Re: undefined behavior in closeout, aggravated by libsigsegv

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 According to Bruno Haible on 11/21/2009 10:11 AM: >> Here's three patches that add test cases, then respectively fixes cygwin >> 1.7 and 1.5 to pass the new tests. I've also tested that on platforms >> like Linux that the new tests pass, as well as te

Re: [PATCH] sh-quote: avoid a warning from -Wstrict-prototypes

Hi Jim, > I get only one -Wstrict-prototypes violation in diffutil's lib/ > directory. This fixes it. Ok to apply? Not in gnulib, but you are free to use this patch as lib/sh-quote.c.diff in a gnulib --local-dir of your package. Rationale: In the C and C++ standards, static void init_sh_qu

Re: uninitialized struct members

Hi Jim, > > - struct partition part; > > + struct partition part IF_LINT (= { .xmid = 0, .ymid = 0 }); > > Any code using a one-argument macro like that is rejected, > due to the comma: > > ../lib/diffseq.h:467:64: error: macro "IF_LINT" passed 2 arguments,\ > but takes just 1

Re: version-etc not working with autoconf 2.61

Eric Blake asked: > Gnulib pre-assumes that you are using Automake, and automake generally > guarantees that PACKAGE is defined. What is guile doing that violates > this assumption? guile's configure.ac is invoking AM_INIT_AUTOMAKE([... no-define ...]) which has the effect that PACKAGE and VERS

Re: problem with getopt-gnu on solaris

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 According to Jean-Louis Martineau on 11/19/2009 1:05 PM: > Hi, > > I updated my gnulib to git 841ad196dd3ca0fbc916b911449067417ac7a7cb and > changed the getopt module to getopt-gnu. > > I get error when compiling on Solaris 10: I can't reproduce the

Re: version-etc not working with autoconf 2.61

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 According to Andy Wingo on 11/21/2009 10:59 AM: > Hello, > > I just applied this to Guile. Perhaps something like it should go into > Gnulib? Gnulib pre-assumes that you are using Automake, and automake generally guarantees that PACKAGE is defined.

[PATCH] sh-quote: avoid a warning from -Wstrict-prototypes

Hi Bruno, I get only one -Wstrict-prototypes violation in diffutil's lib/ directory. This fixes it. Ok to apply? >From 29434090a56d569a307b918dc20aa87f7fb7525e Mon Sep 17 00:00:00 2001 From: Jim Meyering Date: Sat, 21 Nov 2009 19:37:58 +0100 Subject: [PATCH] sh-quote: avoid a warning from -Wst

version-etc not working with autoconf 2.61

Hello, I just applied this to Guile. Perhaps something like it should go into Gnulib? Cheers, Andy commit 624c5c7c8870baf610785a821e8bb40d41db2051 Author: Andy Wingo Date: Fri Nov 20 17:24:40 2009 +0100 fix version-etc for older autoconf * lib/version-etc.c: Don't reference the

problem with getopt-gnu on solaris

Hi, I updated my gnulib to git 841ad196dd3ca0fbc916b911449067417ac7a7cb and changed the getopt module to getopt-gnu. I get error when compiling on Solaris 10: "../gnulib/getopt.h", line 183: (struct) tag redeclared: option "../gnulib/getopt.h", line 232: identifier redeclared: getopt_long

Re: uninitialized struct members

Bruno Haible wrote: > I'm also having similar warnings in the *list modules and in GNU clisp. > >> Since I require warning free compilation with -Wuninitialized, >> I am considering this patch. Would you prefer to avoid the >> warning in a different manner? > > Well, I would prefer if the inaccura

Re: uninitialized struct members

Hi Jim, > While we're on the topic, this is the final warning > I'm seeing in diffutils with most of "manywarnings" enabled: > > In file included from analyze.c:36: > ../lib/diffseq.h: In function 'compareseq': > ../lib/diffseq.h:461: warning: 'part.ymid' may be used uninitialized\ >

Re: undefined behavior in closeout, aggravated by libsigsegv

Eric Blake wrote on 2009-07-23: > Actually, the bug is that libsigsegv is too eager to claim that an SEH > exception is a SEGV, even if cygwin was otherwise prepared to handle it > and not raise SIGSEGV. The fix is to delay handling of the fault until > after the SIGSEGV has actually been raised,

[PATCH] c-stack: avoid "ignoring return value of `write'" warning

It was easy and safe to ignore these four warnings, so... >From df57f4270959d6679422896ca6f579d048777d81 Mon Sep 17 00:00:00 2001 From: Jim Meyering Date: Sat, 21 Nov 2009 17:52:54 +0100 Subject: [PATCH] c-stack: avoid "ignoring return value of `write'" warning * lib/c-stack.c: Include "ignore-v

Re: [PATCH] diffseq: remove useless assignment to "best"

Bruno Haible wrote: > Jim Meyering wrote: >> At that point "best" is already guaranteed to be zero. > > The variable 'best' is used for two different purposes here: for two different > searches in opposite directions. Therefore it does not increase > maintainability > to rely on the result of the

Re: [PATCH] diffseq: remove useless assignment to "best"

Jim Meyering wrote: > At that point "best" is already guaranteed to be zero. The variable 'best' is used for two different purposes here: for two different searches in opposite directions. Therefore it does not increase maintainability to rely on the result of the first search loop for the initial

[PATCH] diffseq: remove useless assignment to "best"

FYI, found by inspection, >From a823652ce27628bbccd09f7614a7508641b4f004 Mon Sep 17 00:00:00 2001 From: Jim Meyering Date: Sat, 21 Nov 2009 13:49:54 +0100 Subject: [PATCH] diffseq: remove useless assignment to "best" * lib/diffseq.h (diag) [USE_HEURISTIC]: Remove useless "best = 0" assignment.