#pentagram@irc.freenode.net logs for 10 Apr 2004 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage

[01:32:31] --> Kirben has joined #pentagram
[01:32:31] --- ChanServ gives channel operator status to Kirben
[01:56:16] <Kirben> disasm tool causes link errors now.
[01:57:20] <Kirben> http://members.optusnet.com.au/scummvm/misc/error.txt
[02:02:02] <Darke> Try adding 'misc/istring.o' to the list of objects.
[02:04:36] --> Sheng_Gradilla has joined #pentagram
[02:04:41] <Sheng_Gradilla> good afternoon :)
[02:06:47] <Darke> *waits a little for the clock to tick past 12* Afternoon!
[02:54:01] --> servus has joined #pentagram
[04:30:38] <-- servus has left IRC (Read error: 54 (Connection reset by peer))
[04:30:39] --> servus_ has joined #pentagram
[05:26:18] --> Kirben2 has joined #pentagram
[05:26:45] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[05:26:55] --- Kirben2 is now known as Kirben
[05:27:00] --- ChanServ gives channel operator status to Kirben
[06:35:26] <-- Sheng_Gradilla has left IRC ("zzz")
[09:29:52] <-- servus_ has left IRC ("Leaving")
[11:25:29] --> Colourless has joined #Pentagram
[11:25:29] --- ChanServ gives channel operator status to Colourless
[11:25:52] <Colourless> hi
[11:26:04] <wjp> hi
[11:29:13] <Colourless> yay.... i broke pentagram somehow :-)
[11:29:26] <wjp> uh, yay! (?)
[11:29:30] <wjp> ;-)
[11:29:59] <Colourless> it fails to load the shapes
[11:31:52] <Colourless> tsk tsk. running in debug mode found 1 problem
[11:32:06] <wjp> what did I do? :-)
[11:32:15] <Colourless> the pointers ShapeInfo are not initialized to 0
[11:32:22] <Colourless> pointers in ShapeInfo
[11:32:26] <wjp> oh, right
[11:32:42] * Colourless writes a constructor
[11:33:04] <wjp> actually only monsterinfo is uninitialized
[11:33:22] <wjp> assuming that TypeFlags::load is the only place ShapeInfo objects are created
[11:33:31] <wjp> but I agree a constructor is probably better :-)
[11:33:40] <Colourless> ok works now :-)
[11:34:34] <Colourless> strange
[11:34:48] <Colourless> it was working earlier today, and stoped working just now
[11:35:04] <wjp> time-delayed bugs?
[11:35:11] <Colourless> :-)
[11:54:42] --> sbx has joined #pentagram
[12:03:25] <wjp> hmmm... the teleporter to the teleport pad opened by the skull of quakes appears to be broken
[12:06:53] <Colourless> it's always been a bit problematic
[12:07:10] <Colourless> but it was working
[12:07:21] <wjp> teleport coords for there: 50 15762 6030 48
[12:07:55] <wjp> the teleport/mark/recall functions make life a lot easier :-)
[12:08:01] <Colourless> indeed they do :-)
[12:10:04] <Colourless> yes it doesn't seem to be working
[12:10:28] <Colourless> got it
[12:10:43] <Colourless> took me ages but eventually it triggered
[12:11:02] <Colourless> it's collision detection related :-)
[12:11:10] <wjp> oh yay
[12:19:54] <wjp> ah, yes, got it too
[12:20:02] <wjp> egg range is only (5,1)
[12:20:11] <Colourless> 5 1!
[12:20:18] <wjp> maybe we're misinterpreting the range again
[12:20:46] <Colourless> that seems far too small
[12:20:54] <wjp> it's multiplied by 32
[12:21:18] <Colourless> *32 would seem about right
[12:24:38] <wjp> hm
[12:24:49] <wjp> when you walk down the stairs, you're at 16090,6111
[12:25:07] <wjp> the y range of the egg is [6047,6111)
[12:25:36] <wjp> (6047 inclusive, 6111 exclusive)
[12:25:37] <Colourless> z might be important
[12:25:43] <wjp> don't check z
[12:25:53] <wjp> we don't, that is
[12:26:53] <wjp> when teleporting back from the shrine, you get stuck in the floor
[12:28:18] <wjp> I wonder if the problem with the egg not getting triggered might be the walking algo
[12:28:44] <wjp> it doesn't seem to be flexible enough about z differences
[12:29:01] <Colourless> might be something like that
[12:29:28] <Colourless> ok
[12:29:29] <Colourless> try this
[12:29:41] <Colourless> press 'c' to enable collision with cursor movement
[12:29:45] <Colourless> then use the cursors to move
[12:29:51] <Colourless> it triggered the egg right away here
[12:30:28] <wjp> hm, yes, same here
[12:30:34] <wjp> ok, so it's most likely the walking
[12:34:50] <Colourless> i'll check it out
[12:35:19] <Colourless> or you can if you want :-)
[12:35:36] <wjp> I won't stop you if you want to do it ;-)
[12:35:38] <Colourless> i honestly can't remember exactly how we trigger eggs
[12:35:54] <wjp> EggHatcherProcess
[12:36:05] <wjp> it's fairly simple
[12:36:41] <Colourless> ok i see
[12:37:40] <Colourless> i wonder why it's not working
[12:38:03] <wjp> most likely there's some garbage lying on the floor that the avatar refuses to step onto
[12:38:05] <Colourless> is it actually an egg or some other trigget type
[12:38:15] <wjp> it's a normal egg
[12:38:42] <Colourless> yeah might be something the avatar isn't stepping over
[12:39:00] <Colourless> or getting stuck on
[12:40:37] * wjp makes middle mouse destroy objects to get a closer look
[12:41:23] <wjp> ummm... it seems explosions actually blow eggs away
[12:41:46] <Colourless> yeah :-)
[12:42:08] <Colourless> we're not handling some of the type flags correctly yet WRT explosions
[12:42:11] <wjp> I don't think we can just make item searches ignore editor items, though
[12:42:40] <Colourless> might be one of the other flags
[12:42:48] <Colourless> might even be a status flag
[12:42:58] <Colourless> (but i don't think that it would be a status flag)
[12:43:37] <Colourless> SI_IGNORE perhaps?
[12:43:47] <wjp> hm, possible
[12:44:24] <wjp> ok, it seems the object blocking the avatar is actually a wall
[12:44:38] <Colourless> we currently 'ignore' SI_IGNORE ;-)
[12:44:39] <wjp> and apparently we're not moving the avatar close enough to the wall
[12:45:15] <Colourless> might be why cursors works. it moves as far as it can
[13:25:29] <wjp> I committed a change to the egghatcher; it now checks the entire footpad of the avatar instead of just the avatar's x,y coords
[13:25:49] <Colourless> that should help out
[13:28:32] <wjp> I wonder how I got stuck in this... I was originally working on dying
[13:33:28] <wjp> other egg troubles: the cheat-teleport egg in the lower catacombs is way too easy to trigger
[13:33:49] <wjp> although this is a real teleporter egg, not a usecode egg
[13:37:35] <wjp> so we may need to check the Z coordinate of eggs anyway
[13:38:43] <Colourless> i would think we have to
[14:20:28] <Colourless> wjp, looking at the execution scene, it now seems like the detection of the npc infront is actually working
[14:20:46] <Colourless> shaana stops when she reaches darion
[14:21:01] <Colourless> instead of overtaking him
[14:21:04] <wjp> wasn't it just one of them that was inverted?
[14:21:43] <Colourless> well, that's what we thought
[14:21:55] <Colourless> but it seems to be working now
[14:23:38] <Colourless> head doesn't seem to be hitting the water anymore either
[14:25:07] <wjp> hm, indeed
[14:25:16] <wjp> no splash
[14:25:33] <Colourless> actually the not overtaking might be because we are doing some collision detection in the animation playing
[14:25:40] <wjp> yes
[14:25:42] <Colourless> rather than the usecode now working correctly
[14:38:39] <-- Kirben has left IRC ("System Meltdown")
[14:39:15] <Colourless> committed some stuff. Among what i committed includes removing the Q_strcasecmp file (the functions have been moved to istring)
[14:39:37] <Colourless> you might need to update some of your build stuff
[14:54:07] <wjp> build times are starting to get really insane :-)
[14:55:02] <Colourless> i think part of that reason is too much of stl is beginning to get pulled in by pent_include.h
[14:55:25] <Colourless> of course building for me is pretty quick because of you know why :-)
[14:56:34] <wjp> anyway, it builds just fine
[14:56:53] <Colourless> i modified one of the module.mk files :-)
[14:57:57] <wjp> oh no! *gasp* :-)
[14:58:37] <wjp> speaking of build files, your msvc workspace/project files aren't in CVS, right?
[14:58:46] <Colourless> nope they aren't
[14:59:11] <Colourless> i was thinking of actually creating an nmake makefile for cvs
[14:59:24] <wjp> should you ever feel the need to commit them, we should probably consider a build/ directory with a msvc (or something) subdir
[14:59:40] <wjp> we could move some of the linux build system support files there too
[14:59:56] <Colourless> having some 'system' directories or something would be useful
[15:00:21] <Colourless> since all win32 builds will (eventually) share some files, but other things will be separate
[15:03:18] <wjp> hm, you disabled the framerate display?
[15:03:31] <Colourless> yeah it's off by default :-)
[15:03:39] <Colourless> there is a console command to turn it back on
[15:03:55] <Colourless> GUIApp::drawRenderStats 1
[15:08:44] <Colourless> you might want to consider adding a class prefix to the teleport, mark and recall commands. they mess up the nice alphabetical sorting that Console::CmdList gives :-)
[15:09:16] <wjp> hm, feel free to add it :-)
[15:09:21] <Colourless> plus it would be more consitent with the other few commands
[15:09:38] <wjp> although we could also remove the class prefixes from the others :-)
[15:10:16] <wjp> brb
[15:11:01] * Colourless adds MainActor:: to the front of them
[15:16:33] <wjp> back
[15:22:37] <Colourless> wb
[16:38:23] <wjp> dinner; bbl
[17:17:28] <wjp> back
[17:19:11] <Colourless> wb
[20:13:50] --> Fingolfin has joined #pentagram
[20:13:50] --- ChanServ gives channel operator status to Fingolfin
[20:14:58] <Fingolfin> hi
[20:16:42] <sbx> hello
[20:26:13] --> Dominus has joined #pentagram
[20:26:14] --- ChanServ gives channel operator status to Dominus
[20:26:42] <Dominus> hi all
[20:27:15] <Dominus> I'm taking the 3 hours pentagram compile time now :-)
[20:27:45] <sbx> hello
[20:35:51] <Dominus> wow, the map change is almost seamless. didn't notice that before
[20:37:21] <Dominus> that is not right. map change should fade out, then grind the hard drive for 5 minutes and then either crash pentagram or actually load the map
[20:39:13] <Dominus> oops, fell off the map
[20:41:22] <Fingolfin> actually, that reminds me, I think pentagram is hanging on exit again for me...
[20:42:40] <-- Colourless has left IRC ("casts invisibility")
[20:49:09] <wjp> Fingolfin: hm, not good :/
[20:55:37] <Fingolfin> I'll do a CVS up and try it with gdb
[20:56:02] <wjp> since when, btw?
[21:00:18] <Fingolfin> I noticed yesterday, didn't try for a week or so (?)
[21:01:02] <wjp> hm, there was a bug with uninitialized pointers getting destructed yesterday
[21:03:24] <Fingolfin> I actually didn't do a gdb run yesterday. I'll do that now
[21:03:44] <Fingolfin> only noticed that pentagram frozen when I quit it, and I had to kill -9 it
[21:05:04] * Fingolfin gets a zillion link errors... hum
[21:07:15] * wjp tries to build latest cvs to see if anything broke
[21:07:40] <Fingolfin> ld: multiple definitions of symbol _ZN15console_ostreamIcSt11char_traitsIcEEC1EP17console_streambufIcS1_E.eh
[21:07:40] <Fingolfin> misc/Args.o definition of absolute _ZN15console_ostreamIcSt11char_traitsIcEEC1EP17console_streambufIcS1_E.eh (value 0x0)
[21:07:40] <Fingolfin> misc/istring.o definition of absolute _ZN15console_ostreamIcSt11char_traitsIcEEC1EP17console_streambufIcS1_E.eh (value 0x0)
[21:07:43] <Fingolfin> and more
[21:07:54] <Fingolfin> all between Args.o and istring.o
[21:08:01] <Fingolfin> I did a "make clean" of course
[21:13:07] <wjp> hm, things link properly for me
[21:15:05] <wjp> it kind of looks like a duplicate template instantiation
[21:17:51] <Fingolfin> yeah
[21:18:00] <Fingolfin> it did link fine yesterday
[21:19:00] * wjp hmms
[21:19:22] <wjp> try moving the explicit template instantiations from misc/Console.h to misc/Console.cpp
[21:19:36] <wjp> (the two lines starting with 'template class')
[21:25:33] <-- sbx has left IRC ("bye")
[21:31:47] <Dominus> ?logs
[21:36:24] <Fingolfin> that didn't really help
[21:37:31] <wjp> not really or really not?
[21:38:30] <Fingolfin> http://fingolfin.dnsalias.net/~maxhorn/pentagram-stuff/build.log
[21:39:00] <Fingolfin> hmm
[21:39:13] <Fingolfin> I just noticed that I also had removed my Apple GCC 3.x work around ...
[21:39:14] <Fingolfin> see http://fingolfin.dnsalias.net/~maxhorn/pentagram-stuff/pentagram.patch
[21:39:22] <Fingolfin> that was OK with the new GCC 3.3
[21:39:29] <Fingolfin> maybe we are now triggering the problem again, or a related bug
[21:51:50] <Fingolfin> nope, doesn't help either
[21:52:46] <Fingolfin> hm number of errors changed, though. This is with a straight from CVS Console.h/.cpp. O get 2833 lines of errors
[21:53:04] <Fingolfin> with the template class stuff move to Console.cpp, "only" 1783 lines :-)
[21:54:09] <wjp> well, sounds like an improvement :-)
[22:01:31] <Fingolfin> yup indeed, 1783 errors, and the Apple GCC 3.x hack wasn't at fault (i.e. I can indeed safely remove it)
[22:02:18] <Fingolfin> erorrs which are ggone include
[22:02:18] <Fingolfin> misc/Console.o definition of console_ostream<char, std::char_traits<char> >::console_ostream[in-charge](console_streambuf<char, std::char_traits<char> >*) in section (__TEXT,__text)
[22:02:22] <Fingolfin> -> which is no surprise
[22:02:29] <Fingolfin> one of the remaining errors
[22:02:29] <Fingolfin> misc/Console.o definition of Pentagram::pent_string<char, std::char_traits<char>, std::allocator<char> >::pent_string[in-charge](char const*) in section (__TEXT,__text)
[22:02:33] * Fingolfin looks at istring.o
[22:02:38] * Fingolfin looks at istring.h
[22:02:39] <Fingolfin> :-)
[22:02:59] <Fingolfin> template pent_string<char>;
[22:03:01] <Fingolfin> hmmm :-)
[22:06:16] <wjp> ah, so Colourless hid more of them in the code :-)
[22:16:29] --> ctzen has joined #pentagram
[22:18:05] <-- ctzen has left #pentagram ()
[23:00:57] <Fingolfin> btw, removing that template fixed all link errors
[23:01:18] <Fingolfin> wonder if there is any reason to keep those template instantiations?
[23:03:16] <Fingolfin> and hang on exit seems gone, too, good :-)
[23:08:55] <wjp> hm, I can't really think of a reason for those
[23:09:02] <wjp> they should be instantiated anyway
[23:54:27] <-- Dominus has left IRC ("a pooka invited me for a drink")