[RFC][PATCH v2 00/11][BZ 10871] Month names in genitive case

[RFC][PATCH v2 00/11][BZ 10871] Month names in genitive case

[Rafal Luzynski]

This is the third version of my attempt to fix the bug 10871. [1]
The first version has been sent to this list in January, [2] the
second version has been sent in March. [3]

The main difference between the previous set of patches and this
new one is that implements the backward compatibility by the function
versioning, as requested in [4] and [5].  Since this concept is new
to me please put more attention on reviewing if I implemented
it correctly.  The backward compatibility is provided in patch 2
and patch 6.

The patch set consists of large number of patches.  This is to
help you put your attention on specific changes and skip some noise.
In the final version probably some of them or all will be squashed.

Here is a short description of the following patches:

1. Add the alternative month names support to nl_langinfo(3) (the
ALTMON_... constants.)  There are no major differences between the
previous [6] version and this one except that it has been rebased.

2. Provide backward compatibility for nl_langinfo() and
nl_langinfo_l().  Please check if I implemented it correctly.
Please note that there is no backward compatibility for __nl_langinfo_l().
Do we need it?  In the meantime joseph@codesourcery.com commented: [7]
> __nl_langinfo_l is at a public symbol version and used by libstdc++ 
> (libstdc++ needs to use internal symbols like that for namespace reasons; 
> indeed, we probably need to add more such exports for libstdc++ use).  It 
> needs to stay backward compatible.

3. Rebuild abilists to reflect nl_langinfo() and nl_langinfo_l()
changes - generated automatically.

4. Add tests for alternative month names.  Actually the same as [8],
the only difference is the Subject line which says it's PATCH 04/11
rather than PATCH 2/6.

5. Implement the %OB specifier.  Actually the same as [9], the only
difference is the Subject line which says it's PATCH 05/11 rather
than PATCH 3/6.

6. Provide backward compatibility for strftime(), strftime_l(),
wcsftime(), and wcsftime_l().  The backward compatible versions ignore
the %OB format specifier and for the format specifier %B return the
same results as %OB does for the new version.
Note that there is no backward compatibility for strptime() and
strptime_l(). Do we need a backward compatible version?  See also [10].

7. Rebuild abilists to reflect the changes in strftime(), strftime_l(),
wcsftime(), and wcsftime_l() - generated automatically.

8. Alternative month names NLS data (Polish).  As previously, feel free
to use this patch only for local tests and reject it from the public
repository.  However, please note that you need some NLS data with the
alternative month names, otherwise you will not see any effect of the
previous patches.  There are no difference between the old version [11]
and this patch except that it has been rebased.

9. Alternative month names NLS data (Russian).  The differences between
the old version [12] and this patch are the same as above.

10. Alternative month names NLS data (Ukrainian).  The differences between
the old version [13] and this patch are the same as above.

11. Alternative month names NLS data (Czech).  This is a new patch,
a bonus for our Czech friends. :-)

I'll appreciate your reviews and comments.

Best regards,

Rafal Luzynski

 [1] https://sourceware.org/bugzilla/show_bug.cgi?id=10871
 [2] https://sourceware.org/ml/libc-alpha/2016-01/msg00143.html
 [3] https://sourceware.org/ml/libc-alpha/2016-03/msg00616.html
 [4] https://sourceware.org/ml/libc-alpha/2016-06/msg00009.html
 [5] https://sourceware.org/ml/libc-alpha/2016-06/msg00019.html
 [6] https://sourceware.org/ml/libc-alpha/2016-03/msg00617.html
 [7] https://sourceware.org/bugzilla/show_bug.cgi?id=10871#c58
 [8] https://sourceware.org/ml/libc-alpha/2016-03/msg00618.html
 [9] https://sourceware.org/ml/libc-alpha/2016-03/msg00619.html
[10] https://sourceware.org/bugzilla/show_bug.cgi?id=10871#c59
[11] https://sourceware.org/ml/libc-alpha/2016-03/msg00620.html
[12] https://sourceware.org/ml/libc-alpha/2016-03/msg00621.html
[13] https://sourceware.org/ml/libc-alpha/2016-03/msg00622.html

Re: [RFC][PATCH v2 00/11][BZ 10871] Month names in genitive case

[Rical Jasan]

Hi Rafal,

I can't comment on much of the mechanical changes, but it would be nice
if the new feature was documented.

On 10/17/2016 04:54 PM, Rafal Luzynski wrote:

> 1. Add the alternative month names support to nl_langinfo(3) (the
> ALTMON_... constants.)  There are no major differences between the
> previous [6] version and this one except that it has been rebased.

I think the ALTMON_* constants should be added to the list of constants
in the description of nl_langinfo in manual/locale.texi.

> 5. Implement the %OB specifier.  Actually the same as [9], the only
> difference is the Subject line which says it's PATCH 05/11 rather
> than PATCH 3/6.

I think OB should be added to the list of conversion specifiers in the
description of strftime in manual/time.texi.

nl_langinfo is annotated as being part of XOPEN, and strftime is marked
as ISO, so if OB and the associated ALTMON_* constants aren't parts of
those standards, they should have an "@comment GNU" line placed above
the "@item" lines.  It should say so in their descriptions as well;
e.g., "This is a GNU extension."

Your introductions to patches 01/11 and 05/11 are good, so modifying
them to sound right in the context of the manual should be fine.

Thanks,
Rical

Re: [RFC][PATCH v2 00/11][BZ 10871] Month names in genitive case

[Joseph Myers]

On Mon, 17 Oct 2016, Rical Jasan wrote:

> Hi Rafal,
> 
> I can't comment on much of the mechanical changes, but it would be nice
> if the new feature was documented.

And in addition to the need for documentation in the glibc manual, a NEWS 
entry is needed as well.  (Typically new NEWS entries would be added at 
the bottom of the entries for the next version, immediately above the 
"Security related changes:" header, unless there is a more logical place 
grouping with related changes.)

-- 
Joseph S. Myers
joseph@codesourcery.com

Re: [RFC][PATCH v2 00/11][BZ 10871] Month names in genitive case

[Rafal Luzynski]

Hi Rical and Joseph,

18.10.2016 17:10 Joseph Myers <joseph@codesourcery.com> wrote:
>
>
> On Mon, 17 Oct 2016, Rical Jasan wrote:
>
> > Hi Rafal,
> >
> > I can't comment on much of the mechanical changes, but it would be nice
> > if the new feature was documented.
>
> And in addition to the need for documentation in the glibc manual, a NEWS
> entry is needed as well. (Typically new NEWS entries would be added at
> the bottom of the entries for the next version, immediately above the
> "Security related changes:" header, unless there is a more logical place
> grouping with related changes.)

Thank you for your comments.  I absolutely agree with you that these
changes should be documented but I deliberately skipped the documentation
update because first I'd like to be sure that the change would be
accepted.  Simply I wouldn't like to document the features which would
be rejected.

Also I'm aware that Changelog files will have to be updated, again it's
too early to update them.  And thank you Joseph for mentioning NEWS,
somehow I wasn't aware of the need to update it.

Please don't take this patch set as finished, all of your updates are
valuable.

Regards,

Rafal Luzynski