Moodle Community Sites

VIP ( = Very Important People) watchers added for further comments about possible strategies.

Oh, VIP watchers, what a responsibility!

This poor developer will try to do his work successfully!

:-D Ciao

Well, I've executed the comparison script with today's HEAD against 3 months ago HEAD and I've got these strings deleted since then:

< blog.php:addotags
< blog.php:addptags
< blog.php:otags
< blog.php:ptags
< countries.php:CS
< countries.php:FX
< countries.php:KO
< countries.php:TP
< countries.php:WA
< countries.php:ZR
< grades.php:aggregationcoefextrahelp
< grades.php:aggregationcoefweighthelp
< grades.php:configbulkcheckboxes
< grades.php:configquickfeedback
< grades.php:errornocategorizedid
< grades.php:overallavg
< moodle.php:nostatstodisplay
< userkey.php:keyiprestrictionhelp
< userkey.php:keyvaliduntilhelp

About these strings... we should test each one to see if they are used in old stable branches:

1) If they are used in stable branches... (say 3 latest branches 1.7, 1.8 and 1.9?) then we must reintroduce them ASAP.
2) If they aren't used, they they only have been in HEAD so, it's safe to keep them deleted, IMO.

Then, two tasks:

A) I think that this task of searching for uses... could be easily automated, to provide the detection script with that useful info.

B The other task would be to detect string files that have changed "a lot" to see if the change is good enough or no. This could be automated too if somebody knows a good algorithm to measure diferences between strings (sure it exists).

In any case... perhaps it would be a goo idea to:

• Move this script to, say, cvs.moodle.org (we need a local checkout of all the branches + head) in the host running the detection script.
• Create some list in lists.moodle.org allowing people to get subscribed to it (Could be 2-3 lists, one with daily changes and another with monthly changes).
• Make the script to send results to that list.

How does it sound?

Ciao

Eloy,

My (modest) +1 for your points 1) and 2) and your suggestions A and B.

For B, Levenshtein algorithm (see http://en.wikipedia.org/wiki/Levenshtein_distance) could be a candidate.

Ciao

Sorry, I can't edit my previous comment, where Jira added a paren at the end of the URL Here is the correct one (sorry for the spam).

http://en.wikipedia.org/wiki/Levenshtein_distance

Fix deleted lang strings added as a new issue - MDL-14866.

Have we this under control with current "Deleted strings" mailout? Must I do anything else here?

Eloy, thanks for the deleted strings mailout - it's great

I don't think there's anything else to do, so I'm going to resolve this issue.

We don't have it at all under control. It happens so often that it is impossible to keep up. But the translation interface now keeps all strings that are not in en_utf8 en saves them with an //orphaned comment.

The only problem I see is that it is e.g. not possible to make a translation for an older version with a 2.0 since not all strings are there, but I think, seen the long process a translation is, that it is not a problem.

I still believe in this context that it would be great to delete all non necessary strings for 2.0 out of the en language pack. David was doing something for that if I recall well..

Yup, he got something in MDL-16769. I guess we need to decide about branched langs soon... ciao