Hi all,
See my comments inline under Julien's.
Le 25/03/2022 à 09:18, Julien Nabet a écrit :
Firebird is not the only pb, charts aren't displayed anymore in
reports and the whole reports part is dependent on old Java external
components.
Yes, the issue of charts in reports no longer being displayed is now a
very old regression bug.
There are also address books pbs:
- Mac one (eg : leaks but not only this, Alex may tell more about
this I suppose)
At least now, a connection to macAB is possible without crashing in the
TDF-provided version - haven't tried Collabora's versions yet.
There are a couple of other Mac-specific issues (tdf#50626, tdf#64641)
with the macOS addressbook.
- Thunderbird one can't be used anymore after Mork->Sqlite migration.
Yes, that's an issue for everyone - note that system support for Sqlite3
on macOS is included by default in the system, but that doesn't help
much if it can't be made to talk to LO over the SDB bridge.
This ties in to a lack of built-in support for SQLite in LO in general.
It is likely that any such integration would be perceived by users as a
very welcome addition (and not just for Moz addressbook support), but my
understanding is that this would not be trivial to implement, as it
would be basically be like redoing all of the work for Firebird over
again, but this time for SQLite. Given that we are already in a mess
with Firebird, having a 2nd mess with incomplete SQLite support might
not be the best allocation of resources.
Much as I hate to say it, _*if*_ resources were to be allocated to Base
development, I would much prefer :
- fixing old regressions, e.g. the chart bug in the report builder;
- making embedded Firebird the functional equivalent of embedded hsqldb
- currently, it is like some awkward reject, shivering in the cold and
dark - lots of incremental improvements to be made here;
- migrating the Java report generator code to C++ - there used to be a
native report writer, and it got killed off in favour of Java - however,
this would not be a small endeavour.
Of course, if the general thinking in the "dev community" is that
database front end support is a dead duck, then it seems unlikely that
even TDF would engage resources in it.
Personally, I would find that incredibly sad, and it would take away my
raison d'être for using LO, but I'm not going to be King Canute either.
Alex
