Re: AIX xlc gets confused by logic in stdio.h

Ralf Wildenhues wrote: > This seems to work fine on AIX with this patch. Thanks! Thanks for testing. I've now applied the patch. Bruno

cosl test failure

Hi Paolo, I added trivial tests for the mathl functions, that verify that 1) there are no link errors, 2) on a randomly picked value, the function's result is correct to 10 decimal digits. For the cosl() function, when compiled from the gnulib sources, this test fails. Seen on OpenBSD/x8

Re: split mathl module

Paolo Bonzini wrote: > The patch is fine OK. Testing revealed a few tweaks were necessary (attached below). > There's quite some duplication in the macros Yes. There's a trade-off between duplication on one side and complexity (pushdef, popdef) that reduces the ease of understanding of the *.m4

new common math function modules

The module 'sqrtl' depends on sqrt. But we don't have an 'sqrt' module yet, which determines where the sqrt function sits. So I'm adding one. While at it, 35 other math.h functions can be treated in the same way. 2010-01-18 Bruno Haible New modules for common functions. * m4

Re: split mathl module (was: Add some #undef's to math.in.h, for AIX with xlc.)

Hi Bruno, * Bruno Haible wrote on Mon, Jan 18, 2010 at 01:27:17AM CET: > Here's a proposed patch, which seeks to fix 1) to 5). Ralf will need to > re-test on AIX. No more problems on AIX with this patch. Thanks! > 2010-01-17 Bruno Haible > > New modules acosl, asinl, atanl, cosl, expl

Re: AIX xlc gets confused by logic in stdio.h

* Bruno Haible wrote on Sun, Jan 17, 2010 at 09:19:33PM CET: > 2010-01-17 Bruno Haible > > Avoid a link error due to the __printf__ symbol. > * lib/stdio.in.h (__attribute__): Define to empty also for gcc 2.5.x > and 2.6.x. > (__format__, __printf__): Remove definitions.

[PATCH 6/N] Detect cache variable collision.

* gnulib-tool (func_cache_lookup_module): Store the module name belonging to the cache variable; error out if two different module names map to the same cache variable name. --- gnulib-tool |6 +- 1 files changed, 5 insertions(+), 1 deletions(-) diff --git a/gnulib-tool b/gnulib-tool inde

[PATCH 7/N] gnulib-tool: new option --no-cache for old extraction code.

* gnulib-tool: Accept option --no-cache, turning off $do_cache. Update matching short versions of --no-changelog. (func_usage): Update. (sed_extract_cache_prog): Renamed from ... (sed_extract_prog): ... this; revert to old extraction script. (func_get_description, func_get_status) (func_get_notice,

Re: [PATCH 0/5] speed up gnulib-tool some more

Hello Bruno, * Bruno Haible wrote on Mon, Jan 18, 2010 at 01:41:57AM CET: > In particular, patch 1 and 2 each removed many blank lines from the output. > One needs to verify whether this is harmless. I did verify that for a number of test cases (one module, several modules, all modules, for diffe

[PATCH] ignore-value: relax license to LGPLv2+

FYI, I wanted to use ignore-value in libvirt, which is LGPLv2+, so have relaxed its license: >From 146d9145073e62a2096a2d6b33f75e93908fedf3 Mon Sep 17 00:00:00 2001 From: Jim Meyering Date: Mon, 18 Jan 2010 11:49:16 +0100 Subject: [PATCH] ignore-value: relax license to LGPLv2+ * modules/ignore-v

Re: split mathl module

On 01/18/2010 01:27 AM, Bruno Haible wrote: Hi Paolo, Ralf Wildenhues wrote: AIX 5.3 needs a number of #undef's in math.in.h, in order to compile, I don't want to consider Ralf's patch, because the 'mathl' module is broken in so many respects that first a cleanup is needed. The patch is fin