Trolltech seems to be hard working to smooth out every imaginable point of criticism and to be the best Free Software team player:

Some progress has been made on the Qt 4 front this week. I've implemented a simple QHttp based version of XMLHttpRequest that will ultimately be bound to KJSEmbed/Qt4. This will mean that the scripts will be able to access web based XML services such as news feeds, weather reports etc. I've now also got a working build environment to begin helping with the kdelibs porting effort.

As already pointed out by others, we started porting kdelibs to Qt4. Thanks to the restless work of Maksim, DCOP now compiles and works, at least the client part of it. He managed to restore compatibility with the protocol used in KDE3, so we now have some means of posting stuff between toolkit versions.

I just had a problem with Knowledge, which as you might know, is my attempt towards a Wikipdia offline reader: I needed to disable lots of GUI stuff in my QMainWindow-derived class as long as no offline image is loaded. Some of the widgets werent even available as object members. All widgets need to be deactivated and then activated again after the book was loaded. The obvious choice would have been to add a bunch of members, one for every object in question, and then call setEnabled( true ) or setEnabled( false ) for each of them respectively. Result: lotsa code.

Qt4 is really progressing well. The only problem at this point is that it still changes a lot even after Beta 2 which originally meant to end the phase of rather radical changes, according to the trolls. But that's fine. Let them get their APIs sorted out, let them react to the extensive feedback and let them time to do it right.

Man it seems like months since I have taken the time to blog, but this quarter just flew by. Since Zack joined the party at Trolltech we have been on the hunt for more developers. Now Zack in Norway is cool, he's doing cool stuff and in the end its all going to help me make better products for my clients.

It has been a while since I blogged - a bit too much real-world work, plus a concerted effort on the Qt Cryptographic Architecture (QCA) are to blame.

Right now, I have QCA2 running, I have two plugin providers working - one based on OpenSSL, and the other based on GNU libgcrypt, and I have 595 passing unit tests. I'm particularly proud of the unit tests, because they give me a lot of confidence in the plugins. In terms of functionality, we have hashing (MD2, MD4, MD5, SHA0, SHA1, SHA256, SHA384, SHA512 and RIPEMD160), keyed hashes (HMAC-SHA1, HMAC-MD5 and HMAC-RIPEMD160) and some ciphers (AES128, AES192 and AES256 in ECB, CBC and CFB modes). Supporting that we have a basic random number provider, a secure byte array, and some arbitrary precision integer classes. It isn't really ready for prime time, but its emerging nicely.

I just committed the Qt Mozilla code. To start testing it you need to: checkout mozilla, configure with --enable-default-toolkit=qt, make and that's it. Please do not, I repeat do not send me wishes or bug reports just yet. I'll simply ignore it unless there's a patch attached to it. There are two basic issues which have to be fixed before I'll be taking any requests (1) the toolbar is waaay to tall, 2) updates aren't sometimes propagated correctly).

The Qt Crytographic Architecture (QCA) was originally written by Justin Karneges as part of the PSI instant messanger. After a little bit of discussion, and an agreement on how it will work out, he's agreed to move it into the kdesupport module (KDE already has two copies - one in kdenetwork, and another in kdenonbeta).