Fwd: textinfo-6.8: Test failing

[Chris Clayton]

Sorry, forgot to reply-all, so not cc'd to the mailing list.


-------- Forwarded Message --------
Subject: Re: textinfo-6.8: Test failing
Date: Mon, 5 Jul 2021 10:57:21 +0100
From: Chris Clayton <chris2553@googlemail.com>
To: Gavin Smith <gavinsmith0123@gmail.com>



On 04/07/2021 22:53, Gavin Smith wrote:
> On Sun, Jul 4, 2021 at 7:04 PM Chris Clayton <chris2553@googlemail.com> wrote:
>> Thanks for the reply, Gavin.
>>
>> Right, thanks. the two commands give:
>>
>> $ locale
>> LANG=en_GB.UTF-8
>> LC_CTYPE="en_GB.UTF-8"
>> LC_NUMERIC="en_GB.UTF-8"
>> LC_TIME="en_GB.UTF-8"
>> LC_COLLATE="en_GB.UTF-8"
>> LC_MONETARY="en_GB.UTF-8"
>> LC_MESSAGES="en_GB.UTF-8"
>> LC_PAPER="en_GB.UTF-8"bug-texinfo@gnu.org
>> LC_NAME="en_GB.UTF-8"
>> LC_ADDRESS="en_GB.UTF-8"
>> LC_TELEPHONE="en_GB.UTF-8"
>> LC_MEASUREMENT="en_GB.UTF-8"
>> LC_IDENTIFICATION="en_GB.UTF-8"
>> LC_ALL=
>>
>> $ locale -a
>> C
>> en_GB
>> en_GB.iso88591
>> en_GB.utf8
>> en_GB.UTF-8
>> en_US.utf8
>> POSIX
>>
>> I've just built texinfo-6.7 and the test suite succeeds.  I'll investigate 
>> what changed in that test. Thanks.
> 
> This isn't what I expected as you have a locale "en_US.utf8" which I
> understand is the same thing as "en_US.UTF-8", which is one of the
> locales tried in 'gdt' in tp/Texinfo/Report.pm.

That's not a surprise to me. My system was created by me using the Linux From 
Scratch guidance.
I freely admit that I was all at sea when it came to which locales I was likely 
to need, so I tried this, that and
t'other and when I found a setup that seemed to suffice, I quit while I was 
ahead.

That was years ago and I've learned alot in that time. However, "good enough is 
good enough" and "if it ain't broke,
don't fix it" where maxims I applied throughout most of my 30+ year career in 
IT. Now it's broke, so I've regenerated
the locale stuff on my system to include en_US.UTF-8 and the texinfo-6.8 now 
completes without error.

So it's close to a case of PEBKAC. It might, howver, be an idea for the test to 
be skipped if none of the locales it
requires to succeed are present.

Thanks for your help.

> You could try the
> following change:-
> 
> diff --git a/tp/Texinfo/Report.pm b/tp/Texinfo/Report.pm
> index d72b5ba7ba..1a30a50f23 100644
> --- a/tp/Texinfo/Report.pm
> +++ b/tp/Texinfo/Report.pm
> @@ -251,7 +251,7 @@ sub gdt($$;$$)
>    # This may not work if a locale named here doesn't exist on the system.
>    # Set LC_ALL rather than LC_MESSAGES as LC_MESSAGES may not be supported
>    # on Perl for MS-Windows.
> -  for my $try ('en_US.UTF-8', 'en_US') {
> +  for my $try ('en_US.utf8', 'en_US') {
>      my $locale = POSIX::setlocale(LC_ALL, $try);
>      last if $locale;
>    }
> 
> - to see if it makes a difference. If it is this part of the code that
> is causing the test failure then it would be due to upgrading the
> bundled version of libintl-perl, due to a fix for the issue at
> https://rt.cpan.org/Public/Bug/Display.html?id=81315
> 
> I reported the issue here
> https://lists.gnu.org/archive/html/bug-texinfo/2020-11/msg00088.html -
> but nobody seemed to have an idea of what should be done.
>