Hello,
I don't think there is any preferred method. You named the advantages of each of the variants accurately, now we only need to decide how much caution we should pay to vecwiki :-). I'm happy to do this with you next week in an extra window, if you have time.
Alternatively, we can add the namespaces to $wgExtraNamespaces, which can be (ab)used to rename core's namespaces (used in zh* wikis), run scripts, let it ride train, and make sure it works properly, but that's probably more time consuming than simply backporting.
Martin
Hello,
There's a request at <https://phab.wiki/255938> to rename several
namespace names for core's MessagesVec.php. See patch at
<https://gerrit.wikimedia.org/r/613110>.
Would it be okay to let those translations go with the usual MediaWiki
Train process, and run the required scripts afterwards
(namespaceDupes.php) or a specific backport window for the change be
preferred?
We've used both approaches in the past (cfr. <https://w.wiki/XE4> for
specific windows) but there are no docs as to what is the preferred
method.
In favour of letting the train take care of things is that the process
is mostly unnatended and we can take care of running the scripts
afterwards. Contras is that I'm not sure vecwiki would be error-free
if we don't run namespaceDupes.php afterwards.
In favour of using a specific backport window is that we make sure
everything gets done, but the contras are that I need to get a
specific window from RelEng, get a deployer that is willing to do that
specific window, two merges (master and wmf/*) and sit while the full
scap and l10nupdate scripts run.
What do you think?
Thanks for your help.
Best regards, M.
--
To unsubscribe from this group and stop receiving emails from it, send an email to sitereq-l+unsubscribe@wikimedia.org.