[PATCH 1/2] alloca: sync with Autoconf master

2020-07-30 Thread Paul Eggert
* m4/alloca.m4 (_AC_LIBOBJ_ALLOCA): Do not define if Autoconf 2.70 or later, since Autoconf master now matches us. --- ChangeLog| 7 +++ m4/alloca.m4 | 14 +- 2 files changed, 16 insertions(+), 5 deletions(-) diff --git a/ChangeLog b/ChangeLog index 95851778c..fa5e1dab1

[PATCH 2/2] largefile: sync with Autoconf master

2020-07-30 Thread Paul Eggert
* m4/largefile.m4 (_AC_SYS_LARGEFILE_TEST_INCLUDES): Avoid undefined behavior on platforms where off_t is 32 bits. See: https://bugs.debian.org/742780 --- ChangeLog | 5 + m4/largefile.m4 | 2 +- 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/ChangeLog b/ChangeLog index

Re: 'fdl' vs. 'fdl-1.3': difference and/or redundant?

2020-07-30 Thread Bruno Haible
Bernhard Voelker wrote: > okay, that means the URL bug should get fixed, right? > Patch attached. This cannot be applied to gnulib. It would need to be applied upstream (maybe by Karl?). > > There is now a notice in the module description > > "Don't use this module! Instead, copy the

Re: 'fdl' vs. 'fdl-1.3': difference and/or redundant?

2020-07-30 Thread Bernhard Voelker
On 2020-07-31 00:04, Bruno Haible wrote: > Bernhard Voelker wrote: >> But once this would be fixed, then the whole 'fdl' module (or the 'fdl-1.3' >> module?) seems to be redundant, doesn't it? >> Couldn't we remove one of them? > > The idea of the 'fdl' module is to point to the current version

Re: Optimize three-valued comparison between integers

2020-07-30 Thread Paul Eggert
On 7/23/20 3:15 PM, Bruno Haible wrote: + AH_VERBATIM([micro_optimizations], +[/* _GL_CMP (n1, n2) performs a three-valued comparison on n1 vs. n2. + It returns + 1 if n1 > n2 + 0 if n1 == n2 + -1 if n1 < n2 While looking into using this in Emacs I noticed that _GL_CMP (A, B)

Re: 'fdl' vs. 'fdl-1.3': difference and/or redundant?

2020-07-30 Thread Bruno Haible
Bernhard Voelker wrote: > But once this would be fixed, then the whole 'fdl' module (or the 'fdl-1.3' > module?) seems to be redundant, doesn't it? > Couldn't we remove one of them? The idea of the 'fdl' module is to point to the current version of the GFDL. A maintainer who does not want to be

Re: 'fdl' vs. 'fdl-1.3': difference and/or redundant?

2020-07-30 Thread Bernhard Voelker
On 2020-07-30 23:48, Bernhard Voelker wrote: > $ diff -u0 doc/fdl.texi doc/fdl-1.3.texi > --- doc/fdl.texi2019-06-23 11:00:48.274552884 +0200 > +++ doc/fdl-1.3.texi2019-06-23 11:00:48.258552658 +0200 > @@ -417 +417 @@ > -@uref{https://www.gnu.org/copyleft/}. >

'fdl' vs. 'fdl-1.3': difference and/or redundant?

2020-07-30 Thread Bernhard Voelker
The 'fdl' module and its file 'doc/fdl.texi' contains version GFDL 1.3 since Nov 2008 [1][2]. Comparing it to the file 'doc/fdl-1.3.texi' of the 'fdl-1.3' module, there is still this single difference: $ diff -u0 doc/fdl.texi doc/fdl-1.3.texi --- doc/fdl.texi 2019-06-23

[PATCH] Work around some Oracle Studio attribute bugs

2020-07-30 Thread Paul Eggert
These were discovered when building bleeding-edge Emacs with Oracle Studio. * m4/gnulib-common.m4 (_GL_ATTRIBUTE_COLD, _GL_ATTRIBUTE_MAY_ALIAS): Port to Oracle Studio 12.6, which mishandles __attribute__ ((__cold__)) and __attribute__ ((__may_alias__)) even though __has_attribute says they work.

Re: iconv replacements

2020-07-30 Thread Bruno Haible
[Dropping bug-bison from CC] > > Yes and no. The code is not making assumptions about a particular iconv() > > implementation. But it needs to distinguish two categories of replacements > > done by iconv(): > > - those that are harmless (for example when replacing a Unicode TAG > >

Re: [musl] Building Bison 3.7 with musl (was Re: portability issues with unicodeio)

2020-07-30 Thread Florian Weimer
* Bruno Haible: > Yes and no. The code is not making assumptions about a particular iconv() > implementation. But it needs to distinguish two categories of replacements > done by iconv(): > - those that are harmless (for example when replacing a Unicode TAG > character U+E00xx with an empty