Re: [PATCH] timevar: add missing timevar.def

Hi Collin, > I've added a small mention in the documentation to make it official. Thanks. Now that I think more about it, I forgot to update gnulib-tool.sh. gnulib-tool.sh does not need to support the new feature, but it should at least not malfunction just because a module description contains t

Re: [PATCH] timevar: add missing timevar.def

Hi Bruno, Bruno Haible writes: > Collin Funk wrote: >> Perhaps there should be an extra status for test modules that should >> generally be excluded. > > I was thinking in the same direction. The 'Status' is already taken (it is > 'deprecated' for some of the relevant modules). Therefore, I'm ad

Re: module snippet/link-warning

Collin Funk writes: >> Do you have time to investigate the cause, and why it depends on >> --single-configure, please? > > Sure, I'll look into it today. The issue is that link-warning is in gltests/ and the following Makefile line: link-warning.h: $(top_srcdir)/build-aux/snippet/link-warni

gnulib-tool.py: New options --extract-[recursive-]dependents

Paul Eggert wrote: > I was testing all modules that depend on c99, this way: > >./gnulib-tool -h --create-testdir --dir foo $(cd modules && grep -lx > 'c99' $(git ls-files)) The libunistring maintenance also needs a way to find all modules that depend on a particular module. I've missed a co

Re: [PATCH] timevar: add missing timevar.def

Paul Eggert wrote: > > What was the immediate problem that you had with this module? > > I was testing all modules that depend on c99, this way: > >./gnulib-tool -h --create-testdir --dir foo $(cd modules && grep -lx > 'c99' $(git ls-files)) >cd foo >./configure >make > > The la

gnulib-tool.py: Fix an exception message

In gnulib-tool.py, error 3 is "configure file does not exist", not "module doesn't exist". 2024-07-28 Bruno Haible gnulib-tool.py: Fix an exception message. * pygnulib/GLModuleSystem.py (GLModuleSystem.find): Raise error 23, not 3. * pygnulib/GLError.py: Provi

Re: module snippet/link-warning

Bruno Haible writes: > The resulting testdir doesn't build, indeed. > > But with option --single-configure, it works fine. Interesting, I didn't know that --single-configure fixed it. > Do you have time to investigate the cause, and why it depends on > --single-configure, please? Sure, I'll lo

Re: module snippet/link-warning

Collin Funk wrote: > IIRC running 'gnulib-tool --create-testdir --dir testdir1' to create a > testdir of all modules without using ./all-modules will fail due to the > 'link-warning' module. The resulting testdir doesn't build, indeed. But with option --single-configure, it works fine. Do you ha

Re: [PATCH] timevar: add missing timevar.def

Hi Bruno and Paul, Paul Eggert writes: > I was testing all modules that depend on c99, this way: > > ./gnulib-tool -h --create-testdir --dir foo $(cd modules && grep -lx > 'c99' $(git ls-files)) > cd foo > ./configure > make > > The last command fails as follows on Fedora 40 x86-64 whe

Re: [PATCH] timevar: add missing timevar.def

On 2024-07-28 12:07, Bruno Haible wrote: What was the immediate problem that you had with this module? I was testing all modules that depend on c99, this way: ./gnulib-tool -h --create-testdir --dir foo $(cd modules && grep -lx 'c99' $(git ls-files)) cd foo ./configure make The last

[PATCH 2/2] maint: remove ‘coding: utf-8’ on ASCII

These days, ‘coding: utf-8’ tags aren’t needed in general, as Emacs typically defaults to UTF-8, and Emacs itself now has many untagged source files in UTF-8. However, for now let’s just remove the tags when they are on ASCII files. * lib/md5-stream.c, lib/sha1-stream.c, lib/sha256-stream.c: * lib

[PATCH 1/2] maint: remove coding: that did not work

* m4/fnmatch.m4, m4/fpieee.m4, m4/mbrlen.m4, m4/mbrtowc.m4: Remove ineffective coding: strings. --- ChangeLog | 4 m4/fnmatch.m4 | 2 +- m4/fpieee.m4 | 2 +- m4/mbrlen.m4 | 2 +- m4/mbrtowc.m4 | 2 +- 5 files changed, 8 insertions(+), 4 deletions(-) diff --git a/ChangeLog b/ChangeLog i

[PATCH] std-gnu23: new module

The plan is to update the c99 module to depend on std-gnu23 instead of on std-gnu11, and to make std-gnu11 obsolete. A benefit of this will be that c99 will no longer affect C++. For now, though, std-gnu23 is simply a new, optional module. * MODULES.html.sh (func_all_modules): Add it. * modules/std

Re: [PATCH] timevar: add missing timevar.def

Hi Paul, > diff --git a/modules/timevar b/modules/timevar > index ef86fedb29..a460a08de8 100644 > --- a/modules/timevar > +++ b/modules/timevar > @@ -4,6 +4,7 @@ A simple self-profiling module based on timers. > Files: > lib/timevar.h > lib/timevar.c > +lib/timevar.def > > Depends-on: > c99

[PATCH] timevar: add missing timevar.def

* modules/timevar (Files): Add timevar.def. --- ChangeLog | 5 + modules/timevar | 1 + 2 files changed, 6 insertions(+) diff --git a/ChangeLog b/ChangeLog index 2b0b7425bc..764de7d2fb 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,8 @@ +2024-07-28 Paul Eggert + + timevar: a

[PATCH] strftime: improve doc

* lib/strftime.h: In comment, use active voice, reword errno description for clarity, and omit reference to draft POSIX 202x to simplify future maintenance. --- lib/strftime.h | 18 +++--- 1 file changed, 7 insertions(+), 11 deletions(-) diff --git a/lib/strftime.h b/lib/strftime.h in

[gnu.org #1541626] [MATERIALS] Typographical errors in Texinfo format licenses

On Tue Jun 02 04:06:31 2020, asd...@posteo.net wrote: > Hello, > > I have discovered some typographical errors in some licenses in Texinfo > format. There are some '.'s that end sentences but follow capital > letters. These must be changed to '@.'. See also here¹. > > Note: Paul Eggert suggests h