Hi all,
I’m very worried about the situation of anjuta in squeeze.
In version 2.26 anjuta moved all the function indexing code in a new
“symbol-db” plugin, which is based on libgda (and in turn sqlite). Don’t
be mislead by the “plugin” name, it’s mandatory. This caused all kinds
of packaging problems, but nothing too dramatic. Until now.
Now, after giving version 2.30 to some of our users with heavy codes in
their hands, it appears that this version is still not ready, causing
important stability problems. It’s hard to tell on what scale, but a
crashing IDE is something that would annoy me a *lot*.
Upstream is well aware of that, and their bugzilla is full of memory
corruption bugs, which, along with huge performance issues, caused them
to rewrite the whole symbol-db code in a cleaner fashion for 2.32,
without providing any bugfixes for this code in the stable 2.30 branch.
To make things worse, the new code requires a new libgda, along with a
few other libraries not in unstable.
I have prepared anjuta/anjuta-extras 2.32.0.0 packages in experimental,
with a few small patches to remove the new libraries requirement, so
that it has the possibility to migrate to squeeze eventually. However
the code changes are fairly large, mostly because of this symbol-db
rewrite.
What does this leave us for squeeze?
* Shipping 2.30.2 with its stability problems, with no upstream
support and no hope to really fix them.
* Shipping this patched 2.32 version which should be much more
stable.
* Going back to anjuta 2.24, which used to be stable too, or even
the 2.4 version from lenny.
* Dropping anjuta from the release, and recommending something
else instead (probably monodevelop).
I’m leaving the decision in the RT’s hands.
Cheers,
--
.''`. Josselin Mouette
: :' :
`. `' “If you behave this way because you are blackmailed by someone,
`- […] I will see what I can do for you.” -- Jörg Schilling
Attachment:
signature.asc
Description: This is a digitally signed message part