Other briefs: I want to give my own shout out to Chris and Dave for such a fantastic semester it has been. It has been real inspiring to learn to open up to the possibilities and concepts that come with Open Source! You guys do a great job pushing the projects, sometimes which scares me (a good thing) and helped me to see the amazing vastness that Mozilla technology is. To me it feels like a new religion in software engineering. This is an experience I won’t forget and won’t let down. As Peter Evanoff mentioned, although we’re done school we’ll eventually bring our project to a version 1! I’m still gonna jump and hang in IRC randomly.. I think I certainly would have time telling students of the next class OSD600 how to build Firefox! - At least I think I contributed in that by writing “Minefield for Dummies”. Anyway it’s time to look for work and I’ve been holding up some other personal learning projects all year - I’ll blog about those things in the future here if I don’t manage to bore you to death.

Anyway though, after the projects were laid out for us, I really wanted to help develop something that everyone can use. That’s my biggest drive in programming also, to make someone that everyone can enjoy. If I didn’t go into programming I might have become an artist! -or something in between like computer graphics design. (Well, there’s still time )

Economically it was a good choice to make also since I wanted to take advantage of buying a USB Key. It was suggested that we get one in order to hold our Mozilla build. As it turns out though I never really did use it for that since the school is so generous with lending out laptops.

I suppose it was these two things that lead me to take on USB Firefox!

It was a little scary, I mean everyone was a bit scared for that matter. This thing is huge and we’re trying to add a brick to this quite solid building. Or another metaphor: It’s like we’re putting a new Janga piece into the tower. Chris told us that the Mozilla community didn’t know how big this project is. Some said it is small, while some others said its huge. On choosing day, I was still trying to measure the success / failure percentage of the project. You never really know until you start hammering it out.

Today though, I consider myself lucky. As I hoped, Peter E and I had gotten a good head start with 0.1. We didn’t take it lightly like we might have for other subject’s first assignments. It was this course that even got me started on using Firefox regularly.. and it has been quite the eye opener.

Next entry, Universal Firefox 0.3 Release!

I started with one random guy, who I can’t remember now, who seemed to dealt with Firefox profiles. I wasn’t getting much information on this projects though, other than forwards for a bug about free t-shirts to extension developers.

So then I switched him out for watching Gijs. Now the bugzilla messages came in a rush haha. I went with Gijs just seeing how he was a nice guy on IRC and helped with building Minefield for the first time.

I can’t so much for the 300 messages I still have saved up in my email folder, that is there’s too much to handle and I’m not going to go into it. But I was surprised at the variety of topics Gijs was involved in. I really had no clue, thinking that he was one of the guys who made Chatzilla. Then he was sometimes mentioned in class throughout the semester. But anyway, this guy was in everything. That’s the interesting part of all this Mozilla stuff is that if you know enough, your not stuck on just Firefox.. they all run on the same backbone technology. So if I caught up with my javascript and C++, maybe I could easily be taking part in multiple applications too. Interesting..

It’s interesting how much we play around that we somehow encounter the problems ourselves before naively releasing a package of what we’ve done so far to have others discover it for us. I guess we’re still good at being end-users! We blew away Firefox Linux from the test machine we have and fired up our USB Firefox. Sure enough that was when we discovered a problem with missing the library libstdc++.so.5 . For a few days we tried making static builds of Firefox to see if that cover up the library issue, but failed. The problem was similar to our dll hell on windows, but instead this is library — something. A machine having a later version of the library did not cover the compatibility for an older one. So I guess it doesn’t work like COM objects in Windows :(. The solution was too easy in the end, I for one didn’t want to believe in it, so I was hesitant. However in the end the best way was to just package the USB Firefox with its own copy of the library to use.

In the process this raised many questions of confusion. It’s still the early stage for our project, in terms of reaching a 1.0 it had only kinda made sense to use the latest official release of Firefox 2 on our USB stick. Later on we’ll move on to Firefox 3 when its safe to assume no random bugs will come in the way of packaging this up. We might actually give it a shot in our next release. What was weird is why isn’t this Firefox using the latest library? If it had be following up with development on CentOS then using version 6 would have been the minimum standard no? We’ll find out more about that soon.

Peter Evanoff has gotten me thinking not only user friendly, but building Grandma-friendly software too. It’s an amusing way to put it, but really Linux needs some work to better appeal to the mass market. The best tester would be Grandma.

Now I imagine the goal of our 0.3 will be to make sure that this is working on Mac too. Hopefully that will be a breeze since its similar to Linux. Will Chris let us blow away the Firefox on one of those brand new shiny Macs? Only time will tell.. (I always remember that line from an episode Ninja Turtles..) . I’m thinking that we seek out Danny Roy or someone to test our USB Firefox on Windows95/98. On the Linux side we already tested on Fedora, Ubuntu and CentOS - so that seems to be well covered. On Mac, Chris Tyler mentioned about testing on a Mac with Intel then PowerPC. Extensions I’ll say are covered, we included a super generic one in the package to provide that proof. If other extensions fail, it’s not our fault they did not make it cross-platform :). Annnnnd finally, the ultimate ideal goal for 0.3 would be to have a single entry point that will automatically fiqure out which OS the stick is on and run the correct Firefox Platform, saving Grandma from any brain work.

Those interested in giving our 0.2 a test run can get it here! Check out our wiki page on zenit as well. We would love to hear some feedback.

Edit: Now I’m very very sure that changing that one name tag to match with the others fixed the problem. It was the only change we made after Chris had a look at it. But now I made another version of the extension with a different name, and it seemed to work! So what was the problem before then? It disappeared? My guess is maybe that “!” cannot go in the name.

I’ve been wondering for a few days now how they got Firefox to recognize that you clicked on an extension to download. I thought the smartest thing they could have done was just have Firefox look for the .xpi at the end of a file. Okay, maybe that’s the lazy way, as extensions can turn out to be for something else. So on the same tutorial page it links to another page about what code to use in an html. I slapped the code into a custom html page, got a Firefox icon for the sample extension as a default, got the Hash value of our xpi (tee-hee) and uploaded it here. Cool ya?

So I clicked the install link on my custom html page now and our sample extension installed as expected. Wait a minute, I was on our USB Firefox on Linux! The extension as I see it is the most minimal needed to get one started and that was made on Windows. Doesn’t matter, it works here and everywhere! The next thing Peter and I checked out was the mozilla folder on this Fedora machine located at /root/.mozilla/extensions/. It had one empty folder in there labelled with a GUID. It’s the same GUID as written in our sample extension’s install.rdf! Cool.. are we gonna do anything about it? It’s not significant at this point in our project, but we learned something.

Then there’s also a few things on the side I think we should look into. Like what files (or even registry keys if any) are being left behind on the host computer. If they are significant then we should document that somewhere for a later polished version. We still have to test how extensions are doing on our USB Key, do they still work?

Finally, I’m sure we’re gonna learn more random stuff as we go. I haven’t really heard from anyone else in class talk about learning useless stuff that has nothing to do with their project. I’m sure it happens often though.. we”re always making that one step forward and then stepping back again. Always there’s that fear of having to step back so far that you end up at the start again. Err.. I think I got off track, so I crossed that all out

It was inspiring to see him dish out the knowledge like he did. We learned about some of the past developments that have ended up in a forked project, got a greater perspective of how other companies especially Google have come into play with Mozilla, generally what is going on with Firefox as it approaches version 3, and what does that mean to us developers for getting code in. It was a very interesting discussion. Thanks.

Our 0.1 is not big, but we got something, and on the side a whole lot of other random knowledge. It’s proof of concept Peter E keeps reminding me. I guess I’ve been to used to having something solid and complete as proof of completion. It almost feels like cheating, but it works! In the process we learned a whole lot about Linux, got some idea of what some of these Firefox components are about, got more efficient with building the source, saw the potential benefits and the strange politics behind bugzilla, (re-)discovered the Mac (since we never or haven’t touched one ages), learned a bit about batch scripting and other random stuff. We had fun. We’re playing around with stuff and learning at the same time.

So far we got a Windows version that reads the common profile on our stick. Actually we just recently got Mac and Linux to read that profile too! I called it the Universal Profile - very original However they’re not yet as nifty as the Windows batch file we made yet! Our Window version covers loading the profile and the best part is that it knows where to find itself on the USB Key so that the user can run this Firefox from any directory.

Our next version, we have yet define exactly. Most likely we want to get the Linux version to load the Universal profile in a more elegant way as the Windows batch file has. Then hopefully the Mac version should not be all that different. Its solution might even be exactly the same, we don’t know yet. You can make assumptions, but no one really knows. We will though. If we got that working perfectly, extensions working by the end of the semester, then we could say we almost have a version 1!