This is from one of the devs in an "off-the-record" QA session here: http://www.shacknews.com/laryn.x?id=17826898#itemanchor_1782...
488 karma · joined May 10, 2007
This is from one of the devs in an "off-the-record" QA session here: http://www.shacknews.com/laryn.x?id=17826898#itemanchor_1782...
What struck me is how technical it was (for a comic book quick overview). It's obviously geared towards developers--they really wanted to plant their flag in the sand and say "we're doing this: here is why it is good."
I can't speak for Mozilla's collaborative culture (and the fact that Firefox emerged from Netscape means it can't be half-bad), but just on an aesthetic level, WebKit C++ is a lot easier on the eyes than Mozilla. One thing that's interesting about WebKit is that in the last 12 months when I've been pulling upstream changesets from them, I've noticed that they're not afraid to do project-wide renames to keep the code clean. It seems like a small thing, but it's a lot more approachable to see clean C++ that uses namespaces and abstracts away platform differences in a coherent way. Mozilla is full of COM interfaces like
#define NS_IHTML_CSS_STYLE_SHEET_IID \
{0xb5cc4ac0, 0xeab6, 0x11d1, {0x80, 0x31, 0x00, 0x60, 0x08, 0x15, 0x9b, 0x5a}}
class nsIHTMLCSSStyleSheet : public nsIStyleSheet
... and so on. edit: To be fair, I'm just a lot more familiar with WebKit, so maybe someone can defend (or pick apart) Mozilla's sources better than I can. If you check out WebKit, you'll also notice that there's an impressive collection of 210MB(!) of individual layout tests.Really, this was inevitable. Their whole tabs-as-process thing is too, when you think about it. Just imagine the potential for a browser-as-OS-with-processes-and-everything and how much they must be salivating to get it on devices and desktops NOT running Windows. It's a hell of a trojan horse in their winning battle to make Microsoft irrelevant...
And open source? <3
I'm so tired of the ___-changing or -changers meme! It's the worst kind of empty meaningless dribble that you'll find in bad technology articles.
A new Ruby VM? It's a gamechanger! Startup X integrated feature Y! This is unbelievable, gamechanging, news!
>>> funcs = []
>>> for x in xrange(3):
>>> funcs.append(lambda: x)
>>> sum(f() for f in funcs)
6
Were you expecting 3? Lambdas (besides having such an awfully long and difficult to type name :P) capture only the "name" of x, so when the for loop changes x later, the lambda sees the new value--so the sum is (2 + 2 + 2) instead of (1 + 2 + 3).What you can do, instead (and I do it often enough that I wish lambdas created their own lexical closure....)
>>> funcs = []
>>> for x in xrange(3):
>>> funcs.append(lambda x=x: x)
>>> sum(f() for f in funcs)
3
By using x as a default argument, you force the x inside the lambda to be the value it was when the lambda was created. It kinda sucks, and always felt to me like a deficiency of the implementation leaking up into the language :[I'd like to poke around with Python's compiler module and look at the AST--you could enforce that a certain class of functions have no side effects, and then spawn multiple processes to execute them semi magically.
Futures like the ones in IO (http://www.iolanguage.com/scm/git/checkout/Io/docs/IoGuide.h...) always struck me as a pretty cool way to abstract away synchronization problems. Using the multiprocess stuff as a background would be fun.
It virtualizes c:\Program Files for "compatibility" so that you can see bizarre situations like writing to a file and then having it not exist where you thought it was.
To avoid these virtual store problems, you have to embed XML in your application's executable to get the infamous ALLOW or DENY window to bother the user. If you use Visual Studio 2005 to do it, watch out! There's a bug in a certain class of unpatched XP that blue screens the computer because of the way the XML is structured that Visual Studio's manifest tool spits out when it merges manifest fragments. It's all overly complicated.
I would never develop Windows software if there wasn't a market for it--it's exciting to see the underdogs picking up marketshare.
Right now I'm experimenting with on an alternative to wxPython called wxpy using a forked version of the SIP bindings generator, instead of SWIG, to see if I can eek out more performance and bring down the memory usage:
But that's like the 10th time hard-to-read-text has happened to me today, and it gives me an idea.
On-the-fly, stored in the cloud site specific CSS customizations would be a killer feature for a Firefox extension. Something that useful must already be in Firebug ;) /goes to check
BUT I don't like that their webpage is kind of sluggish, or that they delegate issue tracking out to a pay service (lighthouse)
in a related note, committing/branching all the time locally and then just doing "git push origin" when I feel like pushing all those changes to github is AMAZING. it really changes the way you deal with rapidly changing code. I'm just learning git but the nerdhype seems strongly justified with this one :)