December 2003 - Cvs-hugs

cvs commit: hugs98/src storage.c
by Sigbjorn Finne 18 Dec '03

18 Dec '03

sof 2003/12/17 21:20:40 PST Modified files: src storage.c Log: setCurrModule(): 'currentModule' might be NIL, so avoid resetting the 'modImports' field in that case. (If you don't, a random word in writeable memory is nulled out. D'oh!) Brings cygwin builds back online. Revision Changes Path 1.84 +7 -3 hugs98/src/storage.c

1 0

cvs commit: hugs98/src config.h.in
by Sven Panne 17 Dec '03

17 Dec '03

panne 2003/12/17 00:40:35 PST Modified files: src config.h.in Log: Synched with aclocal.m4 Revision Changes Path 1.36 +4 -1 hugs98/src/config.h.in

1 0

cvs commit: hugs98/src/unix aclocal.m4
by Sven Panne 17 Dec '03

17 Dec '03

panne 2003/12/17 00:40:04 PST Modified files: src/unix aclocal.m4 Log: Synched with aclocal.m4 from fptools (opengl32.dll fiddling) Revision Changes Path 1.35 +12 -1 hugs98/src/unix/aclocal.m4

1 0

cvs commit: www/Hugs/pages documentation.htm latest.htm libstatus.html
by ross＠glass.cse.ogi.edu 16 Dec '03

16 Dec '03

ross 2003/12/16 07:35:23 PST Modified files: Hugs/pages documentation.htm latest.htm libstatus.html Log: update location of libraries page Revision Changes Path 1.5 +1 -1 www/Hugs/pages/documentation.htm 1.11 +1 -1 www/Hugs/pages/latest.htm 1.4 +2 -2 www/Hugs/pages/libstatus.html

1 0

cvs commit: hugs98/docs/users_guide introduction.xml
by ross＠glass.cse.ogi.edu 16 Dec '03

16 Dec '03

ross 2003/12/16 07:34:35 PST Modified files: docs/users_guide introduction.xml Log: update location of libraries page Revision Changes Path 1.2 +2 -2 hugs98/docs/users_guide/introduction.xml

1 0

hugs w/ cygwin
by Sigbjorn Finne 16 Dec '03

16 Dec '03

FYI, hugs builds cleanly with two cygwin installs I've tested on here: an older one w/ gcc-2.95 and one w/ gcc-3.2 (from a clean cygwin install done on 29/07/03). The reported behaviour suggests a C compiler making life too hard for the conservative GC; one way to find out is to cut back on the use of -O2. Re: Registry gibberish -- this was due to unhygienic macro usage in opts.c. Fixed now. --sigbjorn

1 0

cvs commit: hugs98/src machdep.c
by Sigbjorn Finne 16 Dec '03

16 Dec '03

sof 2003/12/15 21:55:33 PST Modified files: src machdep.c Log: needPrims(): fix buglet introduced by rev 1.110 Revision Changes Path 1.112 +3 -3 hugs98/src/machdep.c

1 0

cvs commit: hugs98/src opts.c
by Sigbjorn Finne 16 Dec '03

16 Dec '03

sof 2003/12/15 21:48:40 PST Modified files: src opts.c Log: - PUTS(s) macro: be defensive & 'linear' in the use of 's' by binding it to a local, first thing. If not, confusion reigns if 's' is side-effecting (as was the case with PUTS() use in optionsToStr().) - PUTStr_aux(): don't rely on PUTS() being well-behaved & bind argument to local first. This latent bug was previously just a minor source of inefficiency (two calls to unlexChar() per character), but with the introduction of the side-effecting ExtractChar() macro, things got a bit uglier. Revision Changes Path 1.15 +8 -3 hugs98/src/opts.c

1 0

table trouble + registry revisited
by Sven Panne 15 Dec '03

15 Dec '03

Hugs runs fine under plain Cygwin now (i.e. via a simple "./configure"), but only when WANT_FIXED_SIZE_TABLES is #defined to 1. Otherwise I get strange 'Undefined constructor function "Deadlock"' errors when starting Hugs, but Deadlock *is* defined (e.g. ":i Exception" shows it). Consistently renaming it to e.g. "Deadlokk" works. Hmmm... I don't understand the current memory management scheme in Hugs in detail, so I can't fix this easily for myself. Explaining DYN_TABLES and WANT_FIXED_SIZE_TABLES a bit more might be a good idea. And another bug: Using :s results in garbage in the registry. Deja vu? Puzzled, S.

1 0

cvs commit: hugs98/src storage.c
by Sven Panne 15 Dec '03

15 Dec '03

panne 2003/12/15 09:58:15 PST Modified files: src storage.c Log: Fixed types of tabObserve/tabBreakpt in table allocation Revision Changes Path 1.83 +4 -4 hugs98/src/storage.c

1 0