Bug 7263 does not include maxlength field sets into a fresh install.
NORMARC is not included into updatedatabase.pl too.
Created attachment 8711 [details] [review]
marcflavour and updatedatabase changed
Related changes for the NORMARC frameworks are made on bug 7850.
QA comment: why do you use an UPDATE instead of updating the the INSERT field ? Being inconsistent like that will undoubtfully have some side-effect on the long term, because we will forget that there is an UPDATE in the middle of the INSERT.
Switching to failed QA
I think the problem is, that the new column for field length is not part of the inserts yet. So this means updating all lines in all sample files to have the additional value.
(In reply to comment #4)
> I think the problem is, that the new column for field length is not part of
> the inserts yet. So this means updating all lines in all sample files to
> have the additional value.
I agree, but I think I still let this patch "failed QA". The proper way to fix this problem is to:
* upload a framework file
* update the default value
* dump the tables again (with --extended-insert option, or something like that, not sure of the exact mysqldump parameter)
* create a patch from this update
That will fix the problem of maxlength not being set AND the problem of those dump not being correct
(PS: in UNIMARC, I just checked, and they are correct)
Katrin, on the chat, promised to take care of a better version of this patch after 3.8
I agree it would be unfair to block it and release a 3.8 version with ppl needing to tune their MARC setup manually just after installing.
So I'll push this patch and set it back to "assigned" for a better fix for 3.10 !
I plan to work on some other things like missing tags in default frameworks - I think I can address this issue together with those problems then. Thanks Paul!
Note that I had to remove the lines for:
because those patches are not here anymore
Switching back to "ASSIGNED" and waiting for a more proper bugfix
At least the translated frameworks can now be generated from the English templates, no problems have occurred since we pushed the initial fix. I am not sure that this still needs to be fixed in a different way (changing the INSERTS).
Maybe just keep it the way it is?