Bug 650028 - RFE allow lists of locales for language groups
Summary: RFE allow lists of locales for language groups
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: Fedora
Classification: Fedora
Component: anaconda
Version: rawhide
Hardware: Unspecified
OS: Unspecified
low
medium
Target Milestone: ---
Assignee: Anaconda Maintenance Team
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks: 667898
TreeView+ depends on / blocked
 
Reported: 2010-11-05 02:57 UTC by Peng Wu
Modified: 2014-01-21 23:04 UTC (History)
5 users (show)

Fixed In Version:
Clone Of:
: 667898 (view as bug list)
Environment:
Last Closed: 2011-03-09 21:13:08 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)

Description Peng Wu 2010-11-05 02:57:01 UTC
In the screen of software selection, there is a category named "Languages", there is a "Chinese Support".
Could we split this into "Simplified Chinese Support" and "Traditional Chinese Support" in Language Categories, as they are two different languages?

Here are the benefits after this split:
1. Only install fonts for one language. For example, wqy-zenhei-fonts for Simplified Chinese, and cjkuni-fonts for Traditional Chinese, no need to install both.
2. Only install input methods for one language. For example, pinyin input method for Simplified Chinese, and chewing input method for Traditional Chinese, also no need to install both.
3. Only install one kind of language pack. For example, *-zh_CN for Simplified Chinese, and *-zh_TW for Traditional Chinese, less packages as above. Example package name: openoffice.org-langpack-zh_CN, koffice-langpack-zh_CN.

Comment 1 Chris Lumens 2010-11-05 14:29:20 UTC
Comps controls what package groups we list.

Comment 2 Bill Nottingham 2010-11-05 19:10:29 UTC
Surely, they are different scripts, not different languages?  (Although, in practice this probably doesn't make any difference.)

Comment 3 Peng Wu 2010-11-08 02:20:16 UTC
yes, they have different writing glyphs, Unicode code points and input methods.

Comment 4 Jens Petersen 2010-11-08 04:47:43 UTC
Ok, I talked to Peng Wu about this, and don't think
the intent was quite communicated fully.

We would like to ask if anaconda could support
lists of locales for language groups.

Then we could have Simplified Chinese for "zh_CN,zh_SG"
and Traditional Chinese for "zh_TW,zh_HK", etc.  Would
that be acceptable.  We might be able to contribute
a patch to do that if it helps.

Otherwise we will just have to duplicate Chinese groups
which will increase the comps maintenance burden.

Comment 5 Jens Petersen 2010-12-02 05:35:17 UTC
Any comments on this proprosal?

Comment 6 Jens Petersen 2010-12-03 03:16:58 UTC
Or can we get yum-langpacks working with anaconda now?

Then we could finally start to remove langpacks from comps!

Comment 7 Peng Wu 2010-12-20 05:10:47 UTC
Here are the draft patch for this package group split:
http://pwu.fedorapeople.org/comps/split-package-group.patch

Feel free to comment it.

Comment 8 Bill Nottingham 2010-12-20 20:24:12 UTC
Jens - what's involved in comment #6?

Comment 9 Jens Petersen 2010-12-21 00:26:03 UTC
(In reply to comment #8)
> Jens - what's involved in comment #6?

Will input from the anaconda team for a start...
and testing I suppose. :)

Comment 10 Jens Petersen 2010-12-21 00:28:50 UTC
> input from the anaconda team for a start...

Would it be better to propose as a Feature?

Comment 11 Ding-Yi Chen 2011-01-03 23:43:22 UTC
(In reply to comment #3)
> yes, they have different writing glyphs, Unicode code points and input methods.

Well, they share roughly 60% of characters (though there are preferences of glyphs, see http://en.wikipedia.org/wiki/Variant_Chinese_character 
and most phrases are same in both places.

However, I do support the split as Peng Wu suggestion. People who need to have both just tick both. :-)

Comment 12 Peng Wu 2011-01-07 05:26:04 UTC
(In reply to comment #11)
> However, I do support the split as Peng Wu suggestion. People who need to have
> both just tick both. :-)

Hi, thanks for commenting. :-)

Comment 13 Peng Wu 2011-01-07 05:30:26 UTC
Polling about this change on Fedora Chinese mailing list is finished.
Some users like this comps split for Simplified Chinese and Traditional Chinese Support, and seems no strong arguments for not splitting.
URL for discussion mail:
http://lists.fedoraproject.org/pipermail/chinese/2010-December/004795.html
Could we try this comps split on Fedora 15?

Comment 14 Peng Wu 2011-01-07 07:40:44 UTC
(In reply to comment #13)
> Polling about this change on Fedora Chinese mailing list is finished.
> Some users like this comps split for Simplified Chinese and Traditional Chinese
> Support, and seems no strong arguments for not splitting.
> URL for discussion mail:
> http://lists.fedoraproject.org/pipermail/chinese/2010-December/004795.html
> Could we try this comps split on Fedora 15?

Sorry, I mistake this bug with comps split. Just cloned this bug to track comps split. (#667898)

This bug is left for anaconda RFE, to let anaconda work with yum-langpacks.

Comment 15 Chris Lumens 2011-03-07 16:47:27 UTC
What exactly do you want anaconda to do?  How would we need to interact with yum-langpacks?

Comment 16 Peng Wu 2011-03-08 02:59:09 UTC
(In reply to comment #15)
> What exactly do you want anaconda to do?  How would we need to interact with
> yum-langpacks?

When anaconda install Fedora, uses yum with langpacks to install packages.
Let yum language packs to handle langpacks rpm package installation, for example,
openoffice.org-langpack-en.

After switching to yum langpacks, it seems that we can removes many entries in comps.xml.in:
      <packagereq type="conditional" requires="autocorr-en">autocorr-zh</packagereq>
      <packagereq type="conditional" requires="eclipse-platform">eclipse-nls-zh</packagereq>

So we can reduces the entries in comps.xml.in, by telling yum langpacks to install the corresponding rpms for the user's languages.

Comment 17 Jens Petersen 2011-03-08 07:51:30 UTC
Sigh - this has nothing to do with yum-langpacks!

Please see comment 4.

The question was: if anaconda could support multiple locales for language groups.

Then we could have a "Simplified Chinese Support" language group for zh_CN,zh_SG;
and "Traditional Chinese Support" group for zh_HK,zh_TW; etc.

Comment 18 Chris Lumens 2011-03-09 21:13:08 UTC
anaconda bases the default package selection on what language you selected at the beginning of installation.  We then go through all the comps groups and select every one that has a langonly= attribute that matches the currently selected language.

You can break up the language groups however you want in comps, and make whatever package changes you want too.  However, you can really only ever have one language selected during anaconda.  If you can come up with a way to use the group definition and langonly= attributes to accomplish what you want to, go for it.  But I'm not going to add yet more language-specific data that only anaconda will know about to solve this problem.  We are trying to get away from that.


Note You need to log in before you can comment on or make changes to this bug.