Guilhem Bichot a écrit :
> Hello Jani,
> The failure of ps_maria which you see during testing of your merge
> (attached output), is ok. In fact, all ps_*.result files (including
> ps_2myisam, ps3_innodb, ps4_heap) were affected by this change done in 5.1:
> Bug#37301 Length and Max_length differ with no obvious reason(2nd version)
> Length value is the length of the field,
> Max_length is the length of the field value.
> So Max_length can not be more than Length.
> The fix: fixed calculation of the Item_empty_string item length
> and you are merging this revision.
If someone is interested, here is how I found out:
- first I wondered: did the merge operation change ps_maria.test?
- answer: "bzr diff ps_maria.test" prints nothing, so no. Thus the
change in result must be caused by a change in real code (not test) as
part of the merge.
- this code change may just as well have affected MyISAM: did the
corresponding MyISAM test (ps_2myisam) experience a similar change in
- answer: "bzr diff ps_2myisam.result" shows a change which is similar
to the difference observed in the failure of ps_maria, so yes. So it's
normal that ps_maria fails and the result file just has to be updated.
- to get more confidence about this, which means, to learn more about
the revision which introduced the change, "bzr gannotate
ps2_myisam.result", click on the changed line, see the revision;
double-click on the line, see the revision's diff: this tells that all
ps_*.result files for all engines were modified the same way
(1365->4096), so it's really clear that there's no problem with ps_maria.