[16:41] <Keybuk> hmm, looks like clang gained support for more gcc extensions
[16:42] <Keybuk> still doesn't support local labels though ;'(
[16:49] <sadmac2> Keybuk: ah. nothing like combining preprocessing and goto :)
[16:49] <Keybuk> sadmac2: hmm?
[16:49] <sadmac2> Keybuk: that's mostly what I use local labels for, when I use them.
[16:50] <Keybuk> they're handy for autogenerated code
[16:50] <sadmac2> that too
[16:50] <Keybuk> where you have local "get out of here on error" code
[16:50] <Keybuk> but you might have a nested loop inside with the same auto-generated code
[16:50] <Keybuk> (which is exactly how they end up in Upstart - the D-Bus bindings code uses them)
[16:51] <sadmac2> almost the same thing really. The preprocessor basically is a (really weak) autogenerator
[17:12] <Keybuk> io.c:681:3: warning: expression result unused [-Wunused-value]
[17:12] <Keybuk>                 NIH_ZERO (nih_io_message_add_control (message,
[17:12] <Keybuk>                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
[17:12] <Keybuk> In file included from io.c:42:
[17:12] <Keybuk> ../nih/macros.h:176:48: note: instantiated from:
[17:12] <Keybuk>         ({ typeof (_e) __ret; while ((__ret = (_e))); __ret; })
[17:12] <Keybuk>                                                       ^~~~~
[17:12] <Keybuk> ..
[17:12] <Keybuk> of course the other problem with clang/llvm at this point is the lack of documentation
[17:12] <Keybuk> so I have no idea how to teach it that it's quite ok for that expression result to be unused
[18:22] <sadmac2> Keybuk: does clang compile to native code or only to that llvm bytecode language?
[18:23] <Keybuk> clang is an llvm frontend
[18:24] <Keybuk> so clang produces bytecode
[18:24] <Keybuk> llvm turns that into assembly
[18:24] <Keybuk> this isn't much different from gcc
[18:24] <sadmac2> ok. the llvm page seemed kind of confusing about that.
[18:36] <Keybuk> llvm isn't purple enough
[18:54] <sadmac2> Keybuk: purple?
[18:58] <Keybuk> sadmac2: I've been dealing all day with minor changes to the splash screen
[18:58] <Keybuk> wrong colour of purple, etc.
[18:59] <sadmac2> Keybuk: purple? Isn't Ubuntu brown?
[18:59] <Keybuk> apparently someone told the design team it was UI Freeze today, so they've been mass-landing the entire Ubuntu/Canonical rebranding in one go
[18:59] <Keybuk> no, it's purple now
[18:59] <sadmac2> gj
[18:59] <sadmac2> the number of people for whom that is the primary Ubuntu complaint is...unfortunate.
[18:59] <Keybuk> http://arstechnica.com/open-source/news/2010/03/ubuntu-dumps-the-brown-introduces-new-theme.ars
[19:00]  * sadmac2 reminds himself to pull an Ubuntu pulseaudio package
[19:01] <sadmac2> I've never heard Ubuntu user's complain about pulse, and I can't for the life of me imagine how you could have made it stop hemorraging failure and disappointment.
[19:01] <Keybuk> oh god
[19:01] <Keybuk> the best bit is that Lennart complains about our pulse packages, saying that they're broken and wrong
[19:01] <Keybuk> and this is why all our users complain about how broken pulse audio is
[19:01] <Keybuk> no, we say, pulse audio is a broken heap of shit
[19:02] <sadmac2> Keybuk: I wanna replace it. Lets write an audio server this weekend
[19:02] <sadmac2> No scratch that. Audio servers are wrong to begin with.
[19:02] <sadmac2> Open source disasters: when the only person willing to do the work is the one least qualified.
[19:03] <sadmac2> What was that quote about being careful around the man who is stupid and industrious?
[19:04] <Keybuk> heh
[19:06] <sadmac2> Wayland has less code than Pulse. its great :)
[19:06] <sadmac2> Of course I think Wayland might have less code than rm
[19:06] <sadmac2> its kind of scary how small that thing is.
[19:13] <sadmac2> http://en.wikipedia.org/wiki/Kurt_von_Hammerstein-Equord