I am pleased to announce that KPhotoAlbum 4.4 is available for download.
The main improvements on this version are support for libkipi 2.0.0 and greatly improved save time for the image database. Now it is also possible to remove duplicate images while merging the tags to the remaining image. And naturally a couple of bugs got fixed as well.
This week we discussed saving times on the KPhotoAlbum IRC channel, which made me give it a shot at rewriting the saving code from QDom to QXmlStreamWritter. I had a hunch that it would give quite some nice improvement in saving speed.
Now, if you have ever participated in one of my Qt training classes, you will likely have heard me say something like NEVER EVER trust your own hunches on where time is spend in your applications, USE A PROFILER!. Rewriting the code resulted in a disappointing 10% time reduction on saving, not the 50% or more I had hoped for. Sigh! Why do I never listen to myself?!
OK, so when I had callgrind and kcachgrind in action anyway, I could just as well look at what made it relatively slow at saving, and I found, not surprisingly, that it spend quite a bit of time in places I wouldn't have guessed in a million years. I got an impression 18% reduction simply by caching whether the XML file should be in compressed or readable format.
In total, depending on format used, and whether your KPhotoAlbum is build in debug or release mode, I got a reduction of time spend when saving between 62% and 73%, which means a speed up of 3-4.
Saving now takes less than 400ms on my database with 13800 images (of which more than 95% are tagged). I guess it will be a while before I need to revisit the speed of saving :-)
I am pleased to announce that KPhotoAlbum 4.3 available for download.
New features were described when announcing the beta version. Only minor bug fixing and addition of pattern based search was done since then.
New features concentrate on video support improvements and background tasks. Also the database backend has been dropped completely leaving the XML backend as the only option. Automatic stacking of images based on file version detection has been brought to the maintenance menu giving control to the user instead of only performing the action upon starting of KPhotoAlbum.
Jesper has posted a video tutorial on YouTube demonstrating the new features that are available on the current development version of KPhotoAlbum.
There are some new features to test out on our git repository (master branch). They are mostly related to how videos are handled but also background job queue is implemented and some refactoring done, including removal of SQL backend. Any feedback is appreciated before pushing out a new release.
Jesper was on a vacation and by tradition this means some news for KPhotoAlbum as well. This time he was hacking together some improvements on video support - and a lot of things under the hood. On visible side it is now possible to select the frame to be used as video thumbnail and while hovering over a thumbnail a few frames from the video are cycled over to give better idea what the video is about. Less visible improvements include e.g. new background tasks manager. It sure is great to see Jesper busy on improving KPhotoAlbum once in a while.
We are happy to announce a new release of KPhotoAlbum. The major changes including stability and speed improvements were described when announcing the release candidate 2. After that only a few minor changes were made.
You can get the KPhotoAlbum 4.2 from KDE mirrors.
Release candidate 2 of our next release is available on KDE FTP site.
Changes since the release of 4.1.1 include the following:
We are working on a new release. New features are frozen and the current GIT master should have all the features to be expected on the release. Some more bug fixes should be done before the release. We are planning on providing a release candidate soonish before the final release, but there is no date set for that.
There has also been a major change in the forces behind KPhotoAlbum. Jesper has decided to step back and has left a huge gap for other people to fill. We want to thank Jesper for his great work in producing the best photo cataloguing software known to exist!