GNU Awk

v4.2.0: * Fixed that loading extensions whose name is longer than 8 chars failed.

by KO Myung-Hun

/*--------------- Release 1 (Tue Jan 22 2000) ---------------------------*/ GNU awk version 3.0.4 + multi-byte extension 1.12 for OS/2 Table of contents ================= Introduction @int Required Executables @req Installation @ins Documents @doc Mannuals @man About Multi-byte extension @mul Compile from Source @com Introduction @int ============ Gawk is the GNU Project's implementation of the AWK programming language. This package contains GNU awk version 3.0.4 with multi-byte extension patch 1.12 by Katsuyuki Okabe <hgc02147@nifty.ne.jp>. Owing to his work, we can process text files written in 8-bit code including Japanese SJIS/EUC. Thanks to Kai Uwe Rommel <rommel@ars.de> 's OS/2 codes, no effort was needed to make `gawk'. If you have questions and suggestions, please mail me. SAWATAISHI JUN, YOKOHAMA JAPAN <jsawa@attglobal.net> http://www2s.biglobe.ne.jp/~vtgf3mpr/indxos2.htm Required Executables @req ==================== HOBBES=ftp://hobbes.nmsu.edu/pub/os2 LEO=ftp://ftp.leo.org/pub/comp/os/os2/leo JSAWA=http://www2s.biglobe.ne.jp/~vtgf3mpr Prerequisite ~~~~~~~~~~~~ `sh' included in KSH package LEO/shells/pdksh-5.2.13-emx.zip HOBBES/util/shell/ksh527rt.zip To view manpages ~~~~~~~~~~~~~~~~ `man' (requires GNU groff) HOBBES/dev/help/gnuman.zip ; v1.0 LEO/gnu/doc/man11a.zip ; v1.1a JSAWA/manpage/manpage-e.htm ; v1.5g JSAWA/manpage/manpage.htm GNU groff Groff v1.10 HOBBES/apps/wp/gnugroff.zip LEO/gnu/doc/gnugroff.zip GNU groff v1.15 (with Japanization patch) JSAWA/manpage/manpage-e.htm [English] JSAWA/manpage/manpage.htm [Japanese] GNU grep HOBBES/util/file/gnugrep.zip ; v2.0 LEO/gnu/systools/gnugrep.zip ; v2.0 JSAWA/gnu/grep.htm ; v2.3h (v2.4 will come soon) Pager : I strongly recommend you to install `less' less 332 with Japanese patch: http://hp.vector.co.jp/authors/VA000199/os2/ You can use this in virtual terminal (xterm etc) of XFree86 for OS/2. You should get this one if you'd like to see manpages in XFree86. less 329 without XFree86 capability LEO/unix/gnu/less329-2.zip HOBBES/unix/util/util/file/less329-2.zip `gzip' and `bzip2' [gzip] HOBBES/util/archiver/gz124_32.zip LEO/gnu/archiver/gz124-32.zip [bzip2] HOBBES/util/archiver/bzip090b.zip LEO/archiver/bzip090b.zip http://www.muraroa.demon.co.uk/ JSAWA/rpm/bzip2.htm ; v0.9.5d --- this version PRESERVE TIME STAMP http://netpassport-wc.netpassport.or.jp/~wmakoto/CLUB-bzip2.html To view Info manual ~~~~~~~~~~~~~~~~~~~ GNU texinfo v4.0 ( '*' denotes modification by me) - can be run both in VIO mode and in *XFree86* terminals - support 7 kind of language including *Japanese* for message - support Info manual written in 8-bit code including Japanese - `texindex' support Japanese SJIS coded texinfo source. - `makeinfo' support HTML output JSAWA/gnu/texinfo.htm [English] JSAWA/jpinfo.htm [Japanese] ( GNU texinfo v3.11 -- obsolete LEO/gnu/doc/gnuinfo.zip HOBBES/apps/wp/tex/util/gnuinfo.zip ) Installation @ins ============ Unzip `gawk304.zip' in a root directory of the drive where emx runtime programs are installed. This is easiest installation. If your emx drive is 'E' do like this. c:\tmp> unzip SOMEWHERE/gawk304.zip c:\tmp> dir README-gawk.os2 emx-gawk.zip c:\tmp> unzip emx-gawk.zip -d e:\ c:\tmp> copy README-gawk.os2 e:\emx\doc\gawk-3.0.4\ Confirm `g:\emx\bin' is included in PATH environment variable. Awk library files are extracted in `/emx/share/awklib' Define AWKPATH environment variable in your `config.sys'. set AWKPATH=e:/emx/share/awklib;x:/foo/bar;.; `gawk' searches given script (by "-f" option) in directories specified by AWKPATH. Documents @doc ========= /emx/doc/GNU COPYING -- GNU GENERAL PUBLIC LICENSE Version 2, June 1991 GNU awk is copyrighted software. You should read this document first. FTP -- How to get GNU Software by Internet FTP or by UUCP /emx/doc/gawk-3.0.4 <Essential documents> README README.FIRST README.pc NEWS ChangeLog POSIX.STD LIMITATIONS PROBLEMS FUTURES ACKNOWLEDGMENT <Japanese documents> OChangeLog.MB, ChangeLog.MB -- history of multi-byte extension OREADME.MB, README.MB -- describes new features and options of multi-byte extension. <OS/2 related> README-gawk.os2 -- this file you should have been copied here Manual @man ====== /emx/info/gawk.info.gz -- gzipped Info manual; you can view this with Info reader `info' ( included in GNU texinfo ). /emx/man/man1/gawk.1.gz -- gzipped manpage; you need `man' ang GNU groff to view this mapage. /emx/doc/gawk-3.0.4/awkcard.ps -- handy and easy-to-use manual; you'd better print this using postsrcipt printer or `ghostsctipt'. About Multi-byte extension @mul ========================== Feature of multi-byte extension is written in Japanese (OREADME.MB, README.MB). For non-Japanese users, I explain essential information on the extention. Default character code of `gawk' included in this package is Japanese SJIS. This `gawk' ALWAYS run in 8-bit "clean" mode. - Options added by multi-byte extension -W ctype=ASCII --ctype=ASCII : Do not recognize type (EUC, SJIS, UTF8) of multi-byte characters (MBC). -W ctype=EUC --ctype=EUC : Recognize MBC as Japanese EUC code -W ctype=SJIS --ctype=SJIS : Recognize MBC as Japanese Shift JIS code -W ctype=UTF8 --ctype=UTF8 : Recognize MBC as UTF8. Both UCS-2 (unicode) and UCS-4 are supported. -W memory --memory : Conserve memory by inhibiting DFA production for regular expression. - New function and extended function RS ; accept MBC (extension) MBC can be used for identifier (extension) substr(STRING, START [, LENGTH]) -- extended function Now substr() recognize MBC as single char according to specified MBC code (SJIS, EUC, UTF8). Devision of one MBC char into two single byte chars is inhibited by correcting START and LENGTH. Example: Interprete a pair of upper case characters represents _single_ MBC character Without multi-byte extension; substr("aXYiABuCDe", 3, 4) returns "YiAB" This means MBC character 'XY' is devided into two single-byte codes (characters if both 'Y' and 'A' are printable ascii code). Extended substr() returns "XYi"; START / LENGTH are corrected from 3 / 4 to 2 / 3. jindex() -- new function This returns the position as ordering number not as byte number ( index() ) jlength() -- new function This returns number of characters. ( length() returns total bytes ) jsubstr(STRING, START [, LENGTH]) -- new function This interpretes START / LENGTH as "ordering number" / "number of charac- ters" instead of "byte offset" / "number of bytes". Compile from source @com =================== GNU awk source, which we can find in ftp://ftp.gnu.org/gnu/gawk/, is distributed by the Free Software Foundation. You should better download source archive via its mirror site. --> Read a file `FTP' included in this package ( \emx\doc\GNU\FTP ). Get `gawk-3.0.4.tar.gz' Get multe-byte extension patch from http://member.nifty.ne.jp/wills/program.html (gawkmb112.diff.gz) To run shell script in `Makefile', `sh.exe' must be available. You can get it from ftp sites: ftp.leo.org/pub/comp/os/os2/leo/shells/pdksh-5.2.13-emx.zip Use GNU patch, patch.exe' or `gpatch.exe': OS/2 default `patch.exe' can't be used. GNU gcc and GNU file utilities must be available. After you got source code, you can easily make `gawk'. SAWATAISHI JUN, YOKOHAMA JAPAN <jsawa@attglobal.net> http://www2s.biglobe.ne.jp/~vtgf3mpr/indxos2.htm /*----- end of document -----------------------------------------------------*/