Re: texinfo-6.0.90 pretest

[Norbert Preining]

Hi Patrice, hi Gavin,

> >  address@hidden erreur->
> >  address@hidden error->
> > and similar things concerning translations.
> 
> Are there really other differences?  @error is handled quite
> specifically since it is the only @-command that is especially translated.

Yes, there are many other differences, all French words were untranslated
in my case.

> What do you install when you make the French locale available?

Good question, on a Debian system one can select the set of
available locales. If one is not available, then calls to
setlocale, or locale etc will not succeed. Perl would give a warning
about unavailable locales etc.

> Normally the in-document translations are not taken from the locales
> directory, but from a list of texinfo specific directories.  And their 
> use is not triggered by the locale setting but by @documentlanguange.

This was my understanding, too, and I think this is correct. So I was
*very* surprised that the available locales chnage the outcome.

> > These two tests succeed when I make a French locale *available*
> > but not set by default.
> 
> Maybe we need a check for the available locales so we can skip those tests.

I agree with Patrice that it would be a *bad* idea to have
tests, or for that matter any level of operation, depending on the
available locales. This should all be completely independent of
the locales.


> >>  test_scripts/formatting_documentlanguage_set_option.sh
> >
> > This error remains, and contains the following diff:
> >  -  Ce document a &eacute;t&eacute; g&eacute;n&eacute;r&eacute; <em>a sunny 
> > day</em> en utilisant <a 
> > href="http://www.gnu.org/software/texinfo/";><em>texi2any</em></a>
> >  +  This document was generated on <em>a sunny day</em> using <a 
> > href="http://www.gnu.org/software/texinfo/";><em>texi2any</em></a>.
> >  </font></p>
> 
> This is a mystery. First of all, what French locales did you install?
> Was it fr_FR or fr_FR.UTF-8? (Don't know if that makes a difference,
> but I'm casting around in the dark.)

As I said, I have used fr_FR.UTF-8. Additionally, I do not have fr_FR
(only, as is) available, although I believe it should work, too.

> Is there anything else in the diff? For example, when I replicate this
> error by changing the language code to "xx" in the test file:

I don't have any other diff, the test-suite.log only gives the
diff listed:
-  Ce document a &eacute;t&eacute; g&eacute;n&eacute;r&eacute; <em>a sunny 
day</em> en utilisant <a 
href="http://www.gnu.org/software/texinfo/";><em>texi2any</em></a>
+  This document was generated on <em>a sunny day</em> using <a 
href="http://www.gnu.org/software/texinfo/";><em>texi2any</em></a>.

> Note the "texi2any: warning: xx is not a valid language code" warning:
> do you get a similar error?

No.

> > Is there anything else in the diff? For example, when I replicate this
> > error by changing the language code to "xx" in the test file:
> 
> I think that this is not really a replication of the issue.  The issue
> is probably along a in-document string translation file not found, or 
> not the correct one, maybe an old one, it is not that the language is
> not found.

That sounds convincing. I would be surprised if texinfo would use
the locale strings from the OS localization.

> And I agree that it is very strange that install-info tests start
> failing now, and, for example not for me though I use a french locale.

I switched now to fr_FR.UTF-8 (LC_ALL=fr_FR.UTF-8) and re-run the check
(make check) with the effect that I see:

FAIL: ii-0018-test
...
+ diff ../../install-info/tests/ii-0018-expected-dir-file ii18-jGczxXVx
1,3c1,4
< This is the file .../info/dir, which contains the
< topmost node of the Info hierarchy, called (dir)Top.
< The first time you invoke Info you start off looking at this node.
---
> Ceci est le fichier .../info/dir, qui contient le nœud le
> plus haut de la hiérarchie Info. Ce nœud est appelé (dir)Top.
> C'est de ce nœud que vous démarrez la première fois que
> vous utilisez Info.
.....

and similar for ii-0019-test (these are the two tests that failed).


Just for completeness, here is the ./configure invocation that is
generated by the debian packaging:
AWK=awk ./configure CFLAGS="-g -O2 -fstack-protector-strong -Wformat 
-Werror=format-security" CPPFLAGS="-D_FORTIFY_SOURCE=2" CXXFLAGS="-g -O2 
-fstack-protector-strong -Wformat -Werror=format-security" FCFLAGS="-g -O2 
-fstack-protector-strong" FFLAGS="-g -O2 -fstack-protector-strong" GCJFLAGS="-g 
-O2 -fstack-protector-strong" LDFLAGS="-Wl,-z,relro" OBJCFLAGS="-g -O2 
-fstack-protector-strong -Wformat -Werror=format-security" OBJCXXFLAGS="-g -O2 
-fstack-protector-strong -Wformat -Werror=format-security"    
--with-external-Text-Unidecode=yes   --with-external-libintl-perl=yes   
--prefix=/usr   --libexecdir='${prefix}/lib'   --infodir='${prefix}/share/info' 
--mandir='${prefix}/share/man'   --build=x86_64-linux-gnu 
--host=x86_64-linux-gnu

make
make check


I have no idea what is wrong here ...


Norbert

------------------------------------------------------------------------
PREINING, Norbert                               http://www.preining.info
JAIST, Japan                                 TeX Live & Debian Developer
GPG: 0x860CDC13   fp: F7D8 A928 26E3 16A1 9FA0  ACF0 6CAC A448 860C DC13
------------------------------------------------------------------------