diff options
Diffstat (limited to 'Build/source/texk/detex/detex-src/README')
| -rw-r--r-- | Build/source/texk/detex/detex-src/README | 195 |
1 files changed, 195 insertions, 0 deletions
diff --git a/Build/source/texk/detex/detex-src/README b/Build/source/texk/detex/detex-src/README new file mode 100644 index 00000000000..a310a7f3fd0 --- /dev/null +++ b/Build/source/texk/detex/detex-src/README @@ -0,0 +1,195 @@ +Detex - Version 2.8 + +Detex is a program to remove TeX constructs from a text file. It recognizes +the \input command. + +This program assumes it is dealing with LaTeX input if it sees the string +"\begin{document}" in the text. It recognizes the \include and \includeonly +commands. + +This directory contains the following files: + +README - you're looking at it. + +COPYRIGHT - NCSA/University of Illinois Open Source License + +Makefile - makefile for generating detex on a 4.2BSD Unix system. + +detex.1l - troff source for the detex manual page. + Assuming you have the -man macros, use "make man-page" to + generate it. + +detex.h - Various global definitions. These should be modified to suit + the local installation. + +detex.l - Lex and C source for the detex program. + +lexout.c - C code generated from detex.l using lex on a Sun (SunOS 4.1.1) + This can be useful for DOS, OS/2 or systems that don't have + lex. + +states.sed - sed(1) script to munge the state names in detex.l so that + reasonable names can be used in the source without causing + lex(1) to overflow. + +os2 - subdirectory containing support for compilation on + OS/2 and DOS systems + + +This software package is distributed under the NCSA/University of +Illinois Open Source License. The latest version is available at +http://www.cs.purdue.edu/homes/trinkle/detex/ Send comments and fixes +to me via email. + +Daniel Trinkle <trinkle@cs.purdue.edu> +Department of Computer Sciences +Purdue University +West Lafayette, IN 47907-1398 + +April 26, 1986 + +Modified -- June 4, 1986 +Changed so that it automatically recognizes LaTeX source and ignores several +environment modes such as array. + + +Modified (Version 2.0) -- June 30, 1984 +Now handles white space in sequences like "\begin { document }". Detex is not +as easily confused by such things as display mode ends and begins that don't +match up. + + +Modified -- September 19, 1986 +Added the "-e <environment-list>" option to ignore specified LaTeX +environments. + + +Modified -- June 30, 1987 +Added the "-n" no-follow option, to allow detex to ignore \input and \include +commands. Also changed the algorithm for locating the input files. It now +interprets the "." more reasonably (i.e. it is not always the beginning of an +extension). + + +Modified -- December 15, 1988 +Added handling of verbatim environment in LaTeX mode and added it to the list +of modes ignored by default. Because of limitations with lex, it was +necessary to shorten the names of some of the existing start states before +adding a new one (ugh). + + +Modified -- January 3, 1988 +Added ignore of \$ inside $$ (math) pair. + + +Modified (Version 2.2) -- June 25, 1990 +Control sequences are no longer replaced by space, but just removed. This +means accents no longer cause output words to be broken. The "-c" option was +added to show the arguments of \cite, \ref, and \pageref macros. This is +useful when using something like style on the output. + + +Modified (Version 2.3) -- September 7, 1990 +Fixed the handling of Ctl mode a little better and added an exception +for \index on suggestions from kcb@hss.caltech.edu (KC Border). Also +changed the value for DEFAULTINPUTS to coincide with a local change. + + +Modified -- February 10, 1991 +Added -t option to force TeX mode even when seeing the "\begin{document}" +string in the input. + + +Modified -- February 23, 1991 +Based on suggestions from pinard@iro.umontreal.ca (Francois Pinard), I +added support for the SysV string routines (-DUSG), added defines for +the flex lexical scanner (-DFLEX_SCANNER), changed NULL to '\0' when +using it as a character (his sys defined NULL as (void *)0), changed +the Makefile to use ${CC} instead of cc, and added comments about the +new compile time options. + + +Modified (Version 2.4) -- September 2, 1992 +Corrected the way CITEBEGIN worked. Due to serious braindeath I had +the condition wrong in the if test. It should be (fLatex && !fCite). +This solves the problem a couple people reported with amstex style +\ref entries. + +Added a preprocessing sed(1) command to replace all the long, easy to +read state names with short two letter state names (SA-S?) so that lex +won't overflow and I don't have to keep shortening the state names +every time I add one. If a state is added, it must also be added to +states.sed (order is important) along with its unique S? replacement. + +Added \pagestyle, \setcounter, and \verb handling from +K.Sattar@cs.exeter.ac.uk (Khalid Sattar). Also allows invocation as +"delatex" to force LaTeX mode. + +Applied patches from queinnec@cenatls.cena.dgac.fr (Philippe Queinnec) +to handle nested {}s in state <LaMacro2> (\bibitem, \cite, \index). + +Added special ligature handling (\aa, \ae, \oe, \l, \o, \i, \j, \ss) +at the suggestion of gwp@dido.caltech.edu (G. W. Pigman II). + +Cleaned up the comments on detex.h, added mathmatica to DEFAULTENV. + + +Modified (Version 2.5) -- January 28, 1993 +Leading spaces in macros are no longer stripped. This means +"foo\footnote{ bar}" comes out as "foo bar" instead of "foobar". + +Fixed special ligature handling so it works in cases line {\ss} instead of +just when followed by a space. + + +Modified (Version 2.6) -- July 30, 1993 +Added OS/2 port from hankedr@mail.auburn.edu (Darrel R Hankerson). + +Added handling of leading and trailing ':' in TEXINPUTS per the latest +TeX as suggested by jnp@tfl.dk (J|rgen N|rgaard). + +Changed the way the input path is constructed in SetInputPaths() so we +never try to modify a constant string. + +Changed the way the the ignore environment list is contructed in +SetEnvIgnore() so we never try to modify a constant string. + +Changed the USG define to HAVE_STRING_H. + +Fixed the states.sed script so it only replaces "Input" in the correct +places. I would like to use the \< \> word separator patterns but +they are not supported by all versions of sed. This as least works. + +Changed the detex.c target in the Makefile to use a temporary file +because I experienced problems (segmentation fault) with lex on +Solaris 2.1 when input was from stdin. + + +Modified (Version 2.7) -- September 10, 1997 +Removed line breaks in detex.l between a few patterns and actions. It +appears that flex is no longer able to handle this. Thanks to Anthony +Harris <harris@hebb.neurology.pitt.edu> and Marty Leisner +<leisner@sdsp.mc.xerox.com> for reporting this. + + +Porting notes -- March 30, 1992 +When using gcc, or compiling on a NeXT, you should compile with +-fwritable-strings. With the change to SetInputPaths() in 2.6 this +should no longer be necessary. + +On a NeXT, it has been reported that lex replaces the '\0' with NULL, +and then the compiler complains about it. I think this is an old bug +that is no longer applicable. + +July 30, 1993 +The flex scanner generator does not work because it does not handle +input buffering the same way as lex. I don't know of any reasonable +way to rewrite detex to get around this problem. + +May 25, 1995 +According to alain@ia1.u-strasbg.fr (Alain Ketterlin), using flex +allows 8-bit characters to be handled correctly. + +Modified (Version 2.8) -- January 1, 2008 +Added NCSA/University of Illinois Open Source License to facilitate +free redistribution. |