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
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
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
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.
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.
-
Christian Rose -
Jeremy Katz -
Sarah Wang