One week after beta 1

asoliverez's picture

It's been 10 days since we released KMyMoney 3.95, and we are still alive, which is not little. The main problems we have had were with KDE SC 4.4 beta. For GPG, we used libkleo, but that API was not stable yet, and it changed in 4.4. So, Thomas changed to use gpgme++ directly instead, skipping libkleo altogether, and now SVN works fine in 4.4. Besides, there were quirks here and there, but nothing serious.

One additional problem is the integration with AqBanking. Most distros still ship the Qt3 version, and KMyMoney fails to compile in those cases. The problem is, there is no actual Qt4 version, rather a script which converts the code to Qt4 on-the-fly. If you are compiling from SVN, that might work, but if you rely on a package, that's a problem. Myself, I haven't been able to build AqBanking for Qt4 under Arch, although I admit I haven't devoted much time to it. Coding, and emails, and a baby, you know, not much spare time to spend on it, and other developers built it and tested it.

Another issue at the moment is our move to the KDE infrastructure. So, we have the KDE mailing list, plus the Sourceforge mailing lists. We had been succesful so far in the separation of the two, but this release triggered a lot of cross-posting, which is not healthy in the long term. Plus, we have to make a better effort to explain the difference in requirements, and pass that information on to the packagers. If you are packaging KMyMoney, please get in touch with us and we will provide as much help as we can. We have already accepted and commited patches in a matter of hours.

Now, after all these hiccups, we are concentrating on fixing the bugs we can, and on porting some of our widgets to Qt4. Most of the custom widgets are still using Qt3Support, but we are slowly moving to the MVC framework. It will get bumpy in SVN for a few weeks, but we expect to sort out most of the problems before the next release, which is going to be sooner than expected, at least by our standards.

UPDATE: explained better what happened to GPG

Comments

GPG encryption

Just to make it clear: I did not revert the change but rather based the implementation on gpgme++ rather than libkleo which in fact uses gpgme++ underneath. This should give us some stability in terms of the API.