About a year ago, I had a good browser at my disposal: it was
simply called Mozilla
, developped by the Mozilla foundation, and every
couple of weeks I would compile a new version of it, test it a little,
and, if it seemed stable, use it. But then Mozilla died and the problems started
for me. Ever since then I've been using the last pre-mortem version
of Mozilla I compiled (actually, on 2005-05-22, so it's a little later
than the official death announcement), but it has various problems
(including known and serious security vulnerabilities) and I would
like to switch to something more recent.
The official replacement for Mozilla is called Firefox. Now some people think a world of good of it: I certainly don't. Basically Firefox was created by taking Mozilla, stripping a whole lot of useful and important stuff from it (such as the nice newsreader, which was seamlessly integrated in the Mozilla suite and which is now a separate program, Thunderbird, with a totally crippled interaction with Firefox). Why anyone would want to remove features from a program and promote it as better totally beats me. Plus, the user interface of Firefox is totally crippled: for example, Mozilla had a nice tabbed sidebar and the Firefox designers somehow decided that tabs are bad for the user and removed them, so the sidebar is now stupidly limited to displaying one kind of view from what was previously an easy to navigate list; another example, is the utterly brain-dead addition of the Google (or search engine) bar: it used to be that I could perform Web searches directly from the URL bar, but someone decided that the two should be separated and invented the idiotic Google bar which eats up precious screen real-estate for no reason[#] at all. Fortunately, various people have hinted ways to get rid of the Firefox Google bar and still be able to search things from the URL bar (one solution is using Web keywords, which I don't like because I don't want to start searching for a term merely because I made a typo in a URL; another, better, solution, is bookmarklets, which aren't perfect, but they're the best I found). So Firefox is somewhat usable: not nearly as good (now) as Mozilla was (a year ago), but still better than nothing.
Another candidate as successor to Mozilla is SeaMonkey: the word
SeaMonkey
used to be the project name for the Mozilla suite
(the one I used a year ago and still use), and the code base has been
continued by a team of developers, not directly connected to the
Mozilla foundation but still using some of their resources (such as
the Mozilla CVS repository), to keep Mozilla alive. Sort
of. So in principle I should be using this. Except that, no, the new
SeaMonkey developers aren't competent enough, or many enough, to drive
the huge Mozilla code base: so they haven't been able to come up with
something which works. I just tried it, compiled it, ran it, and as
soon as I tried performing an interactive search it died with a very
helpful error message,
/opt/mozilla-20060219/lib/seamonkey-1.5a/seamonkey-bin: symbol lookup error: /usr/local/opt/mozilla-20060219/lib/seamonkey-1.5a/components/libctl.so: undefined symbol: pangolite_find_map
Great.
Now to make things more complicated: every version of
Mozilla (since 2003 or so), including Firefox, SeaMonkey and
all derivatives, has very stupid
bug causing unbreakable spaces to be silently transformed into
ordinary spaces in any kind of cut-and-paste or form submission;
fortunately, there is a very
simple fix for that bug (just remove three erroneous lines of
code), which, unfortunately, the developers have refused to include
(and while they agreed in principle to include a more
elaborate—and less correct—, it has been waiting to be
committed for over a year now). I really hate that bug (it totally
breaks up Wikipedia, for example, which is a major disaster, because
it means that if anyone using Firefox makes any
change in a Wikipedia page containing an unbreakable space, that
unbreakable space gets replaced by an ordinary space; and you can't
write either, because if you do a wikirobot
will transform it into a genuine unbreakable space), and certainly I
won't tolerate it on the browser I use. Well, since the
patch is there, all I need to do is patch the source and
recompile.
Except that recompiling Firefox, or SeaMonkey, is a world of pain.
Especially if you're not going to use the latest (CVS
“head”) version, which is basically unstable and often
unusable. These people don't know how to use CVS, and instead of
cleanly tagging every bit of source code used in the tree with some
uniform tag, they disseminate the code across a forest of trees
(mozilla/security/nss mozilla/security/coreconf,
mozilla/modules/libmar, mozilla/db/sqlite3,
SeaMonkeyAll…—what a pleasure!) and don't
tag things uniformly: so just getting the right source code to compile
is a major mess. And that's not counting the quadrillion
configuration options (now, should I use
--enable-fretype2 or --enable-xft? and how
should I checkout libart-lgpl?), with no documentation
wheresoever as to how the release builds were generated, exactly.
Well, the net result is that so far I've been unable to come up with a Mozilla-like browser which works for me. I'm curious to know what other people use.
(Of course, there also exist non-Mozilla-based browsers in this universe. They don't support XUL, however, and, for me, this is a show-stopper.)
Extra note: I've just discovered another problem:
today Debian upgraded the
libxft2 library in its Etch distribution
(the one I use) to version 2.1.8.2-3 (previous was 2.1.7-1). And with
the right setup of fonts this causes a crash of all Mozilla-related
browsers (and various other applications, including things which
worked fine up to yesterday) with another of these helpful error
messages:
/usr/lib/mozilla-firefox/firefox-bin: symbol lookup error: /usr/lib/libXft.so.2: undefined symbol: FT_GlyphSlot_Embolden
Similar problems had previously been reported: it does not appear to be Mozilla's fault in any way, and only marginally Debian's: the idiots in this case are solely the libfreetype/libxft developers who haven't heard of downward compatibility.
[#] To make my point clearer: imagine they had invented 26 different bars, which you had to use according to the first letter of the search or address one wants to perform. That would be unimaginably stupid. Well, the Google bar is perhaps not quite as stupid, but it isn't too far either.