- If destination course is using 'simple' gradebook, restore nothing.
- If destination course has letters defined, don't restore backup letters.
- If one category exists in the course (by name) it's reused.
- Items belonging to non-restored modules are skipped.
- If users aren't restored, exceptions are skipped.
Test, test, test... :-)
- It's automatic. No setting for it. Does it sound ok?
- Any info related to non-exported modules is skipped.
- Any info related to non-exported users is skipped.
by the restore process. Such conversion is performed for
every backup file prior to 2005041100.
Wiki internal links are dinamically changed and then,
converted to Markdown.
quiz module backup and restore is now working (although some important
bugs remains in the module, creating strange versions and hidden questions).
Only some minor final things need to be finished to be 1.5 ready! B-)
some important bugs. Now everything seems to work properly with
lessons and backup/restore.
Only 1 pending change: Should the answer->grade field in old backups
go to the new score field??????
"PROG" (progress) because I haven't been able to reproduce the load
of binary data directly to DB reported by some moodlers in forums.
I've created bug 2634 to see if somebody can point me in the correct
direction to reproduce it.
Used before execute the backup of a course to enforce integrity.
Perhaps this function should go to central code and be executed
at upgrade time (and be available in health center too!).
See bug 2380
(http://moodle.org/bugs/bug.php?op=show&bugid=2380)
for scheduled backups and only done in SITE backups. Structure is:
<MESSAGES>
<MESSAGE>
Content of the message with one flag to difference read & unread
</MESSAGE>
<CONTACTS>
<CONTACT>
Content of the contact
</CONTACT>
</CONTACTS>
</MESSAGES>
message_get_participants() should include contact users too!!
Completely untested by me but it LOOKS like it should work. It's quite
standalone, and is basically a single function call during the restore
precheck which looks in the backup to see if it's a BB file, then
converts it using XSLT.
Apart from a little tidyup and renaming, I added a check so that it
should just silently fail if XSLT functions aren't available
Many thanks to Ziba Scott <ziba@linuxbox.com> who wrote it!
