[15:37]<ashb> where would people think is a good place/name ot put the current os name? [15:37]<ashb> Windows/Darwin/Unix etc [15:37]<ashb> system module? [15:38]<MisterN> system.uname? [15:40]<ashb> hmmm probably as .osName and .machineHardware [15:40]<ashb> or similar [16:15]<ondras> require("os").name [16:15]<ondras> this allows adding some more properties [16:15]<ondras> .version, .arch, ... [16:17]<ashb> ondras: yeah that was the other thought [16:17]<MisterN> ashb: good that we dropped the os module, eh? :P [16:17]<MisterN> so we can use it for this purpose now [16:18]<ashb> yeah [16:18]<ashb> what did it used to even do? [16:18]<ondras> :)) [16:18]<ashb> i've forgotten [16:19]<ashb> oh it had sleep and system() fns [16:36]<MisterN> ashb: i think fork was only in posix [16:51]<Wes-> "software arch" might an interesting property [16:51]<Wes-> unix / os2 / windows / novell [16:51]<Wes-> that sort of thing [16:51]<Wes-> so that linux, solaris, aix, osf/1, xenix, freebsd, etc all fall into the same category [16:52]<Wes-> or maybe "systemArch" [16:53]<ondras> this is a huge bikeshed to be colored! [16:54]<ondras> I bet we will be able to find at least ten different colours for it [16:54]* ondras implementing remote storage for his roguelike [16:54]* ashb is fighting cmd.exe quoting [16:56]<MisterN> Wes-: maybe more like flags [16:56]<MisterN> Wes-: like the posix macros in C [16:57]<Wes-> MisterN: not a bad idea, either! [16:57]* Wes- is learning bsd's tree.h macros [16:58]<Wes-> This is like C++ templates, only without syntax [16:59]<MisterN> Wes-: yeah, you can even do metaprogramming with macros :) [17:00]<Wes-> yep. I know a lot of macro tricks. Better docs for this sucker would be nice, though, it's tricky when you're not the macro author. ;) [17:00]<Wes-> (you should see the macros in gpsee, I have whole classes of polymorphic functions) [17:02]<MisterN> Wes-: have you seen what we do with macros in flusspferd? :) [17:02]<Wes-> no, but I imagine it's equally fun. :) [17:02]<Wes-> You want to see insane, though, check out the generated regexps in gpsee's ffi code [17:02]<Wes-> some of them are > 10K [17:02]<Wes-> I actually had to break some up because of limitations in GNU grep [17:03]<MisterN> oO [17:03]<MisterN> Wes-: this is our insane macro stuff: http://flusspferd.org/docs/group__classes.html#gaa23a1baded44f78a6606863d2cbfac7d [17:04]<MisterN> implementation: http://github.com/ruediger/flusspferd/blob/master/include/flusspferd/class_description.hpp [17:04]<MisterN> Wes-: writing macros with named parameters is a very cool trick IMHO [17:04]<Wes-> oh yeah [17:05]<Wes-> I'm curious why you couldn't use templates in there, mind you [17:05]<MisterN> because we need to generate actual code. [17:05]<MisterN> templates don't do the trick there. [17:05]<Wes-> Oh, right - I bumped into that problem before myself [17:05]<MisterN> in C?? [17:05]<Wes-> I have a nice trick I use a fair bit in GPSEE, I have a "declaration file" which gets include many times in many different files, sometimes mroe than once in each [17:06]<Wes-> the declaration file is actually macro'd to code, or enums, or arrays, or whatever each time it is included [17:06]<Wes-> No, in C++, I was writing a database module class a few years back [17:06]<Wes-> rather, wrapping my C library [17:08]<Wes-> I wrote a beautiful template class that was worthless as a DSO because it didn't instanciate anything [17:18]<MisterN> Wes-: oh no, that was another problem [17:18]<MisterN> Wes-: that's really not the problem we have here [17:19]<MisterN> Wes-: also it's a foolish beginner's mistake that you made there :P [17:19]<Wes-> MisterN: Like I said, I don't actually know C++ :) [17:19]<ashb> i would like it very much if cmd.exe would DIAF [17:20]<MisterN> Wes-: you could probably have solved the problem by putting the whole template in the header. [17:21]<Wes-> Probably, but the whole point was to get 300+ interfaces through to PHP. I just wound up instanciating database objects based on char * arguments instead of compile-time symbols [17:22]<Wes-> Or maybe that was for CXXWRAP (JNI generator) and not PHP? I forget [17:23]<Wes-> All I know is that when I wrote the JSAPI version of that wrapper it was really easy comparitively [17:23]<Wes-> Because I actually knew what I was doing, lol [17:23]<MisterN> Wes-: you wrote a c++ template in order to interface to PHP? [17:23]<Wes-> (and I didn't care at all about host language type safety) [17:24]<Wes-> MisterN: Yes, IIRC the ZEND macros that shipped with PHP5.0 required C++ classes to get PHP objects [17:24]<Wes-> (not really my area of specialty, I just had to whip something together) [17:24]<Wes-> I do recall reading Zend's macros, though. They suck. [17:24]<MisterN> PHP... [17:25]<Wes-> Yeah, everything about that language sucks [17:25]<Wes-> except that it has that one key feature: "ship date" [17:25]<MisterN> ? [17:25]<Wes-> MisterN: "ship date" is the most important product feature. You can leave plenty of other features out, but not ship date. [17:26]<MisterN> Wes-: i don't really understand [17:26]<Wes-> MisterN: The most awesomest piece of software in the world will sell zero copies if you don't release it [17:27]<MisterN> Wes-: yeah, but python has releases, too [17:27]<MisterN> Wes-: i'm not particularly fond of python but i'd use it ten times over php [17:27]<Wes-> You can use python to write server-parsed web pages? [17:27]<MisterN> of course? [17:27]<Wes-> And, more importantly, was released at about the same time as PHP? [17:27]<MisterN> no, but people use PHP for _new_ stuff [17:28]<Wes-> not me [17:28]<MisterN> yeah you use it for old stuff :P [17:28]<Wes-> I'd rather write a whole server-side javascript backend from scratch [17:28]<evilstreak> :D [17:28]<Wes-> more important, I don't write ANY PHP code. Ijust maintain some. [17:30]<ondras> yeah. code purity is an important thing. [17:31]<ondras> never touch those, my baby, they might hurt you! [17:36]<Wes-> I already have brain scars from PHP globals [17:39]<ashb> "Shipping doesn?t mean you win, but not shipping means you lose." [17:41]<Wes-> exactly! [17:45]<ashb> ah - in reference to my earlyier question. [17:45]<ashb> python has sys.platform [23:54]<isaacs> what's the minimum that a Binary implementation has to offer to be useful Binary module?