Regression: GL_STDLIB_H_MODULE_INDICATOR_DEFAULTS is m4_require'd but not m4_defun'd

Regression: GL_STDLIB_H_MODULE_INDICATOR_DEFAULTS is m4_require'd but not m4_defun'd

[Colin Watson]

I recently tried to update man-db to current Gnulib, running in a Debian
unstable chroot, and I ran into some Autoconf macro issues that I don't
entirely understand.  I've boiled it down as far as this, with current
Gnulib master (4bed390be8) and Autoconf 2.71:

  $ ./gnulib-tool --dir x --create-testdir btowc
  [...]
  executing autoconf
  configure.ac:584: warning: GL_STDLIB_H_MODULE_INDICATOR_DEFAULTS is m4_require'd but not m4_defun'd
  glm4/wchar_h.m4:144: gl_WCHAR_H_REQUIRE_DEFAULTS is expanded from...
  glm4/wchar_h.m4:132: gl_WCHAR_MODULE_INDICATOR is expanded from...
  configure.ac:35: gl_INIT is expanded from...
  configure.ac:584: the top level

And indeed the resulting configure script contains an unexpanded macro,
resulting in:

  ./configure: line 8730: GL_STDLIB_H_MODULE_INDICATOR_DEFAULTS: command not found

I initially suspected an incompatibility with Autoconf >= 2.70, since
there were lots of changes there, but I tried the same gnulib-tool test
with Autoconf 2.69 and it does the same thing.

When I "git blame m4/wchar_h.m4", I find the recent commit 4bed390be8,
which added the line
"m4_require(GL_MODULE_INDICATOR_PREFIX[_STDLIB_H_MODULE_INDICATOR_DEFAULTS])".
And indeed, when I back out that commit, things work again; but it was
tagged as a regression fix, so evidently something else breaks.  CCing
Bruno since this was their commit and they might want to have a look.

Thanks,

-- 
Colin Watson (he/him)                              [cjwatson@debian.org]

Re: Regression: GL_STDLIB_H_MODULE_INDICATOR_DEFAULTS is m4_require'd but not m4_defun'd

[Bruno Haible]

Colin Watson wrote:
> I recently tried to update man-db to current Gnulib, running in a Debian
> unstable chroot, and I ran into some Autoconf macro issues that I don't
> entirely understand.  I've boiled it down as far as this, with current
> Gnulib master (4bed390be8) and Autoconf 2.71:
> 
>   $ ./gnulib-tool --dir x --create-testdir btowc
>   [...]
>   executing autoconf
>   configure.ac:584: warning: GL_STDLIB_H_MODULE_INDICATOR_DEFAULTS is m4_require'd but not m4_defun'd
>   glm4/wchar_h.m4:144: gl_WCHAR_H_REQUIRE_DEFAULTS is expanded from...
>   glm4/wchar_h.m4:132: gl_WCHAR_MODULE_INDICATOR is expanded from...
>   configure.ac:35: gl_INIT is expanded from...
>   configure.ac:584: the top level
> 
> And indeed the resulting configure script contains an unexpanded macro,
> resulting in:
> 
>   ./configure: line 8730: GL_STDLIB_H_MODULE_INDICATOR_DEFAULTS: command not found

Thanks for the report and perfectly reproducible test case.

The patch below fixes it, and still works with libidn (scenario reported by
Simon).


2021-09-18  Bruno Haible  <bruno@clisp.org>

	string, wchar: Don't require undefined m4 macros (regr. today).
	Reported by Colin Watson <cjwatson@debian.org> in
	<https://lists.gnu.org/archive/html/bug-gnulib/2021-09/msg00078.html>.
	* m4/string_h.m4 (gl_STRING_H_REQUIRE_DEFAULTS): Don't use m4_require
	directly. Instead, invoke gl_STDLIB_H_REQUIRE_DEFAULTS.
	* m4/wchar_h.m4 (gl_WCHAR_H_REQUIRE_DEFAULTS): Likewise.

diff --git a/m4/string_h.m4 b/m4/string_h.m4
index 9871dac8b..e88ac9ca8 100644
--- a/m4/string_h.m4
+++ b/m4/string_h.m4
@@ -5,7 +5,7 @@
 # gives unlimited permission to copy and/or distribute it,
 # with or without modifications, as long as this notice is preserved.
 
-# serial 33
+# serial 34
 
 # Written by Paul Eggert.
 
@@ -94,7 +94,7 @@ AC_DEFUN([gl_STRING_H_REQUIRE_DEFAULTS],
   ])
   m4_require(GL_MODULE_INDICATOR_PREFIX[_STRING_H_MODULE_INDICATOR_DEFAULTS])
   dnl Make sure the shell variable for GNULIB_FREE_POSIX is initialized.
-  m4_require(GL_MODULE_INDICATOR_PREFIX[_STDLIB_H_MODULE_INDICATOR_DEFAULTS])
+  gl_STDLIB_H_REQUIRE_DEFAULTS
   AC_REQUIRE([gl_STRING_H_DEFAULTS])
 ])
 
diff --git a/m4/wchar_h.m4 b/m4/wchar_h.m4
index d69dbe67d..9529cdd05 100644
--- a/m4/wchar_h.m4
+++ b/m4/wchar_h.m4
@@ -7,7 +7,7 @@ dnl with or without modifications, as long as this notice is preserved.
 
 dnl Written by Eric Blake.
 
-# wchar_h.m4 serial 54
+# wchar_h.m4 serial 55
 
 AC_DEFUN_ONCE([gl_WCHAR_H],
 [
@@ -190,7 +190,7 @@ AC_DEFUN([gl_WCHAR_H_REQUIRE_DEFAULTS],
   ])
   m4_require(GL_MODULE_INDICATOR_PREFIX[_WCHAR_H_MODULE_INDICATOR_DEFAULTS])
   dnl Make sure the shell variable for GNULIB_FREE_POSIX is initialized.
-  m4_require(GL_MODULE_INDICATOR_PREFIX[_STDLIB_H_MODULE_INDICATOR_DEFAULTS])
+  gl_STDLIB_H_REQUIRE_DEFAULTS
   AC_REQUIRE([gl_WCHAR_H_DEFAULTS])
 ])
 

Re: gendering

[Bruno Haible]

This gendering stuff is generally off-topic on this list, but it came up here,
therefore I'm replying here. I hope the thread will be short, since so much
has already been written on this topic.

Hi Colin,

You wrote:
> CCing Bruno since this was their commit and they might want to have a look.

While I understand the motivation of gendered language, and myself use
"they/their/them" when talking about an *unidentified* individual in English
— so as to be inclusive and not offend the Sarahs and Leahs of this world —,
in return I ask that you and others treat me as a human. Not as a number.
Not as an email address. But as an *identified* person. In this role, I have
a first name and a last name. And as usual in Western culture, the first
name gives a good clue about the gender. When my first name is used in the
US, it has 100% probability of being a male's first name [1]. Like yours
has a 99.6% probability [2]. Therefore please apply a bit of common sense
when talking about me.

Some people might say "just add (he/his) to your signature". I say NO.
For the vast majority of people in Western culture, the gender can be
inferred from the first name. The Sarahs and Leahs of this world are a
minority, and it's wrong for a minority to dictate what the majority
should do.

So, please be respectful to me. Thanks.

Bruno

[1] https://git.savannah.gnu.org/gitweb/?p=gettext.git;a=blob;f=libtextstyle/examples/color-hello/names.c;h=bf8af3630a0e86198da3d0f0c93ba8e0318d048d;hb=HEAD#l13244
[2] https://git.savannah.gnu.org/gitweb/?p=gettext.git;a=blob;f=libtextstyle/examples/color-hello/names.c;h=bf8af3630a0e86198da3d0f0c93ba8e0318d048d;hb=HEAD#l18142

Re: gendering

[Jeffrey Walton]

On Sat, Sep 18, 2021 at 9:19 PM Bruno Haible <bruno@clisp.org> wrote:
>
> This gendering stuff is generally off-topic on this list, but it came up here,
> therefore I'm replying here. I hope the thread will be short, since so much
> has already been written on this topic.
>
> Hi Colin,
>
> You wrote:
> > CCing Bruno since this was their commit and they might want to have a look.
>
> While I understand the motivation of gendered language, and myself use
> "they/their/them" when talking about an *unidentified* individual in English
> — so as to be inclusive and not offend the Sarahs and Leahs of this world —,
> in return I ask that you and others treat me as a human. Not as a number.
> Not as an email address. But as an *identified* person. In this role, I have
> a first name and a last name. And as usual in Western culture, the first
> name gives a good clue about the gender. When my first name is used in the
> US, it has 100% probability of being a male's first name [1]. Like yours
> has a 99.6% probability [2]. Therefore please apply a bit of common sense
> when talking about me.
>
> Some people might say "just add (he/his) to your signature". I say NO.
> For the vast majority of people in Western culture, the gender can be
> inferred from the first name. The Sarahs and Leahs of this world are a
> minority, and it's wrong for a minority to dictate what the majority
> should do.

+1.

Using they/them is also disrespective of folks in the LGBTQ community
who wish to be identified in a particular way. For example, if someone
is born a little teapot with a spout but calls themselves Sarah, it is
considered disrespectful to not acknowledge their preferred gender.
Also see https://blogs.scientificamerican.com/voices/actually-we-should-not-all-use-they-them-pronouns/.

I think Dr. Stallman had a pretty good idea... Call the folks who are
not sure what their identity is by "pers". It avoids the grammatical
problems with subject/pronoun disagreement. Subject/pronoun
disagreement is a living hell for me. It is nearly impossible for me
to read such a sentence without breaking and reading the sentence
multiple times. Also see
https://stallman.org/articles/genderless-pronouns.html.

Jeff

Re: Regression: GL_STDLIB_H_MODULE_INDICATOR_DEFAULTS is m4_require'd but not m4_defun'd

[Colin Watson]

On Sun, Sep 19, 2021 at 03:15:45AM +0200, Bruno Haible wrote:
> The patch below fixes it, and still works with libidn (scenario reported by
> Simon).
> 
> 
> 2021-09-18  Bruno Haible  <bruno@clisp.org>
> 
> 	string, wchar: Don't require undefined m4 macros (regr. today).

Thanks for the quick response.  I can confirm that this fixes the
original problem I saw.

-- 
Colin Watson (he/him)                              [cjwatson@debian.org]

Re: gendering

[Colin Watson]

Hi Bruno,

For some years I've been in the habit of using indefinite singular
"they" when I'm not certain what pronouns a given person uses and when
that also isn't directly relevant to the conversation; this is distinct
from its use by non-binary and other people who use it as a definite
pronoun referring to themselves individually.  I've found that this
habit works well as a first approximation among the people I spend most
of my time talking with, and has a lower rate of people telling me I got
it wrong than when I try to guess, so I don't plan to change my default
use of language here.

However, I'm sorry I got it wrong in this case.  When somebody has told
me what pronouns they use then I always try to use those instead of a
placeholder, so I will of course refer to you using "he" etc. pronouns
in future.

Regards,

-- 
Colin Watson (he/him)                              [cjwatson@debian.org]