On Thu, Apr 23, 2020 at 6:30 PM Amit Kapila <amit.kapila16@gmail.com> wrote:
>
> On Thu, Apr 23, 2020 at 5:37 PM Juan José Santamaría Flecha
> <juanjo.santamaria@gmail.com> wrote:
> >
> > I have composed a small set of queries to test the output with different lc_message settings
(lc_messages_test.sql).Please find attached the output from debug3 logging using both EnumSystemLocalesEx
(lc_messages_EnumSystemLocalesEx.log)and _create_locale (lc_messages_create_locale.log).
> >
>
> Thanks, I will verify these.
>
The result looks good to me. However, I think we should test a few
more locales, especially where we know there is some difference in
what _create_locale returns and what we get via enumerating locales
and using GetLocaleInfoEx. Also, we should test some more locales
with the code page. For ex.
Bemba_Zambia
Bena_Tanzania
Bulgarian_Bulgaria
Swedish_Sweden.1252
Swedish_Sweden
Then, I think we can also test a few where you mentioned that the
language tag is different.
Asu asa Asu
Edo bin Edo
Ewe ee Ewe
Rwa rwk Rwa
BTW, we have a list of code page which can be used for this testing in
below link. I think we can primarily test Windows code page
identifiers (like 1250, 1251, .. 1258) from the link [1].
I think we should backpatch this till 9.5 as I could see the changes
made by commit 0fb54de9 to support MSVC2015 are present in that branch
and the same is mentioned in the commit message. Would you like to
prepare patches (and test those) for back-branches?
I have made few cosmetic changes in the attached patch which includes
adding/editing a few comments, ran pgindent, etc. I have replaced the
reference of "IETF-standardized" with "Unix-style" as we are already
using it at other places in the comments as well.
[1] - https://docs.microsoft.com/en-us/windows/win32/intl/code-page-identifiers
--
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com