#pentagram@irc.freenode.net logs for 27 Oct 2002 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage


[00:09:40] --- Darke|zzZ is now known as Darke
[00:10:43] <wjp> morning
[00:10:58] <Darke> Morning. *grin*
[00:13:12] --> Kirben has joined #pentagram
[00:13:12] --- ChanServ gives channel operator status to Kirben
[01:03:34] <wjp> I should be going
[01:03:36] <wjp> night
[01:03:39] <-- wjp has left IRC ("Zzzz...")
[02:36:01] <-- Kirben has left IRC (saberhagen.freenode.net irc.freenode.net)
[02:36:20] --> Kirben has joined #pentagram
[04:12:44] --> Colourless has joined #Pentagram
[04:12:44] --- ChanServ gives channel operator status to Colourless
[06:12:12] <-- Colourless has left IRC ("casts invisibility")
[08:32:07] * Darke watches thousands of merge conflicts as he updates to Colourless' disasm code. Thanks Colourless! *grin*
[08:58:40] * Darke ponders what wine goes best with a medium-rare dragon steak. And whether Colourless will need marinading or not, and for how long.
[11:45:58] --> Colourless has joined #Pentagram
[11:46:21] --- ChanServ gives channel operator status to Colourless
[11:49:27] * Darke hides the barbeque utensils and looks innocent.
[11:50:06] <Kirben> watch out!
[11:50:51] <Colourless> not my fault :-)
[11:51:00] <Colourless> you 'could' have committed your changes first :-)
[11:53:14] --> Dark-Star has joined #pentagram
[11:54:41] <Darke> Colourless: I 'could' if you didn't need to be able to disassemble u8's usecode. *grin* I only found out I'd forgotten to handle u8 in my test scripts today, and found it was a little b0rk3n.
[11:54:58] <Colourless> :-)
[12:05:12] <Colourless> want me to make your life hell?
[12:05:48] <Darke> You seem to do that on a regular basis anyway, but what do you propose to do this time? *grin*
[12:07:04] <Colourless> console redirection. Replaces all cout with pout, cerr to perr, printf to con.printf and so on :-)
[12:07:43] <Darke> Oh, that's not much of a problem, I can just fire up kreplace and handle the edge cases on my own. *grin*
[12:09:18] <Colourless> :-)
[12:10:33] <Darke> What you don't see, is that I've replaced all references of i0/i1/... in my copy of disasm/fold with op.i0/op.i1/... that was only about 10 minutes worth of work/checking. *grin*
[12:11:12] <Colourless> /* Yes, this is icky and evil. *grin* But it works without modifying anything. */
[12:11:12] <Colourless> #ifdef FOLD
[12:11:12] <Colourless> #define printf if(print_disasm) printf
[12:11:12] <Colourless> #endif
[12:11:20] <Colourless> *cough* *cough*
[12:11:31] <Colourless> that is only going to break a little bit :-)
[12:12:10] <Colourless> i guess i'll just add in another hack for you :-)
[12:12:18] <Darke> Umm... I hate to say it, but the entire point of me integrating fold/disasm properly was to get rid of that abombination. *grin*
[12:34:09] <Colourless> this is going to cause just a bit of damage to you :-)
[12:35:06] <Darke> Eh? What is? *grin*
[12:35:30] * Darke likes damage, he has a tendancy to route around it. Nasty hacks, though nasty, are fun. *grin*
[12:36:14] <Colourless> :-)
[12:37:49] --> wjp has joined #pentagram
[12:37:50] --- ChanServ gives channel operator status to wjp
[12:40:13] * Darke wonders how wjp likes his dragon cooked? Medium-rare or well-done? Colourless seems to be asking for a barbeque... *grin*
[12:40:35] <Colourless> better now then later, don't you agree? :-)
[12:40:50] <wjp> I'll have it Medium-well-done, thanks :-)
[12:41:15] <wjp> although I prefer to be able to see my food, so maybe we should skip it this time :-)
[12:47:12] * wjp watches line after line of compile errors
[12:47:26] <Colourless> why what?
[12:47:26] <wjp> or maybe a barbeque wouldn't be such a bad idea after all ;-)
[12:47:26] <Darke> Yeah, Just a sec and I'll commit my correct stuff. *grin*
[12:48:02] <wjp> is <ostream> a standard header?
[12:48:13] <Darke> The phrase 'this code highlights the fact windows uses a case insensitive filesytem' comes to mind. *grin*
[12:48:14] <Darke> Yep.
[12:48:20] <wjp> not on my system :-)
[12:48:35] <Colourless> :-P
[12:48:37] * Darke hmms... gcc version?
[12:48:41] <wjp> old :-)
[12:48:55] <Colourless> ostream very much is standard
[12:49:09] <wjp> I have an ostream.h, but no ostream
[12:49:24] <Colourless> hmm...
[12:49:32] <Darke> Yep. It's very standard. Change it to <iostream>.
[12:50:34] <Darke> AFAIR, the only difference is it drags in <istream> as well. *grin*
[12:51:05] <Colourless> not quite... iostream will define the standard output streams too
[12:52:37] * Darke aiees! Warnings and errors everywhere! Mental note: Doing a cvs update can be damaging to your mental health.
[12:53:12] --- Dark-Star is now known as Dark-Star|away
[12:53:29] <wjp> and then there's the 'misc/Console.h:105: parse error before `<''
[12:54:29] <Colourless> 'that' is the fault of your compiler that has issues with namespaces
[12:54:54] <wjp> not sure
[12:55:07] <Colourless> get rid of all the std:: 's and add in
[12:55:14] <Colourless> using std::char_traits;
[12:55:22] <Colourless> using std::basic_streambuf;
[12:55:34] <Colourless> using std::basic_ostream;
[12:55:47] <wjp> why would that help?
[12:56:45] <Colourless> well the guess is it's choaking on
[12:56:45] <Colourless> std::char_traits<_E>
[12:57:39] * Darke acks at the time and afks for a bit.
[12:57:43] --- Darke is now known as Darke|afk
[12:57:45] <Kirben> hmm should pentagram configure work on mingw ?
[12:59:02] <Colourless> hmm
[12:59:41] <wjp> I guess I should upgrade to gcc3 sometime soon
[12:59:41] <Kirben> I get this when trying bootsrap:
[12:59:44] <Kirben> Initial preparation...this can take awhile, so sit tight...
[12:59:44] <Kirben> aclocal: configure.in: 175: macro `AM_PATH_SDL' not found in library
[12:59:44] <Kirben> autoheader: error: AC_CONFIG_HEADERS not found in configure.in
[12:59:44] <Kirben> automake: configure.in: `AM_INIT_AUTOMAKE' must be used
[12:59:44] <Kirben> automake: no proper implementation of AM_INIT_AUTOMAKE was found,
[12:59:44] <Kirben> automake: probably because aclocal.m4 is missing...
[12:59:46] <Kirben> automake: You should run aclocal to create this file, then
[12:59:50] <Colourless> alternative problem is that is you may not have basic_ostream and basic_streambuf
[13:00:12] <wjp> Kirben: some of those errors are normal
[13:00:18] <wjp> (because we don't use automake in pentagram)
[13:00:26] <wjp> the AM_PATH_SDL is weird
[13:00:36] <Colourless> wjp: upgrading to gcc would be a rather good idea :-)
[13:00:43] <Colourless> s/gcc/gcc3/
[13:00:51] <wjp> problem is that the way I'm planning to do that is upgrading to RH8
[13:01:12] <wjp> however that one seems to have a nasty deadlock in its package manager
[13:01:21] <wjp> (grrr :-) )
[13:02:04] <Colourless> my copy of gcc2.95.x doesn't have the basic_?? classes
[13:05:25] <wjp> neither does mine
[13:05:33] <wjp> it also doesn't have char_traits
[13:05:42] <wjp> hm, it does have string_char_traits
[13:11:09] <wjp> I guess I'll go upgrade to RH8 then this week
[13:11:41] <Colourless> in theory it should be possible to work around that somewhat non standard stl support in gcc2.95.x, but it wouldn't be pretty
[13:12:06] <wjp> better not go there
[13:12:58] * Colourless sees "Copyright (C) 1993 Free Software Foundation" in a header, and wonders no more about why it's so outdated
[13:23:46] <-- Colourless has left IRC (saberhagen.freenode.net irc.freenode.net)
[13:23:46] <-- Kirben has left IRC (saberhagen.freenode.net irc.freenode.net)
[13:24:11] --> Colourless has joined #pentagram
[13:24:11] --> Kirben has joined #pentagram
[13:29:38] <wjp> this is so not going to work...
[13:29:45] <wjp> I'm doing a live upgrade from RH73 to RH80... :-)
[13:29:58] <Colourless> :-)
[13:30:19] <Colourless> sounds like a 1 way trip to a nightmare
[13:30:22] --- Darke|afk is now known as Darke
[13:30:25] <wjp> yes :-)
[13:30:55] * Darke returns with a byte. They were all out of bits for the forceable future, words too, but there's a shipment of them coming in on tuesday.
[13:30:59] <wjp> only need to download 765Mb of packages for it
[13:31:13] * Darke ducks behind a couch before people retailliate for his timelapse pun.
[13:31:43] * wjp didn't see a timelapse pun
[13:32:11] <Colourless> Darke: explain
[13:32:31] <wjp> forceable/forseeable?
[13:32:37] <wjp> although I don't see a pun there either
[13:32:59] <Darke> I said when I left, I'd 'afk for a bit', logically then, I'd 'return with a bit'. So...
[13:33:00] <Colourless> i also don't see the pun
[13:33:10] <wjp> ... :-)
[13:33:15] <Colourless> well, gee Darke, that's off the screen
[13:33:18] * wjp had already forgotten that :-)
[13:33:26] <wjp> where 'forgotten' indeed means it's off the screen :-)
[13:33:45] * Darke did try to 'hint' about the 'timelapse' bit, but it would appear he missed the mark. *grin*
[13:34:02] <wjp> can't blame Darke for not realizing it would be off-screen, though. There's not usually this much conversation :-)
[13:34:23] <Colourless> no, not here
[13:35:04] <wjp> anyway, I'm going to leave my system to blowing itself up, while I'm going to enjoy myself with some analytical geometry
[13:35:25] <Colourless> :-)
[13:35:51] <wjp> or maybe some lunch first... hmmm.... (<-- homework-avoiding behaviour ;-) )
[13:36:36] * Colourless notes that we all seem to excel in Pentagram-programming-avoiding-behaviour
[13:37:02] * Darke is _trying_ to do pentagram-programming, but has a little problem with compiling it at the moment. *grin*
[13:37:23] * Colourless looks around innocently
[13:37:42] <Colourless> what? someone broke it? seems to compile just fine for me
[13:39:00] <Colourless> so, mentioned something about 'case insensitive filesystem'. which filename have i got wrong?
[13:39:41] <Darke> It was in the Q_strcasecmp.cpp file, it referenced "q_strcasecmp.h". *grin*
[13:39:58] <Colourless> oh, i see
[13:42:57] <Colourless> any problems with my console code?
[13:43:55] <Darke> Other then it's missing a <cstdarg> include, disasm seems to output identically with your new code. Congrats. *grin*
[13:44:12] * wjp notices a 'copyright ID software' header :-)
[13:44:19] * Colourless notices a few :-)
[13:44:58] <Colourless> hey, can you blame me from grabbing things from the Quake 2 source :-)
[13:45:06] <wjp> if you want me to :-)
[13:45:28] * Darke guesses he'll have to walk through there doing a s/unsigned int/uint32/, etc to make them 'pentagram compliant'. *grin*
[13:46:45] <Colourless> you can now do a con.Dump("filename.txt"); to dump the contents of the console buffer
[13:47:46] <Colourless> because the console is designed for graphics it has a fixed width. the id software code is nice that it does word wrapping
[13:48:07] <Darke> Neat!
[13:48:30] <Colourless> but, there is a slight problem, it wont work with pout and perr :-)
[13:49:17] <Darke> Which is not as much of a problem as one might expect, since most of the stuff I'll be dumping there won't want to be wordwrapped. *grin*
[13:49:50] <Colourless> should probably add in an option to disable word wrapping
[13:54:37] * Darke nods. Would be handy.
[14:03:46] <Kirben> ok disasm still compile on mingw
[14:04:17] * Colourless celebrates :-)
[14:05:56] <Colourless> i like it when my changes at least compile for 1 other person
[14:05:56] * Darke snickers.
[14:06:10] <Kirben> stack of warnings though
[14:06:30] <Darke> To do with implicit typename depreciation?
[14:07:31] <Colourless> what is an 'implicit typename depreciation'
[14:07:59] <Kirben> yes
[14:13:16] <Colourless> what lines are causing the error?
[14:13:44] <Kirben> 116
[14:13:46] <Kirben> 161
[14:13:59] <Colourless> and which file? :-)
[14:14:03] <Kirben> misc/Console.h
[14:14:19] <Darke> http://gcc.gnu.org/ml/gcc/2002-08/msg00010.html <- A description of the 'problem'.
[14:16:49] <Colourless> ok, i'll see if I can make a fix
[14:17:14] <Colourless> which typename is it complaining about?
[14:18:05] <Kirben> typename console_err_streambuf<_E, _Tr>::int_type
[14:18:09] <Darke> int_type.
[14:18:53] * Darke has 'fixed' it, though it's ugly. *grin*
[14:19:02] <Colourless> ok, try adding the following in the public parts of the classes
[14:19:02] <Colourless> typedef _Tr::int_type int_type
[14:19:04] <Colourless> ;
[14:20:42] <Darke> You actually need: typedef typename _Tr::int_type int_type;
[14:21:04] * wjp needs to read up on the 'typename' keyword
[14:21:23] <Darke> And it's fine for it to be in the protected: area.
[14:22:12] <Colourless> yeah, the protected area is fine, but to be consistent with other streambuf classes, it would normally be in public
[14:22:25] <Colourless> typename identifier;
[14:22:25] <Colourless> Use this keyword only in template definitions. This keyword tells the compiler that an unknown identifier is a type. For example:
[14:22:39] <wjp> yeah, I'm reading stroustrup right now
[14:22:43] <Colourless> template<class T> class X {
[14:22:44] <Colourless> typename T::Y; // treat Y as a type
[14:22:44] <Colourless>
[14:22:44] <Colourless> Y m_y;
[14:22:44] <Colourless> };
[14:26:01] <Kirben> ok, typedef typename _Tr::int_type int_type; in public fixes the warnings
[14:27:08] <Colourless> C++ Classes for Everyone in #pentagram right now!
[14:27:17] * Darke snickers.
[14:28:32] * Darke is actually giving himself perl classes. His testing script is insufficiently flexable. *grin*
[14:28:55] <Colourless> :-)
[14:33:34] <Colourless> someone want to commmit the fixed file?
[14:34:38] * Darke will let someone else do it. *grin*
[14:40:17] <Colourless> i guess it will end up being me :-)
[14:40:35] <wjp> I would do it, but my system is currently not exactly in a stable state :-)
[14:41:35] * Darke would do it, except both is compile trees are in variable amounts of unstable state. *grin*
[14:41:59] <Kirben> committed
[14:42:21] <Kirben> along with mingw makefile
[14:42:24] <Colourless> i guess it wasn't me :-)
[14:42:41] <Colourless> goo
[14:42:42] <Colourless> d
[14:44:03] <Colourless> uh oh kirben, the line ending are messed in the makefile
[14:44:45] <Colourless> looks like you committed a file with Dos/Windows line endings as a file with UNIX endings
[14:45:08] <Kirben> ok
[14:46:08] <Kirben> see if it is alright now
[14:46:42] <Colourless> nope, still bad
[14:47:39] <Kirben> how to fix ? i thought u2d util would do it
[14:48:09] <wjp> it sounds like you need dos2unix for it
[14:48:13] <Colourless> you need to go the other way round
[14:48:59] <Kirben> ok will try d2u
[14:49:50] <Kirben> ok now I hope ?
[14:50:24] <Colourless> yes, that's good
[14:50:27] <wjp> looks ok, yes
[14:50:36] <Kirben> good
[14:57:02] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[15:48:50] --- Darke is now known as Darke|zzZ
[18:17:44] <-- Colourless has left IRC ("casts invisibility")
[18:56:07] --> Dominus has joined #pentagram
[18:59:33] <Dominus> what the... is going on here? so many cvs comitts? something is wrong here :-)
[18:59:56] <wjp> :-)
[19:27:03] --- Dark-Star|away is now known as Dark-Star
[19:54:11] <-- wjp has left IRC ("brb")
[20:02:17] --> wjp has joined #pentagram
[20:02:17] --- ChanServ gives channel operator status to wjp
[20:02:31] * Dark-Star reads the recent discussion...
[20:02:58] <Dark-Star> I want to get a good C++ book, too. Which one ist better? Stroustrup's or "Accelerated C++"??
[20:03:17] <wjp> I don't know the latter
[20:03:37] <Dark-Star> It's relatively new, but from browsing through it, it seems quite good too
[20:03:39] <wjp> Stroustrup is pretty good, but mostly as a reference
[20:06:58] * Dark-Star already has C++ experience, just need to get some hints on templates and such...
[20:12:17] <-- wjp has left IRC ("brb again")
[20:25:24] --> wjp has joined #pentagram
[20:25:24] --- ChanServ gives channel operator status to wjp
[20:38:50] <-- wjp has left IRC (Remote closed the connection)
[20:39:29] --> wjp has joined #pentagram
[20:39:29] --- ChanServ gives channel operator status to wjp
[21:13:04] <-- Darke|zzZ has left IRC (saberhagen.freenode.net irc.freenode.net)
[21:14:02] --> Darke|zzZ has joined #pentagram
[21:18:37] <-- Darke|zzZ has left IRC (saberhagen.freenode.net irc.freenode.net)
[21:18:57] --> Darke|zzZ has joined #pentagram
[22:33:33] <wjp> ok, pentagram compiles for me now :-)
[22:33:55] <Dominus> he he
[22:46:59] <-- Dark-Star has left IRC ()