Wed Jul 28 16:28:42 IDT 2004 ============================ As of gawk 3.1.4, configure should correctly handle HP-UX and I18N issues. -- Arnold -------------------------------------------------------------- 2003-12-10 15:19:38 EST Michael Elizabeth Chastain I built and tested gawk on hppa-hp-hpux11.11 and ia64-hp-hpux11.23. All the tests in the test suite passed. I built with these compilers: gcc 3.3.2 hp ansi C from /opt/ansic/bin hp aCC from /opt/aCC/bin I ran into these problems: NLS does not work; configure with --disable-nls. -D_XOPEN_SOURCE=500 does not work. Multibyte support is not available. To get multibyte support, the following ugly hack might work: --- gawk-3.1.3.orig/custom.h 2003-06-09 17:45:53.000000000 +0200 +++ gawk-3.1.3/custom.h 2003-12-17 15:55:04.000000000 +0100 @@ -101,4 +101,7 @@ #undef HAVE_TZSET #define HAVE_TZSET 1 #define _TZSET 1 +/* an ugly hack: */ +#include +#define HAVE_MBRTOWC 1 #endif ------------------------------- Mon, 27 May 2002 17:55:46 +0800 The network support "|&" may not work under HP-UX 11. An error message appears similar to this: gawk: test_script.awk:3: fatal: get_a_record: iop->buf: can't allocate -61246 bytes of memory (not enough space) Solution: This is a bug in the fstat() call of HP-UX 11.00, please apply the cumulative ARPA Transport patch PHNE_26771 to fix it. The following is the related description in PHNE_26771: Customer's application gets the wrong value from fstat(). Resolution: The value returned via st_blksize is now retrieved from the same info as in 10.20. In case you cannot apply the HP patch, the attached patch to gawk source might work. Xiang Zhao Stepan Kasal diff -ur gawk-3.1.3.a0/posix/gawkmisc.c gawk-3.1.3.a1/posix/gawkmisc.c --- gawk-3.1.3.a0/posix/gawkmisc.c Sun May 25 15:26:19 2003 +++ gawk-3.1.3.a1/posix/gawkmisc.c Fri Jul 11 08:56:03 2003 @@ -126,7 +126,13 @@ * meant for in the first place. */ #ifdef HAVE_ST_BLKSIZE -#define DEFBLKSIZE (stb->st_blksize > 0 ? stb->st_blksize : BUFSIZ) + /* + * 100k must be enough for everybody, + * bigger number means probably a bug in fstat() + */ +#define MAXBLKSIZE 102400 +#define DEFBLKSIZE (stb->st_blksize > 0 && stb->st_blksize <= MAXBLKSIZE \ + ? stb->st_blksize : BUFSIZ) #else #define DEFBLKSIZE BUFSIZ #endif