[PATCH] stdalign: port to stricter ISO C11

2013-03-29 Thread Paul Eggert
ISO C11 says that _Alignof's operand must be a parenthesized type. Problem reported by Eli Zaretskii in . * doc/posix-headers/stdalign.texi (stdalign.h): Document this. * m4/stdalign.m4 (gl_STDALIGN_H): Don't use _Alignof (expr).

Re: Compilation errors when using extended GCC warnings

2013-03-29 Thread Eric Blake
On 03/29/2013 04:27 PM, Assaf Gordon wrote: > Hello, > > I'm trying to add "gnulib" to a new project. > My project enables gcc's "-Wswitch-enum -Wswitch-default -Werror" settings. In general, gnulib does NOT try to cater to all possible gcc warnings. You may want to take the approach that coreuti

Compilation errors when using extended GCC warnings

2013-03-29 Thread Assaf Gordon
Hello, I'm trying to add "gnulib" to a new project. My project enables gcc's "-Wswitch-enum -Wswitch-default -Werror" settings. When adding some gnulib modules, compilation fails. When adding the "propername" module, I get the following error: === unistr/u8-uctomb-aux.c: In function 'u8_uctomb_au

Re: stat-time.h vs. pthread-win32

2013-03-29 Thread Ludovic Courtès
Sorry, there was a thinko in the patch I posted: that doesn’t solve anything since the user’s gets included anyway. The root of the problem is that Guile’s ‘configure’ was defining HAVE_STRUCT_TIMESPEC, hence the problem: http://git.sv.gnu.org/cgit/guile.git/commit/?h=stable-2.0&id=7e7c6f6a93

Re: stat-time.h vs. pthread-win32

2013-03-29 Thread Ludovic Courtès
Paul Eggert skribis: > That workaround looks OK to me. pthread-win32 > seems like it's in pretty bad shape, so hacks like > that appear to be necessary. Here’s a complete patch. Ludo’. >From 1147bf0f9efe58680f571b87907a741ce2d93a24 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ludovic=20Court=C3=A