Python 2.7 is in bugfix-only mode.

Could you expand on the feature request?  The Python locale module delegates to the operating system’s locale system; to the best of my knowledge there is no registry of locale names at the Python level.

In the file locale.py there is a var locale_alias that map the locale names, es_cu isn't there.

Ah, ok.  Sounds good to me.

(I think this can go in stable branches, like other update to registries (e.g. mimetypes), but other core devs may disagree.)

LGTM.

> LGTM.

Thanks for the review, Éric. Do you have time to commit the patch?

I have a busy week-end ahead but I could sneak a little time for this.

Benjamin, does this have to wait for 2.7.5?

Is there some sort of reference for these aliases? Where do they come from?

They come from the X.org project.  See comments in http://hg.python.org/cpython/file/default/Lib/locale.py#l601

I had forgotten there was a makelocalealias.py script; maybe we could run it instead of adding an entry manually.

Yes, please. See what makelocalealias.py does.

Éric Araujo wrote:
> 
> Éric Araujo added the comment:
> 
> They come from the X.org project.  See comments in http://hg.python.org/cpython/file/default/Lib/locale.py#l601
> 
> I had forgotten there was a makelocalealias.py script; maybe we could run it instead of adding an entry manually.

Hi Eric,

let me know if you need help with the script.

The es_CU locale has been added to GNU libc (in version 2.15)[1], but
Tools/i18n/makelocalealias.py script uses the /usr/share/X11/locale/locale.alias file to generate the locale_alias dictionary. I think
/usr/share/X11/locale/locale.alias needs to be updated to recognize
the new GNU libc locales.

See for more information:

* https://bugs.freedesktop.org/show_bug.cgi?id=1544
* http://translate.sourceforge.net/wiki/guide/locales_x11

[1] http://sourceware.org/git/?p=glibc.git;a=commitdiff;h=1eb0b490809a681ea801a1cbd93df5b51a4a47e0;hp=c24a9a8aaea37ac630611eb2c75a6517802a1a9d

On Debian testing, I get a few changes that look like fixes but not es_CU yet:

@@ -807,0 +818,1 @@ locale_alias = {
+    'bokm\xef\xbf\xbd':                     'nb_NO.ISO8859-1',
@@ -822,0 +834,1 @@ locale_alias = {
+    'c.ascii':                              'C',
@@ -936,0 +949,3 @@ locale_alias = {
+    'en_dk':                                'en_DK.ISO8859-1',
+    'en_dk.iso88591':                       'en_DK.ISO8859-1',
+    'en_dk.iso885915':                      'en_DK.ISO8859-15',
@@ -976,1 +991,1 @@ locale_alias = {
-    'english.iso88591':                     'en_EN.ISO8859-1',
+    'english.iso88591':                     'en_US.ISO8859-1',
@@ -1114,0 +1130,1 @@ locale_alias = {
+    'fran\xef\xbf\xbdis':                   'fr_FR.ISO8859-1',
@@ -1164,2 +1180,2 @@ locale_alias = {
-    'hebrew':                               'iw_IL.ISO8859-8',
-    'hebrew.iso88598':                      'iw_IL.ISO8859-8',
+    'hebrew':                               'he_IL.ISO8859-8',
+    'hebrew.iso88598':                      'he_IL.ISO8859-8',
@@ -1246,0 +1263,1 @@ locale_alias = {
+    'km':                                   'km_KH.UTF-8',
@@ -1413,1 +1430,1 @@ locale_alias = {
-    'russian':                              'ru_RU.ISO8859-5',
+    'russian':                              'ru_RU.KOI8-R',
@@ -1427,0 +1445,1 @@ locale_alias = {
+    'sid_et':                               'sid_ET.UTF-8',

I could hunt down the very latest upstream X.org locale.alias file and try the script again.  And/or we can do this one addition (es_CU) now, unless there are issues with Python recognizing a locale that the OS doesn’t know about yet.


MAL: the script needed no change at all to find the file and run! :)  We could make it more similar to token/keyword/etc and have it update locale.py in 3.4+.  Also maybe make updating the aliases a part of feature release workflow, or even bugfix releases.

This issue is superseded by issue20079.