0.2 Ubiquity in Thunderbird is up.

Link to 0.2 release:

http://zenit.senecac.on.ca/wiki/imgs/Ubiquity-0.1.2pre.zip

Link to project page:

http://zenit.senecac.on.ca/wiki/index.php/Make_Ubiquity_Work_In_Thunderbird


Well... I sort of had big plans for 0.2. Thankfully, one of those plans came through. I was really hoping that I could get what I was planning to do for my 0.3 into my 0.2 instead, but I’m having too many problems getting it to work properly.

Oh well no worries. What my 0.2 essentially does, is it allows Ubiquity to pop up in both Firefox and Thunderbird using the same XPI file. That means, you can install Ubiquity in either application using the same extension. There’s no need to create two separate extensions and maintain them both. In my eyes, that’s a huge hassle, and I can see why everyone wants it done this way. To be quite honest, getting it to run in two separate applications wasn’t too difficult, especially after I David Humphrey gave me some tips on a direction to take. What IS hard is finding a way to get EVERYTHING to work properly without bloating the code.

** Rant about my experience on the project. Skip this if you don’t want to read! **

What do parts of the code do I bloat, and what don’t I? Sadly, I think that if I want to get Ubiquity working in both Firefox and Thunderbird, one way or another I’m going to have to add lots of code. I don’t think there’s really a way around it. I’ve been making an attempt to use Javascript’s inheritance feature to help Ubiquity manage itself depending on its environment. But you know... sometimes I just look at things I’m doing and say to myself, “It would be a lot easier if I just used an if / else if here to do things, but it’s easier here to inherit things.” Of course, if I go ahead and slap ifs everywhere, I have to be careful. What if this thing gets ported to another program? That means someone will have to go through all the code and manage any ifs I put in. That’s not cool, especially considering that there’s almost NO COMMENTING throughout the Ubiquity code.

I can’t stand uncommented code. We’ve been taught that it’s a standard to comment your code and keep your comments up to date every time you change something. Sadly, in an open source environment, you probably have a hard time forcing people to comment things. Another problem I’m having with separating the Ubiquity code is that I have to learn everything that Ubiquity does. If I screw something up it breaks the program. Fortunately, I’m not as lost as I was when I first started. I’m to the point now where I can figure out what does what without asking a ton of questions. In fact, sometimes I find myself editing my old code and saying to myself, “Why did I do this? It’s stupid. It works but it’s still dumb.”

I think at this point my biggest problem is fighting against the code that we created. As I mentioned above, this includes my own code. We cut out so much of the code just to get Ubiquity to fire up in Thunderbird, it’s very hard to figure out what was removed. In a lot of cases what I would consider important code was just ripped out of the program. I’m guilty of this too. I did a lot of code ripping. “If Ubiquity works without this code, and this code is making it break, get rid of it!” That was my philosophy while getting this to work in TB. I should have just commented things out.

I have a pretty good example of why I wish we had just commented out (with comments saying WHY we were removing something). Last week one of our team members, Justin, couldn’t get previewBlock to work properly for the map command. It kept returning null and he couldn’t figure out what was wrong. I checked out the Thunderbird code, and it looked like previewBlock WAS getting set properly. I then decided to check out the Firefox code and compare the two. Well, the line that previewBlock used to set itself to the browser’s current window was ripped out. That’s why it was null, and why it wouldn’t work.

Sadly, despite my rant on commenting, I still haven’t added a single comment. I really need to get around to doing that, but I’ve been caught up in the “rush to release” momentum.

** END RANT. **

In a few cases, I haven’t been able to get some things to stop throwing errors up in Firefox. It’s not a big deal though, they still run with the errors. I’m just going to leave them for now and work on “organizing” things. I’d really like to eliminate the last few TB_ files I added. Just completely merge everything.

Ok so what can you expect from this build? It runs in both Firefox and Thunderbird. Is it fully functional in Firefox? Close. Is it fully functional in Thunderbird? Nah, not really. It’s getting a lot better though. I’m not sure how complete our new annotation service is for Thunderbird. I do know one thing though, it blows up in Firefox. Not sure what I’m going to do about that. Currently I’m making Firefox use its built in annotation service (what Ubiquity did originally). So that works fine as long as I keep that functionality separate.

At this point someone has made Ubiquity force Thunderbird to choose a web browser whenever you want to use a command that requires a tab. I’m not too keen on this idea but I’m not touching it unless everyone wants it changed. So it’s going to stay this way for a while unless everyone decides they want to change it. The biggest problem I have with this method is that if you don’t have Ubiquity installed in Firefox, any commands that require you to choose to run Firefox will blow up. Firefox won’t be able to find the internal URL and will say it is invalid. So if you want your command-editor from Thunderbird to pop up in Firefox using this method, make sure Firefox has Ubiquity installed.

As far as what’s inside the code in this release, I’m not releasing any code with my attempts at using JS inheritance. There’s just way too many problems, and removing things that break just to get a release out is kind of pointless (and a monumental waste of time), considering the inheritance won’t actually DO anything if I go that route. So I’m releasing our “cleanest” build, straight off of our repository.

If I manage to get my structure worked out without bloating our code too much within the next week or two, I might just release a 0.2.1 for anyone that’s interested.