Bugzilla Bug 64201

Gentoo documentation team is using incorrect directory name for Korean
documentation.

Reproducible: Always
Steps to Reproduce:
1. Visit http://www.gentoo.org/doc/ko/index.xml
2. Visit http://www.gentoo.org/doc/kr/index.xml
3.

Actual Results:  
I can see http://www.gentoo.org/doc/kr/index.xml only.

Expected Results:  
I heard that ko stands for Korean(a language) and kr stands for Korea(a
country). I think ko, the name of Korean documentation directory, should be kr.

Debian people use ko for Korean documentation. (See
http://www.debian.org/index.ko.html)
And Gentoo developers are using ja for Japanese. (ja stands for Japanese and jp
stands for japan.)
Even though it's a minor problem, I think it's not difficult to fix.

Additionally, some of the documentation is being translated with Korean Gentoo
Translation Wiki. Some of them will be bug-reported soon.
The address of the translation wiki is:
http://www.gentoo.or.kr/~nabugoon/wiki/wiki.php/GentooDocs

You're right. Your report prompted me to check all dirs against ISO-639 and we
have been naughty.
BTW, we look forward to receiving your Korean translations. Good luck.

FYI, ISO-639 is the reference for language codes, not debian.
http://www.loc.gov/standards/iso639-2/

`curl -s http://www.loc.gov/standards/iso639-2/ISO-639-2_values_8bits.txt|awk
-F'|' '{if ($3) print toupper($3),$4}'|sort|uniq`
will give you the official list of 2-letter codes sorted on the code.

Here are the (potentially) wrong directories:
CN is not an ISO639 code, let's keep it for simplified Chinese which is not
listed on the ISO page
CZ has been replaced by CS, should be removed (Czech)
GR has been replaced by EL, should be removed (Greek)
KR should be KO, Korean,  KR is for Kanuri
SE should be SV, Swedish,  SE is for Northern Sami
TW should be ZH, Chinese,  TW is for Twi

Fixing it is not as straightforward as you think. Well, fixing it is easy, but
committing will not be as easy because of obsolete documents that do not follow
the current DTD. Non-DTD-compliant docs will be rejected by CVS. I suggest we
put those invalid obsolete docs in the attic as they are likely to contain
outdated or even wrong information anyway.

Is there anyway to get rid of an unwanted dir in CVS, or at least put it in the
attic so it does not show up online anymore?

As far as I know, removing directory is not a feature of CVS. You could put
files in unwanted directories into the attic, instead. (I can't see anything
when I type http://www.gentoo.org/doc/cn/ because there isn't index.xml in that
directory. So I think putting files in the attic would be enough.)

I agree with Xavier that we need to be consistent in our approach. I also agree
on moving the old files to the Attic and let the translation teams themselves
fix the non-DTD validity.

BennyC is busy with renaming cn and tw respectively to zh_cn and zh_tw
This is consistent with the pt and pt_br that we already use to distinguish Portuguese variants

Kurt, CC'ing to you so that you can update group names and cvs access rights
accordingly.
This should have been noticed earlier.
Sorry about the extra work.

BTW, the GWN is using BE instead of NL for Dutch. It would be nice to be
consistent and have that fixed, but I believe I have no access rights in there.

Movable files have been moved to the right location and outdated ones have been
sent to the attic.

Those with no CVS access can see the changes and find oudated files at
http://www.gentoo.org/cgi-bin/viewcvs.cgi/?root=doc

I realized that just copying unwanted files into the attic doesn't work
perfectly. I can see http://www.gentoo.org/doc/kr/index.xml even though
index.xml in kr is in the attic, because it's just copied in the attic, not
removed. I think you should delete unwanted files also.

They're not copied, they're removed. But the webservers themselves don't remove
files that aren't in CVS anymore apparently (i.e. they just let those linger).

Kurt, do the web nodes clean their files or are the old files just lingering in
axkit's cache? When we remove a file, we mean it and we expect it to be removed
from the wwww.g.o as well.
FYI, users find those files via Google.

Kurt: could you purge all files on the webnodes that aren't in CVS anymore? It
seems that files we remove from CVS stay on the webnodes.

sorry -- didn't see this bug earlier.  I added the --delete option to the rsync
command that updates the nodes.  This should fix it.  Let me know if it
doesn't.

It's fixed. Thanks. I suppose the bug's now closeable