Accidental string changes in anaconda?

[Christian Rose]

Ok, so why someone seems to have reverted all string changes that have
happened the last months in anaconda beats me.

Anaconda has had many string changes of the type:

mention of tool "redhat-config-xxx" => mention of tool
"system-config-xxx"

mention of "Red Hat Linux" => mention of %s (which might be replaced by
"Red Hat Linux" or "Fedora Core" by a compile time switch)

Now all those string changes are reverted, at least in the anaconda po
files, so we're back to square zero: anaconda happily tells users to use
redhat-config-xxx tools, and calls the product "Red Hat Linux" in a
hardcoded way.

Accident, or some strange strategy that beats me?


Christian

[Federico Bernardi]

I saw it and I submitted a bug report
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=117759
Maybe was there a problem with CVS server?
Could someone commit old code and overwrite the new one?
I really don't know... I saw in the translations some %s but in the original
part were "Red Hat"... :(

Federico

----- Original Message ----- 
From: "Christian Rose" <menthos(a)menthos.com&gt;
To: <fedora-trans-list(a)redhat.com>; <fedora-devel-list(a)redhat.com&gt;
Sent: Monday, March 08, 2004 11:29 PM
Subject: Accidental string changes in anaconda?


>  Ok, so why someone seems to have reverted all string changes that
> have
>  happened the last months in anaconda beats me.
>
>  Anaconda has had many string changes of the type:
>
>  mention of tool "redhat-config-xxx" => mention of tool
>  "system-config-xxx"
>
>  mention of "Red Hat Linux" => mention of %s (which might be replaced by
>  "Red Hat Linux" or "Fedora Core" by a compile time switch)
>
>  Now all those string changes are reverted, at least in the anaconda po
>  files, so we're back to square zero: anaconda happily tells users to use
>  redhat-config-xxx tools, and calls the product "Red Hat Linux" in a
>  hardcoded way.
>
>  Accident, or some strange strategy that beats me?
>
>
>  Christian
>
>
>  --
>  Fedora-trans-list mailing list
>  Fedora-trans-list(a)redhat.com
>  http://www.redhat.com/mailman/listinfo/fedora-trans-list
>
>

[Jeremy Katz]

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.  

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).

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
:-)

Jeremy

[Sarah Wang]

Hi Jeremy,

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?

> 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?

> 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?

Regards,

Sarah

>   
>

[Jeremy Katz]

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.

[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.
>
>
>   
>

[Bart Martens]

On Tue, 2004-03-09 at 01:01, Sarah Wang wrote:
>  Who wants to work on 
>  outdated files? 
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=109374
Translations bug for anaconda. This bug is NEW since november. Yesterday
Stephane submitted a fixed fr.po file to add missing translations and
fix broken translations.

[Simos Xenitellis]

Please inform the list as soon as the changes take place so that to
continue translating.

simos

On Tue, 2004-03-09 at 07:39, Jeremy Katz wrote:
>  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.  
>  
>  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).
>  
>  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
>  :-)
>  
>  Jeremy
>  
>  
>  --
>  Fedora-trans-list mailing list
>  Fedora-trans-list(a)redhat.com
>  http://www.redhat.com/mailman/listinfo/fedora-trans-list
>

[Rodolfo M. Raya]

On Mon, 2004-03-08 at 19:29, Christian Rose wrote:

>  Ok, so why someone seems to have reverted all string changes that
> have
>  happened the last months in anaconda beats me. 

Another change in CVS. The fixes I committed yesterday are in limbo
now.  :(

Rodolfo

-- 
Rodolfo M. Raya <rodolfo(a)heartsome.net&gt;
Heartsome Holdings Pte. Ltd.