Well, for the final release of the semester I’m planning on having tinderbox work dynamically off of the patch that I’ve fixed up since my 0.2 release.  This will involve me reworking the code that I’ve written for the 0.2 release and shoving some more javascript into the tinderbox index page from the 0.1.


After a good 5 hours online with Rob Helmer I finally figured it out. I got Tinderbox working locally on my Ubuntu computer. The main pitfalls were finding which packages were needed to get the install actually working correctly. Afterwards, there were a few files that I needed to manually copy over by hand (thanks to broken makefiles) and it was golden.

I’ll post up a complete guide when I get the time to do so, otherwise anyone is welcome to come to me with questions. Note that I only got Bonsai (which is another required program) working just enough to get by.. So I won’t be of much help in that area.

Well, my tinderbox presentation on Friday went nice and smoothly. The class seemed to understand the project well and didn’t have too many questions. Despite my lack of contribution ideas I was presented with some good ones by Chris Tyler.

The main stumbling point for my project right now is getting a running instance of tinderbox on my home machine. This will allow me to test my new code that I’m trying to contribute to tinderbox that will make my front page possible. So far it’s been quite a challenge. The contribution idea was if anyone wanted to try their hand at setting up a Bonsai server or Tinderbox server then that would be greatly appreciated.

Well, unlike my previous release there are no flashy screen shots this time I’m afraid. The release for me this time was hacking away at the tinderbox Perl code *shudder*.

<tangent>I grew up on PHP and Perl is radically different to me. Never really taking the proper time to learn Perl seems to be a mistake at this point. Although, this past release I’ve revisited Perl with a trusty reference book by my side and realized its power yet again… Which would be files</tangent>

The idea with this release was to get a simple JSON listing for all tinderbox trees. In layman’s terms: Figure out how to get Perl to output something called JSON (A markup style for JavaScript) for all the build projects that Mozilla has. Well, lucky for me a JSON output method existed already. But it was full of overly detailed information that was not needed, and this was where I came in.

During my last release (which was turned down) there was concern expressed in the amount of server requests and processing time required to generate the relatively complex index page. Therefore, by trimming the “fat” out I’m hoping to speed things up.

Note that installing and running your own tinderbox is a monumental challenge (As is stated in the INSTALL file) and I never actually got it working.. at least yet.. (Needs bonsai as well, which is another challenge in itself) Therefore testing was not preformed.

The hardest part was probably just learning what all the code really did as I knew what I needed was there, somewhere. This also saved me from re-inventing the wheel and getting slapped on the wrist in a code review…

I’m interested to see how this one turns out. Let the fun begin 🙂

Patch can be seen here

Well, as the 0.2 deadline rolls around I’ll start by laying out the groundwork for what needs to be done to Tinderbox. The main thing now is instead of making it look nice I’ll have to make it more efficient.  The last release I made looked great,  however, had too many server calls to really be acceptable.. one call per tree which would put it at around 25 maybe.

 The thing is, the current tinderbox code doesn’t have any way for me to extrat the critical information that I need to make it less server intensives.  Therefore I will be adding JSON output per tree to make my 0.1 release functional.

Looks like things were not so easy after all for me.. Thats ok, I like a good challenge! Erm, anyone know JSON? (a joke, I swear 🙂

Last weeks class on tweaking Thunderbird was interesting.  Since we haven’t worked on Thunderbird before it gave the whole class a chance to see how that program worked in comparison to the all mighty Firefox.

Well, as it seems, working on Thunderbird feels exactly the same as Firefox!  This might have a lot to do with the fact that they are both based on the same framework.  But that’s the beauty of it!  It’s just another example of how easy it is to work with Mozilla products yet again.

Over the past few weeks since I built FF3 on Ubuntu and blogged about it I’ve received a few comments on it. Surprisingly, people have actually used my how-to! I can only imagine how many others have used it and just not left any comments.

When I first created the post I thought to myself “How could this possibly be useful? Mozilla already has some documentation on the process.” But the documents are quite generic and made for all operating systems.  And it seems that some Ubuntu  users were more than happy to see a direct how-to on this process.

I guess my point is, even if you think that something might not be useful to you it may just be the saving piece of info for someone else lost out there in cyberspace..

The Free Software and Open Source Symposium this past week was an interesting experience for all who attended. With speakers ranging from IBM CEO’s and Microsoft directors to long time teachers on the subject, it seems that there was a place for everyone during this event.

The goals of FSOSS are simple. Gather the greatest minds that drive open source and listen to their ideas and theories on where it’s taken us and where it will. Its a guaranteed great learning experience for people at all computer skill levels.

A few organizations that attended FSOSS as speakers include Facebook, Mozilla, IBM, Microsoft, Open Office, Source Forge, Drupal, Miro.. the list could really go on.. check out the FSOSS 2007 Agenda for a detailed list.

For the measly price of $40 dollars to pre register it is quite a “steal” in terms of educational and informational value. It was also really close for me being in the Toronto area (Just over at York U). The speakers, however, flew in from various parts of North America to talk here so I believe I got my moneys worth on this one.

A highly recommended symposium to anyone hacking away at code and is interested in open source or how you can get involved. (+ you get free lunch too! :).

Well, with a lot of help from the Mozilla community and especially Rob Helmer I’ve finally got a 0.1 to show for it. Its a index page for Tinderbox for those of you who don’t know. Here’s a screen shot snippet:

Tinderbox 0.1 Release

You can find the full screenshot and files at Bug #366784 for viewing / downloading. Now, it’s not exactly beautiful code and I had to make some compromises in terms of code flexebility.. but its here none the less. Nothing that can’t be “cleaned” later.

Well, its actually been submitted for review! So soon you might even see this as the default tinderbox index. (Which I find quite cool) To be able to see your work actually fit into the grand scheme of things like this gives me a good feeling. Like I’ve actually done something worthwhile. 🙂

Feedback always welcome!

Well, I’ve been in contact with Rob Helmer about redoing (or should I just say DOING) an index.html page for the tinderbox.mozilla.org server. As it sits right now tinderbox redirects your to showbuilds.cgi which is a huge and ugly list of trees available for viewing..

What the new index page will provide is a user friendly look at these trees that is useful to the developers using them. Bug 366784 deals with the topic talked about and has the following example by Rob Helmer:

Example Tinderbox layout

This example will be expanded and consist of multiple rows and columns showing tinderbox check in information much more pleasantly. I’ll go about building this page using panels that each trees generates dynamically, probably using iFrames or the like. Ideas are welcome for iFrame alternatives or server performance ideas.. (since this page will be pulling quite some info)