From: Steve Kowalik <steven@wedontsleep.org> To: Ying Chun Guo/China/IBM@IBMCN, "Elizabeth K. Joseph" <lyz@princessleia.com> Cc: Openstack-i18n Openstack-i18n <openstack-i18n@lists.openstack.org> Date: 08/21/2015 01:50 PM Subject: Re: [Openstack-i18n] Zanata trial starts ! Please register and try - requirements before production
Hi Daisy,
On 10/08/15 19:43, Ying Chun Guo wrote:
1. Language codes
In Transifex, there are 44 languages which have progress. Some of the language codes only contain language codes, the others contain language codes plus locale codes. The language code format in Transifex is language codes ( in small case ) + underline ("_") + locale codes ( in big case ), for example, ko_KR. The language code format in Zanata is language codes ( in small case )
Thank you, Steve. Could you add three more languages? These languages have good progress in Horizon (only) but bad progress in the overall view. So I missed these languages in the previous list. Please add them. Filipino ( fil ) Georgian ( ka_GE) Telugu (India) (te_IN) BTW, I don't know if changing underscore to dash will cause any unexpected issues. Have you ever tried to pull translations from Zanata and then store in git repo ? I don't know if the folder names ( same as language codes ) would be changed. Or you use some special codes to change dash back to underscore while pulling translations and before storing in git repo ? Regards Daisy Steve Kowalik <steven@wedontsleep.org> wrote on 08/21/2015 01:50:24 PM: +
hyphen ("-") + locale codes ( in small case ), for example, ko-kr. I don't know if the different format will cause issues because language codes are the folder names in OpenStack git repo.
It did not with Zanata 3.6.2, since we were telling the Zanata client to
map from ko-kr to ko_KR. Since the move to 3.7.1, we can not perform the
mapping, so moving forward we will use ko-KR.
Refer to Horizon repo [1]. Please help to verify that.
Besides the format differences, I also notice some languages are missed in Zanata. They are: bg_BG, gu, ks, kok, mai, mni, mr, pa_IN, ta, ur. These languages should be added. And in Zanata, there are some languages which should be removed from Zanata. They are: bn, nl, en-au, en-Au, en-aU, en-gb, fi, ko, pl-P, pl-p, pt-B, sl-, te.
Considering the format differences and the language lists differences, I suggest that we update the database to configure the language codes just same as Transifex. And then export all the translations of these languages into Zanata.
I have written down the list of language codes in Transfiex which have any progress in the trail feedback page. [2]
I have gone through your list, and converted them into the locale code that Zanata expects (a dash, rather than an underscore).
When Elizabeth and I set up the new translate-dev, I will be creating the following locales on it:
http://paste.openstack.org/show/423139/
Can you look over that list and let me know if there are any you wish to
add or remove?
Cheers, -- Steve <bob2> joy hasn't spoken on IRC in 45 seconds, he must be MIA!