Building Mozilla, or geeky gluttony for punishment

[lethargic_man]

Mozilla is a good web browser, in many ways. However, it does have some flaws in it, one of which is that it doesn't render vowellated Hebrew properly in Linux: the vowels render between the letters, rather than above, below and within them.

Turns out, I learned from Bugzilla, that this is a known problem and at some point support for the native Linux renderer, Pango (which is capable of rendering Hebrew correctly) will be built in, though as the above bug has been open for five years now, I'm not holding my breath.

The solution, it seems, is to build Mozilla myself with --enable-pango, rather than just downloading a prebuilt executable.

So, I download the Mozilla source at work, stick it onto my USB fob, and bring it home. I open up the tarball, and start writing a .mozconfig file, as recommended, starting it with . $topsrcdir/browser/config/mozconfig. Only browser/ doesn't exist at all. I can't figure out what's going on here. The Mozilla source code tarball appears not to contain the source code for the Mozilla browser! Am I missing something here (apart from the code)? Apparently not: make -f client.mk build fails.

Running make -f client.mk checkout MOZ_CO_PROJECT=browser appears to be starting to get the missing code. But doing a CVS checkout of even only a portion of the codebase over a narrowband connection is almost certainly not a good idea (the part I already have comes to 319M uncompressed!).

Looks like I'm going to be taking my USB fob back along with me to work tomorrow. (Then rinse, lather and repeat.) Possibly the solution is to try building it there, then zip up the whole code hierarchy to bring home once I've got to the point where it builds successfully.

Comments

[compilerbitch.livejournal.com]

Mozilla is Not Small. I have built it from source myself (can't remember why now) some time ago, and I do remember it being something of a pain.

I can also recommend Bulldog 8MBit/s broadband...

[liv]

I think this does rather prove my point that handling Hebrew on the web requires a lot more work than just making sure everything supports Unicode. Good luck getting it sorted, though.