[Kde-extra-gear] [PATCH] cvsExtract.sh

Jeroen Wijnhout Jeroen.Wijnhout at kdemail.net
Fri Feb 13 11:55:41 CET 2004


On Friday 13 February 2004 11:27, Klas Kalass wrote:
> Hmm, I think you did not understand me. If someone adds some more support
> for CVS Options as a parameter (s)he will most certainly set
> CVS_CHECKOUT_OPTIONS in the command line handling.  Now, if the user
> chooses to use
> --noi18nbasetag  all those values will be lost. It is really easy to avoid
> this future bug by introducing $TAG_OR_BRANCH now and using it in addition
> to CVS_CHECKOUT_OPTIONS, which is more logical too.

Ah, now I understand. Indeed this would result in a much better and 
hacker-proof script.

> The semantics of the --noi18nbasetag and --noadminbasetag is to "not use
> the specified tag or branch" for i18n and admin. You should implemented it
> like this, not make "--i18n-ignore-cvs-options" out of it.
>
> Anyways, code is clearer than words so go ahead and commit. I will change
> it :-)

Ok :-)

> Regards,
>   Klas
>
> P.S: I am not that happy about the two options (admin and i18n). The
> tagging script should rather be completed and used to tag admin and i18n.
> IMHO at least i18n *should* be tagged, because you will want to change
> strings during development and will end up with different strings in branch
> and head eventually.

i18n can still be tagged, i18nbasetag is only about the kde-i18n/subdirs 
files. Which KEG people shouldn't tag. For example with Kile the kile.po 
files are tagged as release_1_6_1 for example, however the kde-i18n/subdirs 
files is not tagged by me. This would cause the script to not find the 
kde-i18n/subdirs files, resulting in an empty list of available languages. 
The same holds for the --noadmintag option. However, I do agree, it is a bit 
of a hack.

best,
Jeroen
-- 
Kile - an Integrated LaTeX Environment for KDE
http://kile.sourceforge.net


More information about the Kde-extra-gear mailing list