lang=$1;;
*) error 1 "invalid output format: $1";;
esac
}
# out_lang_tex
# ------------
# Return the tex output language (DVI or PDF) for $OUT_LANG.
out_lang_tex ()
{
case $out_lang in
dvi | ps | dvipdf ) echo dvi;;
pdf ) echo $out_lang;;
html | info | text ) echo $out_lang;;
*) error 1 "invalid out_lang: $1";;
esac
}
# out_lang_ext
# ------------
# Return the extension for $OUT_LANG.
out_lang_ext ()
{
case $out_lang in
dvipdf ) echo pdf;;
dvi | html | info | pdf | ps | text ) echo $out_lang;;
*) error 1 "invalid out_lang: $1";;
esac
}
## ------------------------- ##
## TeX auxiliary functions. ##
## ------------------------- ##
# Save TEXINPUTS so we can construct a new TEXINPUTS path for each file.
# Likewise for bibtex and makeindex.
tex_envvars="BIBINPUTS BSTINPUTS DVIPSHEADERS INDEXSTYLE MFINPUTS MPINPUTS \
TEXINPUTS TFMFONTS"
for var in $tex_envvars; do
eval ${var}_orig=\$$var
export $var
done
# absolute_filenames TEX-PATH -> TEX-PATH
# ---------------------------------------
# Convert relative paths to absolute paths, so we can run in another
# directory (e.g., in tidy build mode, or during the macro-support
# detection). Prepend ".".
absolute_filenames ()
{
# Empty path components are meaningful to tex. We rewrite them as
# `EMPTY' so they don't get lost when we split on $path_sep.
# Hopefully no one will have an actual directory named EMPTY.
local replace_empty="-e 's/^$path_sep/EMPTY$path_sep/g' \
-e 's/$path_sep\$/${path_sep}EMPTY/g' \
-e 's/$path_sep$path_sep/${path_sep}EMPTY:/g'"
local res
res=`echo "$1" | eval $SED $replace_empty`
save_IFS=$IFS
IFS=$path_sep
set x $res; shift
res=.
for dir
do
case $dir in
EMPTY)
res=$res$path_sep
;;
*)
if test -d "$dir"; then
res=$res$path_sep`absolute "$dir"`
else
# Even if $dir is not a directory, preserve it in the path.
# It might contain metacharacters that TeX will expand in
# turn, e.g., /some/path/{a,b,c}. This will not get the
# implicit absolutification of the path, but we can't help that.
res=$res$path_sep$dir
fi
;;
esac
done
echo "$res"
}
# output_base_name FILE
# ---------------------
# The name of FILE, possibly renamed to satisfy --output.
# FILE is local, there is no directory part.
output_base_name ()
{
case $oname in
'') echo "$1";;
*) local out_noext
out_noext=`noext "$oname"`
local file_ext
file_ext=`echo "$1" | $SED 's/^.*\.//'`
echo "$out_noext.$file_ext"
;;
esac
}
# destdir
# -------
# Return the name of the directory where the output is expected.
destdir ()
{
case $oname in
'') echo "$orig_pwd";;
*) dirname "$oname";;
esac
}
# move_to_dest FILE...
# --------------------
# Move FILE to the place where the user expects it. Truly move it, that
# is, it must not remain in its build location unless that is also the
# output location. (Otherwise it might appear as an extra file in make
# distcheck.)
#
# FILE can be the principal output (in which case -o directly applies), or
# an auxiliary file with the same base name.
move_to_dest ()
{
# echo "move_to_dest $*, tidy=$tidy, oname=$oname"
# If we built in place and have no output name, there is nothing to
# do, so just return.
case $tidy:$oname in
false:) return;;
esac
local destfile
local destdir
local destbase
local sourcedir
local sourcebase
for file
do
test -f "$file" \
|| error 1 "no such file or directory: $file"
case $tidy:$oname in
true:) destdir=$orig_pwd
destfile=$destdir/$file;;
true:*) destfile=`output_base_name "$file"`
destdir=`dirname "$destfile"`;;
false:*) destfile=$oname
destdir=`dirname "$destfile"`;;
esac
# We want to compare the source location and the output location,
# and if they are different, do the move. But if they are the
# same, we must preserve the source. Since we can't assume
# stat(1) or test -ef is available, resort to comparing the
# directory names, canonicalized with pwd. We can't use cmp -s
# since the output file might not actually change from run to run;
# e.g., TeX DVI output is timestamped to only the nearest minute.
destdir=`cd "$destdir" && pwd`
destbase=`basename "$destfile"`
sourcedir=`dirname "$file"`
sourcedir=`cd "$sourcedir" && pwd`
sourcebase=`basename "$file"`
if test "$sourcedir/$sourcebase" != "$destdir/$destbase"; then
verbose "Moving $file to $destfile"
rm -f "$destfile"
mv "$file" "$destfile"
fi
done
}
## --------------------- ##
## Managing xref files. ##
## --------------------- ##
# aux_file_p FILE
# ---------------
# Return with success if FILE is an aux file.
aux_file_p ()
{
test -f "$1" || return 1
case $1 in
*.aux) return 0;;
*) return 1;;
esac
}
# bibaux_file_p FILE
# ------------------
# Return with success if FILE is an aux file containing citation
# requests.
bibaux_file_p ()
{
test -s "$1" || return 1
if (grep '^\\bibstyle[{]' "$1" \
&& grep '^\\bibdata[{]' "$1" \
## The following line is suspicious: fails when there
## are citations in sub aux files. We need to be
## smarter in this case.
## && grep '^\\citation[{]' "$f"
) >&6 2>&1;
then
return 0
fi
return 1
}
# index_file_p FILE
# -----------------
# Return with success if FILE is an index file.
index_file_p ()
{
test -f "$1" || return 1
case $in_lang:$latex2html:`out_lang_tex`:`$SED '1q' "$1"` in
# When working with TeX4HT, *.idx are created by LaTeX. They must
# be processed to produce *.4ix, *.4dx files. The *.4dx file is
# passed to makeindex to produce the *.ind file. This sequence is
# handled by run_index, so we are only interested in the *.idx
# files, which have each "\indexentry" preceded by a
# "\beforeentry".
latex:tex4ht:html:"\\beforeentry {"*) return 0;;
# When index.sty is used, there is a space before the brace.
latex:*:*:"\\indexentry{"*|latex:*:*:"\\indexentry {"*) return 0;;
texinfo:*:*:"\\entry{"*) return 0;;
*) return 1;;
esac
}
# xref_file_p FILE
# ----------------
# Return with success if FILE is an xref file (indexes, tables and lists).
xref_file_p ()
{
test -f "$1" || return 1
# If the file is not suitable to be an index or xref file, don't
# process it. It's suitable if the first character is a
# backslash or right quote or at, as long as the first line isn't
# \input texinfo.
case `$SED '1q' "$1"` in
"\\input texinfo"*) return 1;;
[\\''@]*) return 0;;
*) return 1;;
esac
}
# generated_files_get FILENAME-NOEXT [PREDICATE-FILTER]
# -----------------------------------------------------
# Return the list of files generated by the TeX compilation of FILENAME-NOEXT.
generated_files_get ()
{
local filter=true
if test -n "$2"; then
filter=$2
fi
# Gather the files created by TeX.
(
if test -f "$1.log"; then
$SED -n -e "s,^\\\\openout.* = \`\\(.*\\)'\\.,\\1,p" "$1.log"
fi
echo "$1.log"
) |
# Depending on these files, infer outputs from other tools.
while read file; do
echo $file
case $in_lang in
texinfo)
# texindex: texinfo.cp -> texinfo.cps
if index_file_p $file; then
echo ${file}s
fi
;;
latex)
if aux_file_p $file; then
# bibtex: *.aux -> *.bbl and *.blg.
echo $file | $SED 's/^\(.*\)\.aux$/\1.bbl/'
echo $file | $SED 's/^\(.*\)\.aux$/\1.blg/'
# -recorder: .fls
echo $file | $SED 's/^\(.*\)\.aux$/\1.fls/'
fi
;;
esac
done |
# Filter existing files matching the criterion.
#
# With an input file name containing a space, this produces a
# "command not found" message (and filtering is ineffective).
# The situation with a newline is presumably even worse.
while read file; do
if $filter "$file"; then
echo $file
fi
done |
sort |
# Some files are opened several times, e.g., listings.sty's *.vrb.
uniq
}
# xref_files_save
# ---------------
# Save the xref files.
xref_files_save ()
{
# Save copies of auxiliary files for later comparison.
xref_files_orig=`generated_files_get "$in_noext" xref_file_p`
if test -n "$xref_files_orig"; then
verbose "Backing up xref files: $xref_files_orig"
# The following line improves `cp $xref_files_orig "$work_bak"'
# by preserving the directory parts. Think of
# cp chap1/main.aux chap2/main.aux $work_bak.
#
# Users may have, e.g., --keep-old-files. Don't let this interfere.
# (Don't use unset for the sake of ancient shells.)
TAR_OPTIONS=; export TAR_OPTIONS
tar cf - $xref_files_orig | (cd "$work_bak" && tar xf -)
fi
}
# xref_files_changed
# ------------------
# Whether the xref files were changed since the previous run.
xref_files_changed ()
{
# LaTeX (and the package changebar) report in the LOG file if it
# should be rerun. This is needed for files included from
# subdirs, since texi2dvi does not try to compare xref files in
# subdirs. Performing xref files test is still good since LaTeX
# does not report changes in xref files.
if grep "Rerun to get" "$in_noext.log" >&6 2>&1; then
return 0
fi
# biblatex report of whether rerunning is needed.
if grep "biblatex.*(re)run" "$in_noext.log" >&6 2>&1; then
return 0
fi
# If old and new lists don't have the same file list,
# then something has definitely changed.
xref_files_new=`generated_files_get "$in_noext" xref_file_p`
verbose "Original xref files = $xref_files_orig"
verbose "New xref files = $xref_files_new"
if test "x$xref_files_orig" != "x$xref_files_new"; then
return 0
fi
# Compare each file until we find a difference.
for this_file in $xref_files_new; do
verbose "Comparing xref file `echo $this_file | $SED 's|\./||g'` ..."
# cmp -s returns nonzero exit status if files differ.
if cmp -s "$this_file" "$work_bak/$this_file"; then :; else
verbose "xref file `echo $this_file | $SED 's|\./||g'` differed ..."
if $debug; then
diff -u "$work_bak/$this_file" "$this_file"
fi
return 0
fi
done
# No change.
return 1
}
## ----------------------- ##
## Running the TeX suite. ##
## ----------------------- ##
# run_tex ()
# ----------
# Run TeX as "$tex $in_input", taking care of errors and logs.
run_tex ()
{
case $in_lang:$latex2html:`out_lang_tex` in
latex:*:dvi|latex:tex4ht:html)
tex=${LATEX:-latex};;
latex:*:pdf)
tex=${PDFLATEX:-pdflatex};;
texinfo:*:dvi)
# MetaPost also uses the TEX environment variable. If the user
# has set TEX=latex for that reason, don't bomb out.
case $TEX in
*latex) tex=tex;; # don't bother trying to find etex
*) tex=$TEX
esac;;
texinfo:*:pdf) tex=$PDFTEX;;
*) error 1 "$out_lang not supported for $in_lang";;
esac
# do the special catcode trick for ~ in filenames only for Texinfo,
# not LaTeX.
if test x"$in_lang" = xtexinfo && test $catcode_special = maybe; then
catcode_special=true
else
catcode_special=false
fi
# Beware of aux files in subdirectories that require the
# subdirectory to exist.
case $in_lang:$tidy in
latex:true)
$SED -n 's|^[ ]*\\include{\(.*\)/.*}.*|\1|p' "$in_input" |
sort -u |
while read d
do
ensure_dir "$work_build/$d"
done
;;
esac
# Note that this will be used via an eval: quote properly.
local cmd="$tex"
# If possible, make TeX report error locations in GNU format.
if $line_error; then
if test "${tex_help:+set}" != set; then
# Go to a temporary directory to try --help, since old versions that
# don't accept --help will generate a texput.log.
tex_help_dir=$t2ddir/tex_help
ensure_dir "$tex_help_dir"
tex_help=`cd "$tex_help_dir" >&6 && $tex --help </dev/null 2>&1 || true`
fi
# The mk program and perhaps others want to parse TeX's
# original error messages.
case $tex_help in
*file-line-error*) cmd="$cmd --file-line-error";;
esac
fi
# Tell TeX about TCX file, if specified.
test -n "$translate_file" && cmd="$cmd --translate-file=$translate_file"
# Tell TeX to make source specials (for backtracking from output to
# source, given a sufficiently smart editor), if specified.
test -n "$src_specials" && cmd="$cmd $src_specials"
# Tell TeX to allow running external executables
test -n "$shell_escape" && cmd="$cmd $shell_escape"
# Tell TeX to be batch if requested.
if $batch; then
# \batchmode does not show terminal output at all, so we don't
# want that. And even in batch mode, TeX insists on having input
# from the user. Close its stdin to make it impossible.
cmd="$cmd </dev/null '${escape}nonstopmode'"
fi
# we'd like to handle arbitrary input file names, especially
# foo~bar/a~b.tex, since Debian likes ~ characters.
if $catcode_special; then
# $normaltilde is just to reduce line length in this source file.
# The idea is to define \normaltilde as a catcode other ~ character,
# then make the active ~ be equivalent to that, instead of the plain
# TeX tie. Then when the active ~ appears in the filename, it will
# be expanded to itself, as far as \input will see. (This is the
# same thing that texinfo.tex does in general, BTW.)
normaltilde="${escape}catcode126=12 ${escape}def${escape}normaltilde{~}"
cmd="$cmd '$normaltilde${escape}catcode126=13 ${escape}let~\normaltilde '"
fi
# Other special (non-active) characters could be supported by
# resetting their catcodes to other on the command line and changing
# texinfo.tex to initialize everything to plain catcodes. Maybe someday.
# append the \input command.
cmd="$cmd '${escape}input'"
# TeX's \input does not (easily or reliably) support whitespace
# characters or other special characters in file names. Our intensive
# use of absolute file names makes this worse: the enclosing directory
# names may include white spaces. Improve the situation using a
# symbolic link to the filename in the current directory, in tidy mode
# only. Do not alter in_input.
#
# The filename is almost always tokenized using plain TeX conventions
# (the exception would be if the user made a texinfo.fmt file). Not
# all the plain TeX special characters cause trouble, but there's no
# harm in making the link.
#
case $tidy:`func_dirname "$in_input"` in
true:*["$space$tab$newline\"#\$%\\^_{}~"]*)
_run_tex_file_name=`basename "$in_input"`
if test ! -f "$_run_tex_file_name"; then
# It might not be a file, clear it.
run rm -f "$_run_tex_file_name"
run ln -s "$in_input"
fi
cmd="$cmd '$_run_tex_file_name'"
;;
*)
cmd="$cmd '$in_input'"
;;
esac
verbose "$0: Running $cmd ..."
if eval "$cmd" >&5; then
case $out_lang in
dvi | pdf ) move_to_dest "$in_noext.$out_lang";;
esac
else
error 1 "$tex exited with bad status, quitting."
fi
}
# run_bibtex ()
# -------------
# Run bibtex on (or biber) current file.
# - If its input (AUX) exists.
# - If some citations are missing (LOG contains `Citation').
# or the LOG complains of a missing .bbl
#
# Don't try to be too smart:
# 1. Running bibtex only if the bbl file exists and is older than
# the LaTeX file is wrong, since the document might include files
# that have changed.
#
# 3. Because there can be several AUX (if there are \include's),
# but a single LOG, looking for missing citations in LOG is
# easier, though we take the risk of matching false messages.
run_bibtex ()
{
case $in_lang in
latex) bibtex=${BIBTEX:-bibtex};;
texinfo) return;;
esac
# "Citation undefined" is for LaTeX, "Undefined citation" for btxmac.tex.
# The no .aux && \bibdata test is also for btxmac, in case it was the
# first run of a bibtex-using document. Otherwise, it's possible that
# bibtex would never be run.
if test -r "$in_noext.aux" \
&& test -r "$in_noext.log" \
&& ( (grep 'Warning:.*Citation.*undefined' "$in_noext.log" \
|| grep '.*Undefined citation' "$in_noext.log" \
|| grep 'No file .*\.bbl\.' "$in_noext.log") \
|| (grep 'No \.aux file' "$in_noext.log" \
&& grep '^\\bibdata' "$in_noext.aux") ) \
>&6 2>&1; \
then
bibtex_aux=`generated_files_get "$in_noext" bibaux_file_p`
for f in $bibtex_aux; do
run $bibtex "$f"
done
fi
# biber(+biblatex) check.
if test -r "$in_noext.bcf" \
&& grep '</bcf:controlfile>' "$in_noext.bcf" >/dev/null; then
run ${BIBER:-biber} "$in_noext"
fi
}
# run_index ()
# ------------
# Run texindex (or makeindex or texindy) on current index files. If
# they already exist, and after running TeX a first time the index
# files don't change, then there's no reason to run TeX again. But we
# won't know that if the index files are out of date or nonexistent.
run_index ()
{
local index_files
index_files=`generated_files_get $in_noext index_file_p`
test -n "$index_files" \
|| return 0
: ${MAKEINDEX:=makeindex}
: ${TEXINDEX:=texindex}
: ${TEXINDY:=texindy}
local index_file
local index_noext
case $in_lang:$latex2html:`out_lang_tex` in
latex:tex4ht:html)
for index_file in $index_files
do
index_noext=`noext "$index_file"`
run tex \
'\def\filename{{'"$index_noext"'}{idx}{4dx}{ind}}
\input idxmake.4ht'
run $MAKEINDEX -o $index_noext.ind $index_noext.4dx
done
;;
latex:*)
if $TEXINDY --version >&6 2>&1; then
run $TEXINDY $index_files
else
run $MAKEINDEX $index_files
fi
;;
texinfo:*)
run $TEXINDEX $index_files
;;
esac
}
# run_tex4ht ()
# -------------
# Run the last two phases of TeX4HT: tex4ht extracts the HTML from the
# instrumented DVI file, and t4ht converts the figures and installs
# the files when given -d.
#
# Because knowing exactly which files are created is complex (in
# addition the names are not simple to compute), which makes it
# difficult to install the output files in a second step, it is much
# simpler to install directly the output files.
run_tex4ht ()
{
case $in_lang:$latex2html:`out_lang_tex` in
latex:tex4ht:html)
: ${TEX4HT:=tex4ht} ${T4HT:=t4ht}
run "$TEX4HT" "-f/$in_noext"
# Do not remove the / after the destdir.
run "$T4HT" "-d`destdir`/" "-f/$in_noext"
;;
esac
}
# run_thumbpdf ()
# ---------------
run_thumbpdf ()
{
if test `out_lang_tex` = pdf \
&& test -r "$in_noext.log" \
&& grep 'thumbpdf\.sty' "$in_noext.log" >&6 2>&1; \
then
thumbpdf=${THUMBPDF_CMD:-thumbpdf}
thumbcmd="$thumbpdf $in_dir/$in_noext"
verbose "Running $thumbcmd ..."
if $thumbcmd >&5; then
run_tex
else
report "$thumbpdf exited with bad status." \
"Ignoring its output."
fi
fi
}
# run_dvipdf FILE.dvi
# -------------------
# Convert FILE.dvi to FILE.pdf.
run_dvipdf ()
{
# Find which dvi->pdf program is available.
if test -z "$dvipdf"; then
for i in "$DVIPDF" dvipdfmx dvipdfm dvipdf dvi2pdf dvitopdf; do
if findprog $i; then
dvipdf=$i
fi
done
fi
# These tools have varying interfaces, some 'input output', others
# 'input -o output'. They all seem to accept 'input' only,
# outputting using the expected file name.
run $dvipdf "$1"
if test ! -f `echo "$1" | $SED -e 's/\.dvi$/.pdf/'`; then
error 1 "cannot find output file"
fi
}
# run_tex_suite ()
# ----------------
# Run the TeX tools until a fix point is reached.
run_tex_suite ()
{
# Move to the working directory.
if $tidy; then
verbose "cd $work_build"
cd "$work_build" || exit 1
fi
# Count the number of cycles.
local cycle=0
while :; do
# check for probably LaTeX loop (e.g. varioref)
if test $cycle -eq "$max_iters"; then
error 0 "Maximum of $max_iters cycles exceeded"
break
fi
# report progress
cycle=`expr $cycle + 1`
verbose "Cycle $cycle for $command_line_filename"
xref_files_save
# We run bibtex first, because it's more likely for the indexes
# to change after bibtex is run than the reverse, though either
# would be rare.
run_bibtex
run_index
run_core_conversion
xref_files_changed || break
done
# If we were using thumbpdf and producing PDF, then run thumbpdf
# and TeX one last time.
run_thumbpdf
# If we are using tex4ht, call it.
run_tex4ht
# Install the result if we didn't already (i.e., if the output is
# dvipdf or ps).
case $latex2html:$out_lang in
*:dvipdf)
run_dvipdf "$in_noext.`out_lang_tex`"
move_to_dest "$in_noext.`out_lang_ext`"
;;
*:ps)
: ${DVIPS:=dvips}
run $DVIPS -o "$in_noext.`out_lang_ext`" "$in_noext.`out_lang_tex`"
move_to_dest "$in_noext.`out_lang_ext`"
;;
esac
cd_orig
}
## -------------------------------- ##
## TeX processing auxiliary tools. ##
## -------------------------------- ##
# A sed script that preprocesses Texinfo sources in order to keep the
# iftex sections only. We want to remove non-TeX sections, and comment
# (with `@c _texi2dvi') TeX sections so that makeinfo does not try to
# parse them. Nevertheless, while commenting TeX sections, don't
# comment @macro/@end macro so that makeinfo does propagate them.
# Unfortunately makeinfo --iftex --no-ifinfo doesn't work well enough
# (yet), makeinfo can't parse the TeX commands, so work around with sed.
#
# We assume that `@c _texi2dvi' starting a line is not present in the
# document.
#
comment_iftex=\
'/^@tex/,/^@end tex/{
s/^/@c _texi2dvi/
}
/^@iftex/,/^@end iftex/{
s/^/@c _texi2dvi/
/^@c _texi2dvi@macro/,/^@c _texi2dvi@end macro/{
s/^@c _texi2dvi//
}
}
/^@ifnottex/,/^@end ifnottex/{
s/^/@c (_texi2dvi)/
}
/^@ifinfo/,/^@end ifinfo/{
/^@node/p
/^@menu/,/^@end menu/p
t
s/^/@c (_texi2dvi)/
}
s/^@ifnotinfo/@c _texi2dvi@ifnotinfo/
s/^@end ifnotinfo/@c _texi2dvi@end ifnotinfo/'
# Uncommenting is simpler: remove any leading `@c texi2dvi'; repeated
# copies can sneak in via macro invocations.
uncomment_iftex='s/^@c _texi2dvi\(@c _texi2dvi\)*//'
# run_makeinfo ()
# ---------------
# Expand macro commands in the original source file using Makeinfo.
# Always use `end' footnote style, since the `separate' style
# generates different output (arguably this is a bug in -E). Discard
# main info output, the user asked to run TeX, not makeinfo.
run_makeinfo ()
{
test $in_lang = texinfo \
|| return 0
# Unless required by the user, makeinfo expansion is wanted only
# if texinfo.tex is too old.
if $expand; then
makeinfo=${MAKEINFO:-makeinfo}
else
# Check if texinfo.tex performs macro expansion by looking for
# its version. The version is a date of the form YEAR-MO-DA.
# We don't need to use [0-9] to match the digits since anyway
# the comparison with $txiprereq, a number, will fail with non-digits.
# Run in a temporary directory to avoid leaving files.
version_test_dir=$t2ddir/version_test
ensure_dir "$version_test_dir"
if (
cd "$version_test_dir"
echo '\input texinfo.tex @bye' >txiversion.tex
# Be sure that if tex wants to fail, it is not interactive:
# close stdin.
$TEX txiversion.tex </dev/null >txiversion.out 2>txiversion.err
); then :; else
report "texinfo.tex appears to be broken.
This may be due to the environment variable TEX set to something
other than (plain) tex, a corrupt texinfo.tex file, or
to tex itself simply not working."
cat "$version_test_dir/txiversion.out"
cat "$version_test_dir/txiversion.err" >&2
error 1 "quitting."
fi
eval `$SED -n 's/^.*\[\(.*\)version \(....\)-\(..\)-\(..\).*$/txiformat=\1 txiversion="\2\3\4"/p' "$version_test_dir/txiversion.out"`
verbose "texinfo.tex preloaded as \`$txiformat', version is \`$txiversion' ..."
if test "$txiprereq" -le "$txiversion" >&6 2>&1; then
makeinfo=
else
makeinfo=${MAKEINFO:-makeinfo}
fi
# If TeX is preloaded, offer the user this convenience:
if test "$txiformat" = Texinfo; then
escape=@
fi
fi
if test -n "$makeinfo"; then
# in_src: the file with macros expanded.
# Use the same basename to generate the same aux file names.
work_src=$workdir/src
ensure_dir "$work_src"
in_src=$work_src/$in_base
local miincludes
miincludes=`list_prefix includes -I`
verbose "Macro-expanding $command_line_filename to $in_src ..."
# eval $makeinfo because it might be defined as something complex
# (running missing) and then we end up with things like '"-I"',
# and "-I" (including the quotes) is not an option name. This
# happens with gettext 0.14.5, at least.
$SED "$comment_iftex" "$command_line_filename" \
| eval $makeinfo --footnote-style=end -I "$in_dir" $miincludes \
-o /dev/null --macro-expand=- \
| $SED "$uncomment_iftex" >"$in_src"
# Continue only if everything succeeded.
if test $? -ne 0 \
|| test ! -r "$in_src"; then
verbose "Expansion failed, ignored...";
else
in_input=$in_src
fi
fi
}
# insert_commands ()
# ------------------
# Used most commonly for @finalout, @smallbook, etc.
insert_commands ()
{
if test -n "$textra"; then
# _xtr. The file with the user's extra commands.
work_xtr=$workdir/xtr
in_xtr=$work_xtr/$in_base
ensure_dir "$work_xtr"
verbose "Inserting extra commands: $textra"
local textra_cmd
case $in_lang in
latex) textra_cmd=1i;;
texinfo) textra_cmd='/^@setfilename/a';;
*) error 1 "internal error, unknown language: $in_lang";;
esac
$SED "$textra_cmd\\
$textra" "$in_input" >"$in_xtr"
in_input=$in_xtr
fi
case $in_lang:$latex2html:`out_lang_tex` in
latex:tex4ht:html)
# _tex4ht. The file with the added \usepackage{tex4ht}.
work_tex4ht=$workdir/tex4ht
in_tex4ht=$work_tex4ht/$in_base
ensure_dir "$work_tex4ht"
verbose "Inserting \\usepackage{tex4ht}"
perl -pe 's<\\documentclass(?:\[.*\])?{.*}>
<$&\\usepackage[xhtml]{tex4ht}>' \
"$in_input" >"$in_tex4ht"
in_input=$in_tex4ht
;;
esac
}
# compute_language FILENAME
# -------------------------
# Return the short string describing the language in which FILENAME
# is written: `texinfo' or `latex'.
compute_language ()
{
# If the user explicitly specified the language, use that.
# Otherwise, if the first line is \input texinfo, assume it's texinfo.
# Otherwise, guess from the file extension.
if test -n "$set_language"; then
echo $set_language
elif $SED 1q "$1" | grep 'input texinfo' >&6; then
echo texinfo
else
# Get the type of the file (latex or texinfo) from the given language
# we just guessed, or from the file extension if not set yet.
case $1 in
*.ltx | *.tex | *.drv | *.dtx) echo latex;;
*) echo texinfo;;
esac
fi
}
# run_hevea (MODE)
# ----------------
# Convert to HTML/INFO/TEXT.
#
# Don't pass `-noiso' to hevea: it's useless in HTML since anyway the
# charset is set to latin1, and troublesome in other modes since
# accented characters loose their accents.
#
# Don't pass `-o DEST' to hevea because in that case it leaves all its
# auxiliary files there too... Too bad, because it means we will need
# to handle images some day.
run_hevea ()
{
local hevea="${HEVEA:-hevea}"
local run_hevea="$hevea"
case $1 in
html) ;;
text|info) run_hevea="$run_hevea -$1";;
*) error 1 "run_hevea: invalid argument: $1";;
esac
# Compiling to the tmp directory enables to preserve a previous
# successful compilation.
run_hevea="$run_hevea -fix -O -o '$out_base'"
run_hevea="$run_hevea `list_prefix includes -I` -I '$orig_pwd' "
run_hevea="$run_hevea '$in_input'"
if $debug; then
run_hevea="$run_hevea -v -v"
fi
verbose "running $run_hevea"
if eval "$run_hevea" >&5; then
# hevea leaves trailing white spaces, this is annoying.
case $1 in text|info)
perl -pi -e 's/[ \t]+$//g' "$out_base"*;;
esac
case $1 in
html|text) move_to_dest "$out_base";;
info) # There can be foo.info-1, foo.info-2 etc.
move_to_dest "$out_base"*;;
esac
else
error 1 "$hevea exited with bad status, quitting."
fi
}
# run_core_conversion ()
# ----------------------
# Run the TeX (or HeVeA).
run_core_conversion ()
{
case $in_lang:$latex2html:`out_lang_tex` in
*:dvi|*:pdf|latex:tex4ht:html)
run_tex;;
latex:*:html|latex:*:text|latex:*:info)
run_hevea $out_lang;;
*)
error 1 "invalid input/output combination: $in_lang/$out_lang";;
esac
}
# compile ()
# ----------
# Run the full compilation chain, from pre-processing to installation
# of the output at its expected location.
compile ()
{
# Source file might include additional sources.
# We want `.:$orig_pwd' before anything else. (We'll add `.:' later
# after all other directories have been turned into absolute paths.)
# `.' goes first to ensure that any old .aux, .cps,
# etc. files in ${directory} don't get used in preference to fresher
# files in `.'. Include orig_pwd in case we are in clean build mode, where
# we have cd'd to a temp directory.
common="$orig_pwd$path_sep$in_dir$path_sep"
#
# If we have any includes, put those at the end.
# Keep a final path_sep to get the default (system) TeX directories included.
txincludes=`list_infix includes $path_sep`
test -n "$txincludes" && common="$common$txincludes$path_sep"
#
for var in $tex_envvars; do
eval val="\$common\$${var}_orig"
# Convert relative paths to absolute paths, so we can run in another
# directory (e.g., in clean build mode, or during the macro-support
# detection). ".:" is added here.
val=`absolute_filenames "$val"`
eval $var="\"$val\""
export $var
eval verbose \"$var=\'\$${var}\'\"
done
# --expand
run_makeinfo
# --command, --texinfo
insert_commands
# Run until a fix point is reached.
run_tex_suite
}
# remove FILES
# ------------
remove ()
{
verbose "Removing" "$@"
rm -rf "$@"
}
# mostly_clean
# ------------
# Remove auxiliary files and directories. Changes the current directory.
mostly_clean ()
{
cd_orig
set X "$t2ddir"
shift
$tidy || {
local log="$work_build/$in_noext.log"
set X ${1+"$@"} "$log" `generated_files_get "$work_build/$in_noext"`
shift
}
remove ${1+"$@"}
}
# cleanup ()
# ----------
# Remove what should be removed according to options.
# Called at the end of each compilation cycle, and at the end of
# the script. Changes the current directory.
cleanup ()
{
case $build_mode in
local) cd_orig; remove "$t2ddir";;
clean) mostly_clean;;
tidy) ;;
esac
}
## ---------------------- ##
## Command line parsing. ##
## ---------------------- ##
# Push a token among the arguments that will be used to notice when we
# ended options/arguments parsing.
# Use "set dummy ...; shift" rather than 'set - ..." because on
# Solaris set - turns off set -x (but keeps set -e).
# Use ${1+"$@"} rather than "$@" because Digital Unix and Ultrix 4.3
# still expand "$@" to a single argument (the empty string) rather
# than nothing at all.
arg_sep="$$--$$"
set dummy ${1+"$@"} "$arg_sep"; shift
#
# Parse command line arguments.
while test x"$1" != x"$arg_sep"; do
# Handle --option=value by splitting apart and putting back on argv.
case "$1" in
--*=*)
opt=`echo "$1" | $SED -e 's/=.*//'`
val=`echo "$1" | $SED -e 's/[^=]*=//'`
shift
set dummy "$opt" "$val" ${1+"$@"}; shift
;;
esac
case "$1" in
-@ ) escape=@;;
-~ ) catcode_special=false;;
# Silently and without documentation accept -b and --b[atch] as synonyms.
-b | --batch) batch=true;;
--build) shift; build_mode=$1;;
--build-dir) shift; build_dir=$1; build_mode=tidy;;
-c | --clean) build_mode=clean;;
-D | --debug) debug=true;;
-e | -E | --expand) expand=true;;
-h | --help) usage;;
-I) shift; list_concat_dirs includes "$1";;
-l | --lang | --language) shift; set_language=$1;;
--mostly-clean) action=mostly-clean;;
--no-line-error) line_error=false;;
--max-iterations) shift; max_iters=$1;;
-o | --out | --output)
shift
# Make it absolute, just in case we also have --clean, or whatever.
oname=`absolute "$1"`;;
# Output formats.
-O|--output-format) shift; out_lang_set "$1";;
--dvi|--dvipdf|--html|--info|--pdf|--ps|--text)
out_lang_set `echo "x$1" | $SED 's/^x--//'`;;
-p) out_lang_set pdf;;
-q | -s | --quiet | --silent) quiet=true; batch=true;;
--src-specials) src_specials=--src-specials;;
--shell-escape) shell_escape=--shell-escape;;
--tex4ht) latex2html=tex4ht;;
-t | --texinfo | --command ) shift; textra="$textra\\
"`echo "$1" | $SED 's/\\\\/\\\\\\\\/g'`;;
--translate-file ) shift; translate_file="$1";;
--tidy) build_mode=tidy;;
-v | --vers*) version;;
-V | --verb*) verb=true;;
--) # What remains are not options.
shift
while test x"$1" != x"$arg_sep"; do
set dummy ${1+"$@"} "$1"; shift
shift
done
break;;
-*)
error 1 "Unknown or ambiguous option \`$1'." \
"Try \`--help' for more information."
;;
*) set dummy ${1+"$@"} "$1"; shift;;
esac
shift
done
# Pop the token
shift
# $tidy: compile in a t2d directory.
# $clean: remove all the aux files.
case $build_mode in
local) clean=false; tidy=false;;
tidy) clean=false; tidy=true;;
clean) clean=true; tidy=true;;
*) error 1 "invalid build mode: $build_mode";;
esac
# Interpret remaining command line args as filenames.
case $# in
0)
error 2 "Missing file arguments." "Try \`--help' for more information."
;;
1) ;;
*)
if test -n "$oname"; then
error 2 "Can't use option \`--output' with more than one argument."
fi
;;
esac
# We can't do much without tex.
#
if findprog ${TEX:-tex}; then :; else cat <<EOM
You don't have a working TeX binary (${TEX:-tex}) installed anywhere in
your PATH, and texi2dvi cannot proceed without one. If you want to use
this script, you'll need to install TeX (if you don't have it) or change
your PATH or TEX environment variable (if you do). See the --help
output for more details.
For information about obtaining TeX, please see http://tug.org/texlive,
or do a web search for TeX and your operating system or distro.
EOM
exit 1
fi
# We want to use etex (or pdftex) if they are available, and the user
# didn't explicitly specify. We don't check for elatex and pdfelatex
# because (as of 2003), the LaTeX team has asked that new distributions
# use etex by default anyway.
#
# End up with the TEX and PDFTEX variables set to what we are going to use.
if test -z "$TEX"; then
if findprog etex; then TEX=etex; else TEX=tex; fi
fi
#
if test -z "$PDFTEX"; then
if findprog pdfetex; then PDFTEX=pdfetex; else PDFTEX=pdftex; fi
fi
# File descriptor usage:
# 0 standard input
# 1 standard output (--verbose messages)
# 2 standard error
# 3 some systems may open it to /dev/tty
# 4 used on the Kubota Titan
# 5 tools output (turned off by --quiet)
# 6 tracing/debugging (set -x output, etc.)
# Main tools' output (TeX, etc.) that TeX users are used to seeing.
#
# If quiet, discard, else redirect to the message flow.
if $quiet; then
exec 5>/dev/null
else
exec 5>&1
fi
# Enable tracing, and auxiliary tools output.
#
# This fd should be used where you'd typically use /dev/null to throw
# output away. But sometimes it is convenient to see that output (e.g.,
# from a grep) to aid debugging. Especially debugging at distance, via
# the user.
#
if $debug; then
exec 6>&1
set -vx
else
exec 6>/dev/null
fi
#
# input_file_name_decode
# ----------------------
# Decode COMMAND_LINE_FILENAME, and compute:
# - COMMAND_LINE_FILENAME clean of TeX commands
# - IN_DIR
# The directory to the input file, possibly absolute if needed.
# - IN_DIR_ABS
# The absolute directory of the input file.
# - IN_BASE
# The input file base name (no directory part).
# - IN_NOEXT
# The input file name without extensions (nor directory part).
# - IN_INPUT
# Defaults to COMMAND_LINE_FILENAME, but might change if the
# input is preprocessed. With directory, possibly absolute.
input_file_name_decode ()
{
# See if we are run from within AUC-Tex, in which case we are
# passed `\input{FOO.tex}' or even `\nonstopmode\input{FOO.tex}'.
case $command_line_filename in
*\\nonstopmode*)
batch=true;;
esac
case $command_line_filename in
*\\input{*}*)
# Let AUC-TeX error parser deal with line numbers.
line_error=false
command_line_filename=`\
expr X"$command_line_filename" : X'.*input{\([^}]*\)}'`
;;
esac
# If the COMMAND_LINE_FILENAME is not absolute (e.g., --debug.tex),
# prepend `./' in order to avoid that the tools take it as an option.
echo "$command_line_filename" | LC_ALL=C $EGREP '^(/|[A-Za-z]:/)' >&6 \
|| command_line_filename="./$command_line_filename"
# See if the file exists. If it doesn't we're in trouble since, even
# though the user may be able to reenter a valid filename at the tex
# prompt (assuming they're attending the terminal), this script won't
# be able to find the right xref files and so forth.
test -r "$command_line_filename" \
|| error 1 "cannot read $command_line_filename, skipping."
# Get the name of the current directory.
in_dir=`func_dirname "$command_line_filename"`
in_dir_abs=`absolute "$in_dir"`
# In a clean build, we `cd', so get an absolute file name.
if $tidy; then
in_dir=$in_dir_abs
fi
# Strip directory part but leave extension.
in_base=`basename "$command_line_filename"`
# Strip extension.
in_noext=`noext "$in_base"`
# The normalized file name to compile. Must always point to the
# file to actually compile (in case of recoding, macro-expansion etc.).
in_input=$in_dir/$in_base
# Compute the output file name.
if test x"$oname" != x; then
out_name=$oname
else
out_name=$in_noext.`out_lang_ext`
fi
out_dir=`func_dirname "$out_name"`
out_dir_abs=`absolute "$out_dir"`
out_base=`basename "$out_name"`
out_noext=`noext "$out_base"`
}
## -------------- ##
## TeXify files. ##
## -------------- ##
for command_line_filename
do
verbose "Processing $command_line_filename ..."
input_file_name_decode
# `texinfo' or `latex'?
in_lang=`compute_language "$command_line_filename"`
# An auxiliary directory used for all the auxiliary tasks involved
# in compiling this document.
case $build_dir in
'' | . ) t2ddir=$out_noext.t2d ;;
*) # Avoid collisions between multiple occurrences of the same
# file, so depend on the output path. Remove leading `./',
# at least to avoid creating a file starting with `.!', i.e.,
# an invisible file. The sed expression is fragile if the cwd
# has active characters. Transform / into ! so that we don't
# need `mkdir -p'. It might be something to reconsider.
t2ddir=$build_dir/`echo "$out_dir_abs/$out_noext.t2d" |
$SED "s,^$orig_pwd/,,;s,^\./,,;s,/,!,g"`
esac
# Remove it at exit if clean mode.
trap "cleanup" 0 1 2 15
ensure_dir "$build_dir" "$t2ddir"
# We will change directory, better work with an absolute path...
t2ddir=`absolute "$t2ddir"`
# Sometimes there are incompatibilities between auxiliary files for
# DVI and PDF. The contents can also change whether we work on PDF
# and/or DVI. So keep separate spaces for each.
workdir=$t2ddir/`out_lang_tex`
ensure_dir "$workdir"
# _build. In a tidy build, where the auxiliary files are output.
if $tidy; then
work_build=$workdir/build
else
work_build=.
fi
# _bak. Copies of the previous auxiliary files (another round is
# run if they differ from the new ones).
work_bak=$workdir/bak
# Make those directories.
ensure_dir "$work_build" "$work_bak"
case $action in
compile)
# Compile the document.
compile
cleanup
;;
mostly-clean)
mostly_clean
;;
esac
done
verbose "done."
exit 0 # exit successfully, not however we ended the loop.