Accidental string changes in anaconda?

[Sarah Wang]

Hi Jeremy,

Thanks for clarifying the issue and understanding our point of view :)

Sarah

Jeremy Katz 写道:

>On Tue, 2004-03-09 at 10:01 +1000, Sarah Wang wrote:
>  
>
>>Jeremy Katz 写道:
>>    
>>
>>>On Mon, 2004-03-08 at 17:29, Christian Rose wrote:
>>>      
>>>
>>>>Ok, so why someone seems to have reverted all string changes that have
>>>>happened the last months in anaconda beats me.
>>>>        
>>>>
>>>Hrmm?  I haven't touched the .pot file recently (since test1-ish). 
>>>Granted, I _should_ push what's current but I haven't recently.  There
>>>haven't been really many changes at all, though.    Hrmm... actually, I
>>>think I might see what happened -- it looks like I somehow missed
>>>committing the changed anaconda.pot file when I updated stuff and then
>>>sarahs remerged stuff last night.  
>>> 
>>>
>>>      
>>>
>>Oops! I thought pot is always the *master*... I thought I was doing the 
>>right thing by keeping all po files consistent with pot files! All 
>>translators are very keen to keep up with the master files. The current 
>>status page showed that the total number of strings across all languages 
>>are greatly varied. If we can't be sure the pot files are most up2date 
>>how can we be sure the po files will be most up2date?
>>    
>>
>
>They're supposed to be.  Unfortunately, anaconda is "special" in how its
>translations are handled which sometimes leads to bogons and things not
>matching up and the like.  The problems with the manual process that is
>anaconda translations...  I could probably write a nice paper on why
>it's bad to try to split source and translations based on it ;)
>
>  
>
>>>Ugh.  I'll clean this up tonight after I get home as best as I can. 
>>>It's going to require going back to the state of things before Sarah's
>>>merging last night, remerge with what's in devel CVS and then push
>>>again.  Which means that anything from today will be lost :/  I _think_
>>>that's probably the best plan, though (you can always then go back a
>>>revision to before my merging).
>>> 
>>>      
>>>
>>Would it be easier just to push the updated pot files to 
>>rhlinux.redhat.com each time when they are updated? Who wants to work on 
>>outdated files?
>>    
>>
>
>I usually push it when I update it, I just managed to not commit or copy
>or something last time.  For a while it was on my todo list to automate
>the process a bit.  But with Fedora and the ability to be more open
>about release stuff in advance, I've switched instead to the plan of
>"push anaconda CVS to be public".  I just need to finish getting a few
>things in order before I can do that.
>
>  
>
>>>Yet another reason why I need to finish getting to where I can put
>>>anaconda's CVS repository in general on rhlinux.redhat.com and just be
>>>done with this nonsense of duplicate CVS repositories for translations
>>>:-)
>>>
>>>      
>>>
>>That would make a lot of sense to have one single CVS repositories. Is 
>>this the case for all other packages as well? Can we have automatica 
>>sync if we must maintain two CVS repositories?
>>    
>>
>
>There are only a very very small number of things that are split like
>this at this point.  And most are for hystorical reasons that will
>slowly migrate over.  So I'm not sure how much it's worth putting effort
>into setting up the automation at this point. 
>
>Jeremy
>
>PS Didn't get to cleaning up the anaconda stuff tonight...  I will take
>care of it first thing in the morning.
>
>
>  
>