Yahoo Webcam Fixes for Kopete
So i just committed a rather nice patch from new contributor Michael Cole that fixes a few things in Yahoo with regard to webcams and avatars.
You can now accept webcams from Yahoo users and send your webcam to other users. Avatars that are now also downloaded and stored properly.
Michael mentions that the other user doesn't see an notification when you send your webcam to somebody (IOW, they have to request it still) and that avatars only download when they're changed while online.
Either way, it's nice to see some webcam love come our way. Thanks Michael!
First version of QtPersistence released
So, awhile back, I asked about an ORM framework for Qt. I didn't find much else in that category, and nobody else had any thing that was feasible either. I appreciated the suggestions about QDataStream, the XML stuff the Kadu folks are doing, etc, but I really wanted a database - perhaps because I'm crazy like that. 
Anyways, I spent about 6 weeks of spare time, and started the QtPersistence project - which is releasing it's first version today - and because I'm blogging about it, it's now official.
The release notes are here and the tarball is here. I'll probably do another release in about two weeks that does a better job of setting up various bits of project infrastructure but I got something that works passes its test suite, so I'm putting it out there. As always, constructive feedback is welcome.
Update: I left out the most important thing! The source code repository is on Gitorious at http://www.gitorious.org/qtpersistence
kde bugzilla ui changes: feedback wanted
If you've been reading the kde-core-devel mailing list, there's been some talk recently about bugzilla, how much it sucks, whether we should switch, and alternative tools to switch to. However, only one person told me about something that could be improved. So, in order to get feedback on whether or not our customizations that are currently live on bugs.kde.org are hindering vs. helping, I've changed the layout on bugstest.kde.org back to the standard default bugzilla layout.
Please, if you use kde's bugzilla currently, take some time to test out bugstest.kde.org. When you do this, please send feedback. The feedback should be something that we're doing on the live site that is done better by the test site. It can be good or bad!This can be in the form of a bug report in bugs.kde.org against the bugs.kde.org product, or sending me a simple email. Please include changes that you want made, or specific issues with the site so that I can at least look at fixing them.
I get that a lot of KDE folks are not happy with our current Bugzilla. I'm glad that you want to have good tools to work with. So far, I haven't really seen anything that's better. I've only seen systems that do things differently, which means that we can emulate that with Bugzilla, most likely. That's after using Jira, Mantis, Flyspray, Bugzilla, and Redmine, and an old version of RT. Why not work on improving our current tool? If you think that Bugzilla has fundamental problems that require switching, I want to know what those are too.
i can haz object persistence with Qt?
ok, so before i spend time going and creating something totally awesome to do object persistence with Qt, I have a question.
Have people already done something like this?
What I Know:
I know about Kandau already. I looked at it, and it's Qt/KDE 3 only, and I don't really want to spend the time to port it to Qt/KDE 4, mostly because I didn't like the way you had to use it.
I'm not too interested, at the moment, in using the Ruby bindings to Qt (which would then give me access to DataMapper and/or ActiveRecord), but depending on what's out there, I may go that route anyways.
Akonadi:
Before somebody suggests Akonadi (which is awesome), it doesn't really fit my use case at the moment. I'm looking for something pretty simple, and I'm not concerned with exporting data to other apps right now.
devenv 1.0 released
Just made the first release of a small project of mine that I wrote to handle my various projects and development environments. It's called devenv and the homepage is at Gitorious. 1.0 is the "it works good enough for me so here it is" release. It's designed only for zsh, so if you try to run it on not zsh, it will crash and burn.
Would love some feedback, patches, etc. Future ideas include commands/scripts for updating source trees for a project, doing a better job of handling branches, and anything that I notice that I start to do a lot that will keep me from typing, cause I'm lazy.
Patch Friday – 10/09/2009
I got patch friday done this week. I committed some header cleanups and a user-submitted patch to fix unsetting your Jabber mood. The jabber mood fix will be in KDE 4.3.3. The last two patch fridays before this one have been a wash. The first was because I was traveling over the weekend, and the next because my daugther was sick (but is better now).
I haven't checked reviewboard or bugs.kde.org in awhile for patches that are ready, so if you have one that you think is ready (especially if you're the patch author) then feel free to email me a link to the bug or reviewboard post. 
Patch Friday – 09/18/2009
boot camp workout class this morning. more heavy weights today, although in a bit different style. ended up doing step ups with 65 lbs of weight in each hand. could barely lift the weights up enough to make the step.
Got Fedora 11 reinstalled on the laptop while at work. Always nice when you can install 1.2GB of updates during a meeting.
2 patches committed today for Patch Friday. One adds a warning on 'clear' to the chat window, and the other fixes up a ton of bugs in the Gadu-Gadu contact list support. Thanks to Dominik Geyer and Jakub Grandys for the patches!
Patch Friday!
I've deemed Fridays to be "Patch Fridays". This means that I will be doing various things with patches. I review patches, test patches, commit patches, etc. So, if you've sent me a patch, or stuck one on review board, or I've seen one from you on bugs.kde.org, then if I haven't done anything with it before a Friday, it'll most likely be handled on a Friday.
For this Patch Friday, I've (so far) got four patches to commit: two gadu gadu patches and two chatwindow patches.
Why Friday?
Because I had a lot of patches that I knew I needed to take care of, and today just happened to be the first day I could get to them. So I've decided that I'm going to stick with it.
Can I send you just any patch?
Well, not really. I'd prefer it if they were areas that I spent a lot of time on. For now, that would be:
- Kopete
- bugs.kde.org
- Basket
- Anything for the kdesdk module
- KatePart
- KDevelop
The last two areas are only areas that I'm marginally involved in, and so I'll be passing patches for those two pieces of code through the proper channels before committing them. Basket patches will go through my git repository of basket on GitHub before going into Basket's master repository that Kelvie Wong, the Basket maintainer, has.
Every Friday?
Yeah, pretty much, unless I'm travelling or otherwise indisposed.
Got Patches?
Friday's not over yet! Send me your patches!
Kopete and Yahoo!
I've just committed the implementation of the new Yahoo! protocol authentication to Subversion and wanted to provide more information on what the problem was and how to get the fix.
Hey, why can't I login in anymore?
The problem that affected Kopete is with the authentication mechanism. Yahoo Messenger 6 used a heinously complicated password obfuscation method to "encrypt" the password as it was being sent over the wire to Yahoo's servers. Way back when, Cerulean Studios, the creators of the Trillian client, were kind enough to implement this authentication mechanism. Pidgin got it, as well as the libyahoo2 library we were using at the time. When we switched to our own Yahoo! library some time later, we carried it over. As part of this change, we began to identify to the Yahoo servers as Yahoo Messenger 6. Everything was working, and everybody was happy.
The real problem came relatively recently. As time went by, we started updating which version of Yahoo Messenger we presented ourselves as. When those changes were made, the authentication code was never updated. So, even though we said, hey we're version 15, we were using the older scheme. I even found out in February that Yahoo! was going to be changing their authentication mechanism. I didn't do anything because I was too busy with other stuff, and this fell by the wayside.
Yahoo began upgrading their servers at some point recently to phase out the older clients. It became a problem for us because when they upgraded, they started requiring protocol version 15 clients to speak the version 15 authentication scheme, which we never implemented. Since we still spoke version 13's authentication, this cut us off entirely.
So where do I get the fix?
A nice person pointed the Pidgin folks at some documentation, and when this problem reared its ugly head a few days ago, they got to work on a fix. I pulled the fix from their code and now we authenticate the same way Pidgin does. That code is now in KDE's subversion repositories in trunk, the 4.3 branch, and the 4.2 branch. I've sent a message to the packager mailing list with the revision number from the 4.2 branch that needs to be integrated. I've also asked for a retagging of KDE 4.3 RC1 so that the fix can be included there as well.
So, you'll need to wait for your distribution to provide package upgrades or compile from Subversion yourself.
What the fix does
It does three things:
- Implements the new authentication mechanism - which is much simpler
- Sets the default server to scsa.msg.yahoo.com - this is what Pidgin is using now as well
- Converts any configuration that was using a yahoo.com server to login to point at scsa.msg.yahoo.com instead, so that you have the highest possibility of logging in
Other Yahoo! login problems
If after you've upgraded, you still can't login, please file a new bug so that it can be investigated and handled properly. Adding comments to the already existing bug are not as likely to net a solution that will actually work.
Credits
Thanks to the Pidgin folks for the code and the new release which is how I actually noticed the problem in the first place. I didn't see the original bug report come in.
Thanks in particular to John Bailey, whose post on the subject from Pidgin's point of view provided the inspiration and some small bits of content for this blog post and its format.
new on bugs.kde.org – aliases
I turned on aliases awhile ago on bugs.kde.org. Aliases are a way for people to reference a bug number based on some nice text that is given to the bug rather than by the bug number. You don't have to give every bug an alias, but it seemed like a useful feature to have. If nobody likes it, then we can turn it off, but hopefully people are already using it and i'm just not aware of it.