#pentagram@irc.freenode.net logs for 31 Aug 2003 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage


[00:01:21] <wjp> ok.. got the dagger
[00:02:43] <wjp> heh, the ritual with the dagger is executed outside :-)
[00:04:35] <DarkeZzz> Eh?
[00:04:45] * DarkeZzz has never played u8. *grin*
[00:04:49] <wjp> ah :-)
[00:04:54] <wjp> you should fix that :-)
[00:05:37] * DarkeZzz 's problem is he's never had a machine that could play u8, still doesn't if he wants proper sound/music. *Grin*
[00:05:56] <DarkeZzz> Have you tried to get u8 working in dosbox yet? *grin*
[00:05:58] * wjp points at pentagram :-)
[00:06:03] <wjp> u8/dosbox doesn't work yet
[00:06:24] <wjp> u7 does, but no u8 :/
[00:06:36] <wjp> (funny, that... I would've expected it to be the other way around)
[00:06:52] <DarkeZzz> I'll let you trip over all the bugs first, then I'll give it a go. *grin*
[00:08:00] <DarkeZzz> Nah. U8's memory manager is probably quite a bit more complex and picky then u7's, which just seems to brute force a solution. *grin*
[00:08:17] <wjp> u7 uses some rather dirty 386 tricks, though
[00:10:00] <wjp> ugh, another segfault just after saving
[00:10:41] <wjp> 0x0806f000 in SoftRenderSurface<unsigned>::Paint(Shape*, unsigned, int, int) (
[00:10:42] <wjp> this=0x810ff60, s=0x811fb20, framenum=15, x=-1486, y=-135387606)
[00:10:42] <wjp> at SoftRenderSurface.inl:193
[00:10:44] <wjp> nice coordinates :-)
[00:12:36] <wjp> _now_ vividos is standing by Lothian's side...
[00:12:40] <wjp> just a tad too late :-)
[00:14:43] <DarkeZzz> Heh.
[00:15:59] --> Kirben has joined #pentagram
[00:15:59] --- ChanServ gives channel operator status to Kirben
[00:16:06] <wjp> hi Kirben
[00:16:54] <wjp> ok, entered the catacombs :-)
[00:17:34] * wjp cheats a bit to avoid climbing
[00:19:47] <wjp> yay! entered the necromancer's test :-)
[00:20:15] <wjp> reagent combining isn't really working yet...
[00:20:20] <wjp> it's too picky
[00:22:38] <wjp> and of course item un-combining would be nice too :-)
[00:22:41] <DarkeZzz> Heh.
[00:24:20] <wjp> ok, Dead Speak works
[00:25:51] <wjp> being invulnerable is quite convenient here :-)
[00:31:59] <DarkeZzz> It would be. *grin*
[00:32:19] * DarkeZzz officially disappears. *poof*
[00:33:31] <wjp> whee! finished the tests :-)
[00:33:45] <wjp> and drowned in lava! :-)
[00:33:49] <wjp> grrr :-)
[00:36:06] <Kirben> Hi
[00:40:06] <wjp> stone cove...
[00:44:37] <wjp> hm, time to go
[00:44:38] <wjp> bye
[00:44:41] <-- wjp has left IRC ("Zzzz...")
[07:49:08] <-- Kirben has left IRC ("brb")
[07:49:44] --> Kirben has joined #pentagram
[07:49:44] --- ChanServ gives channel operator status to Kirben
[07:50:15] <-- Kirben has left IRC (Client Quit)
[07:50:40] --> Kirben has joined #pentagram
[07:50:40] --- ChanServ gives channel operator status to Kirben
[09:12:54] --> Colourless has joined #Pentagram
[09:12:55] --- ChanServ gives channel operator status to Colourless
[09:13:20] <Colourless> hi
[09:57:26] --> Dark-Star has joined #pentagram
[10:11:28] <DarkeZzz> ?date
[10:11:28] <exultbot> It is now Sun Aug 31 10:11:28 2003 (GMT).
[10:13:15] * DarkeZzz furtivelooks. No, he's not making a commit. Really. He's just... err... wondering what the date is, yeah.
[10:18:05] <-- Colourless has left IRC (Read error: 110 (Connection timed out))
[11:30:35] --> wjp has joined #pentagram
[11:30:35] --- ChanServ gives channel operator status to wjp
[11:41:05] <wjp> DarkeZzz: committed _again_, I see? :-)
[11:41:10] <wjp> what is the world coming to... :-)
[11:41:32] <-- Dark-Star has left IRC (adams.freenode.net irc.freenode.net)
[11:42:56] --> Dark-Star has joined #pentagram
[11:43:01] * DarkeZzz denies it!
[11:43:19] <wjp> ah, in that case somebody else seems to be working on fold
[11:44:33] <wjp> I got as far as Stone Cove last night :-)
[11:44:45] <wjp> (not that I got stuck there, but it was getting too late :-) )
[11:44:45] <DarkeZzz> Cool. *grin*
[11:45:09] <wjp> and not being able to un-stack reagents makes spellcasting a bit hard
[11:48:20] <wjp> hm, good thing you denied committing things
[11:48:31] <wjp> I'd have to blame you for breaking he build otherwise
[11:48:36] <wjp> s/he/the/
[11:49:13] <DarkeZzz> Eh? I did? Where?
[11:49:14] <wjp> is there maybe a certain file a certain person forgot to commit? :-)
[11:49:29] <wjp> (not you, of course :-) )
[11:52:48] <DarkeZzz> Umm... quotes.txt?
[11:53:03] <wjp> I was thinking more of tools/disasm/module.mk actually
[11:54:15] <DarkeZzz> Ahh. That's 'cause it's not modified in my version, since I haven't recompiled the entire thing, just fold. *grin*
[11:55:51] * DarkeZzz makes changes and waits for compile to complete. *clawtap*
[11:57:43] <DarkeZzz> Incredibly complex file commit done. Things should work again now. *grin*
[11:58:04] <DarkeZzz> (Not that I actually did the original commit, or anything, or for that matter the last commit either. It was... err... somebun else!)
[12:03:26] <wjp> hm, interesting... this monster egg here seems to have spawned quite a large number of ghosts :-)
[12:05:16] <DarkeZzz> Our cache in/out stuff isn't working correctly?
[12:06:38] <wjp> hard to say
[12:07:19] <DarkeZzz> I suppose it's quite possible it's supposed to do that...
[12:07:22] <wjp> it's spawning a couple of ghosts per second
[12:07:29] <wjp> so _something_ is going wrong :-)
[12:07:41] <DarkeZzz> Hrm... that definately sounds Quite Wrong(tm). *grin*
[12:08:19] <wjp> I thought monster eggs were supposed to mark themselves hatched
[12:11:04] <wjp> interesting
[12:15:45] * DarkeZzz earperks?
[12:22:22] <wjp> it seems this monsteregg has an x/y-range of 1
[12:23:07] <wjp> so if you step exactly on it, it will hatch
[12:23:38] <wjp> and since monstereggs have no built-in protection like being hatched multiple times like normal eggs, they keep getting hatched
[12:25:24] <wjp> I'm starting to think monster eggs are just not supposed to be hatched
[12:25:34] <wjp> so the npcnum field would be for something else instead of range
[12:31:36] <DarkeZzz> Hrm... makes sense.
[12:32:01] <wjp> leaves the question just _what_ the npcnum field is, though :-)
[12:32:41] --> Colourless has joined #Pentagram
[12:32:41] --- ChanServ gives channel operator status to Colourless
[12:32:49] <wjp> hi
[12:34:08] <Colourless> hi
[12:36:16] <wjp> I'm starting to wonder if a MonsterEgg should be a subclass of Egg at all :-)
[12:36:36] <DarkeZzz> Hi.
[12:38:34] <Colourless> i don't think it should be
[12:38:43] <Colourless> it's just a container i think
[12:38:55] <wjp> container?
[12:38:57] <Colourless> that holds the info for the unkegg that does the actual spawninng
[12:39:07] <wjp> oh, that kind of container :-)
[12:39:18] <Colourless> yeah, wrong word to use :-)
[12:42:10] <wjp> that leaves just one Egg subclass
[12:43:03] * DarkeZzz watches wjp become One With The Ovum.
[12:43:06] <wjp> (TeleportEgg)
[12:43:14] <wjp> hmm...
[12:43:28] * wjp waves bye-bye to his Stone Cove savegame
[12:43:59] <Colourless> you didn't have to kill your save game :-)
[12:44:52] <wjp> it might be salvageable
[12:45:16] <wjp> but that would probably be more work than replaying it :-)
[12:50:50] <wjp> well, the fact that nothing breaks when changing the parent from Egg to Item is a good indication that it probably shouldn't be an Egg :-)
[12:56:10] <wjp> ok, no more infinitely many ghosts now :-)
[13:01:19] <wjp> combat...
[13:01:31] <wjp> I wonder what needs to be done to be able to hit monsters
[13:01:31] <Colourless> oh no :-)
[13:02:14] <Colourless> probably more than you'd think
[13:02:19] <wjp> (fun subtask: use wpnovlay.dat :-) )
[13:02:47] <wjp> Item::receiveHit is one thing
[13:03:58] <wjp> Item::gotHit
[13:04:25] <wjp> death...
[13:04:43] <wjp> alignments can probably wait
[13:05:17] <Colourless> gotHit is collision detection related
[13:06:27] <wjp> probably calls the corresponding usecode event?
[13:06:44] <wjp> ok, so receiveHit is probably the main issue
[13:07:01] <wjp> and probably figuring out the various animation flags related to combat
[13:07:39] <wjp> receiveHit(uword, byte, word, uword)
[13:08:26] <wjp> the uword is an objid
[13:08:30] <wjp> the byte is a direction
[13:09:58] <wjp> other two seem to indicate type/strength of hit?
[13:10:08] <wjp> (not too sure...)
[13:10:15] <wjp> the word is often a call to random
[14:37:46] --> SB-X has joined #pentagram
[16:14:13] <wjp> Colourless: if I wanted to create an explosion effect, what would I need to do?
[16:14:18] <wjp> just create a SpriteProcess?
[16:14:23] <Colourless> yes
[16:14:33] <wjp> very convenient :-)
[16:15:05] <Colourless> with the correct parameters though obviously
[16:15:20] <Colourless> all it will do is play the animation, nothing else though
[16:15:21] <wjp> somehow I expected that :-)
[16:15:51] <wjp> "nothing else"? are you referring to anything specifically that it doesn't do?
[16:17:10] <wjp> (I'm not expecting it to deliver hits to other items or anything, if that's what you mean :-) )
[16:17:29] <wjp> (or interact with anything in the world, for that matter)
[16:55:31] <Colourless> both :-)
[17:08:20] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[17:43:46] --> RaT_HaCk has joined #pentagram
[18:00:44] <wjp> uh oh
[18:01:07] <wjp> good news: I found a way to dramatically increase our activity rating
[18:01:50] <Colourless> explain
[18:02:07] <wjp> it seems we got "infected" by exult :-)
[18:02:35] <wjp> about 90-100 files... *sigh* :-)
[18:03:03] <Colourless> oh no :-)
[18:03:57] <Colourless> it's in old too :-)
[18:04:36] <Colourless> what it looks like is that 'bit' was copied from exult into one pentagram file. then it was copied into every new file too
[18:04:54] <wjp> not every file, though
[18:10:44] <wjp> ok, committed that too
[18:11:53] <Colourless> so, am i going to get more cvs emails than a certain worm sent me?
[18:12:32] <wjp> you get cvs emails from a worm? :-)
[18:12:54] <Colourless> actually, im' amazed i didn't :-)
[18:12:58] <SB-X> Fexult.w32
[18:13:30] <Colourless> then again, as you said, it seems we got infected by exult :-)
[18:14:19] <wjp> it's fairly localized in pentagram
[18:19:36] <Colourless> but you have eradicated the infection now
[18:19:51] <wjp> yes :-)
[18:25:53] <Colourless> you know, we can use cvs history to find who was the 'carrier'
[18:26:40] <wjp> sure :-)
[19:02:43] <wjp> heh, you can now cause a nice explosion when you drop an oil flask into the campfire
[19:07:03] <Colourless> :-)
[19:07:33] <wjp> hm, I should try putting some more oil flask around the campfire when doing that :-)
[19:09:45] <wjp> or even better, cast conflagration :-)
[19:10:36] <wjp> lol!
[19:10:45] <wjp> the explosions get hurled away by other explosions :-)
[19:10:55] <Colourless> hehe
[19:13:24] <Colourless> now i have to compile like every file in pentagram
[19:14:47] <wjp> been there... done that... :-)
[19:17:03] <-- RaT_HaCk has left IRC (Read error: 54 (Connection reset by peer))
[19:18:15] <Colourless> you know, i like pentagram's ultra responsive moving much more than the originals really lagg
[19:18:18] <Colourless> y movement
[19:18:27] <Colourless> of course that really just has to do with the turning delay
[19:18:32] <wjp> hm, Item::grab() isn't working in all cases yet
[19:18:43] <wjp> yes, turning delay isn't in yet :-)
[19:18:51] <wjp> (I wasn't in any particular hurry to implement that...)
[19:21:33] <Colourless> i think it should be something that we can toggle
[19:21:59] <Colourless> the purist can play with turning delay on, the rest will play without
[19:23:23] <wjp> z-differences need some work
[19:23:31] <wjp> it's insane to get blocked by a tiny object on the ground while running
[19:23:45] <Colourless> you don't say :-)
[19:26:47] <Colourless> it's funny, i actually imagine that every 3d game engine that's developed has problems with tiny things on the ground blocking movement
[19:29:27] <wjp> yeah... it's a tricky issue :-)
[19:43:46] <wjp> ok, committed some receiveHit/explode stuff
[20:19:18] --> Dominus has joined #pentagram
[20:19:20] --- ChanServ gives channel operator status to Dominus
[20:19:46] <Dominus> hey ho, you busy people
[20:19:59] <Colourless> hey
[20:24:03] <Dominus> phew, anonymous cvs access on sf is really not working for me atm (trying to get dosbox)
[20:31:07] <wjp> hi
[20:31:15] <wjp> same for me
[20:31:39] <wjp> there haven't been many dosbox commits lately though
[20:32:38] <Dominus> mhm, just wanted to get the latest...
[20:33:26] <Dominus> there is also a daily (or 2-days-apart) binary snapshot availlable for Win32 (if someone like Colourless has interest)
[20:33:55] * Colourless is busy
[20:34:27] * Colourless is doing one of two things
[20:34:28] <wjp> BG2? ;-)
[20:35:05] * Colourless is either doing what wjp says, or coding shaders. Though he has finished doing the latter for now. :-)
[20:35:29] <Dominus> http://vogons.zetafleet.com/showthread.php?s=&threadid=2023 link to the dosbox snapshot forum stuff
[20:36:09] <wjp> I've been playing Planescape: Torment a bit recently
[20:36:13] <Colourless> thinking about it, i should probably commit the exult (and probably pentagram old) 2xSaI fix, that i mentioned in #exult 'yesterday' (my time)
[20:36:43] <Colourless> Torment is next on my list of old games to play again :-)
[20:36:44] <Dominus> I did have problems with running u7 in dosbox as a new game would for some reason not include a certain file. starting one from u7for xp and then continuing in dosbox worked though
[20:37:41] <wjp> Colourless: you mean that the rectangle that needs to be updated needs to be enlarged by one or two pixels?
[20:37:56] <Colourless> that isn't actually the problem :-)
[20:38:09] <Colourless> it's more subtle than that
[20:38:21] <wjp> in that case I'm definitely curious :-)
[20:38:28] <Colourless> enlargine the rectangle will never actually fix the problem
[20:38:42] <wjp> Dominus: oh? I think I managed to start a new game
[20:39:16] <Colourless> the problem is related to the code the is used to do clipping when attempting to get pixels that are off the left, or top of the screen
[20:39:27] <Colourless> s/the is used/that is used/
[20:41:16] <Colourless> when accessing a pixel that is to the left of the current it does this:
[20:41:17] <Colourless> colorG = *(bP - prev1_xoff);
[20:41:22] <Colourless> (or similar)
[20:41:44] <Colourless> prev1_xoff is normally set to 1, so it will read the pixel to the left (from the source obviously)
[20:42:06] <Colourless> when clipping is required because the current pixel is the left most, prev1_xoff is set to 1
[20:42:11] <Colourless> set to 0 i mean
[20:42:44] <wjp> so that pixel gets twice the 'weight'?
[20:42:58] <Colourless> the problem is of course that prev1_xoff is never actually set to 1, when it's no longer being clipped
[20:43:24] <wjp> ah :-)
[20:43:41] <Colourless> same thing happens with prev1_yoff too
[20:44:07] <wjp> so it's only looking to the right and bottom most of the time?
[20:44:16] <Colourless> yes :-)
[20:44:34] <wjp> oops :-)
[20:44:43] <wjp> hm, SDL 1.2.6 is out
[20:44:45] <Colourless> the ironic thing, the area around the mouse cursor was the only part of the screen being properly scaled, the rest of the screen wasn't
[20:45:06] <Colourless> while we were thinking the opposite :-)
[20:45:34] <Colourless> ah oh well. Only requires 2 extra lines of code to fix
[20:45:39] <wjp> since it wasn't clipping and so it was actually set to 1... funny :-)
[20:45:50] <wjp> hm, SDL 1.2.6 has mmx/3dnow alpha blitters
[20:46:08] <wjp> dynamic object loading functions
[20:46:24] <wjp> hm, "greatly improved X11 DGA video speed"
[20:46:32] <wjp> mmx audio mising code
[20:46:38] <wjp> s/mis/mix/
[20:47:00] <wjp> FSAA support for opengl
[20:47:05] <Colourless> well, i think you would really hope that the mmx audio code wasn't 'missing' anything :-)
[20:47:12] <wjp> (and lots and lots of others :-) )
[20:47:12] <Dominus> he
[20:47:24] <Colourless> 2xSaI scaling will probably look better after I commit too, since it that clipping problem was really messing up the algorithm
[20:47:49] <wjp> do you have the author's email?
[20:47:53] <wjp> (or was this one of mame's?)
[20:48:42] <Colourless> kreed hacked in the clipping code into exult after we were getting crashes
[20:48:52] <Colourless> (normal 2xSaI code didn't have clipping)
[20:50:59] <Colourless> Super2xSaI and SuperEagle work properly though
[21:09:29] <Colourless> ok, committed the fixes
[21:14:09] <Dominus> got to reboot.
[21:14:14] <-- Dominus has left IRC ("a pooka invited me to Charlie's")
[21:15:58] <Colourless> i should be going
[21:16:00] <Colourless> cya
[21:16:03] <-- Colourless has left IRC ("casts invisibility")
[21:56:51] <SB-X> where did dominus go?
[22:03:54] <wjp> hm, yes, that reboot is taking a bit long
[22:21:54] <wjp> you'll love the new function I mapped to the middle mouse button here :-)
[22:21:59] <wjp> item->explode(); ;-)
[22:37:03] <SB-X> i expect that to go in the release version *nods*
[22:37:35] <SB-X> you could battle with explosions, if one hurls another one
[22:42:07] <wjp> oh wow... I started a (anon)cvs update an hour ago and it actually started now
[22:42:23] <wjp> (well, not really "now", but it did start :-) )
[22:51:40] <wjp> funny... the exult ML has been getting spam with a _really_ wrong date for a while now
[22:51:51] <wjp> mails from 22 march 2002 today
[22:52:11] <wjp> also got mails from 16, 19 and 20 march 2002
[22:52:27] <wjp> so it seems somebody with a really wrong system clock is spamming :-)
[23:02:51] <SB-X> or the spam got caught in a temporal causality loop
[23:02:57] <SB-X> some new countermeasure?
[23:03:42] <wjp> heh :-)
[23:04:01] <wjp> "Let them deal with it in 18 months when they probably have much more advanced spamfilters"? :-)
[23:06:22] <SB-X> those anomalies are unpredictable, couldve been intended to stay there 18 years
[23:06:28] * SB-X shrugs.
[23:06:32] <wjp> hm, indeed :-)