#exult@irc.freenode.net logs for 31 Jan 2004 (GMT)

Archive Today Yesterday Tomorrow
Exult homepage

[00:13:09] --> Fingolfin has joined #exult
[00:13:09] --- ChanServ gives channel operator status to Fingolfin
[00:24:49] <wjp> hi
[01:15:21] <Fingolfin> rehi
[01:26:52] <Fingolfin> wjp: having a shifted down "zero" shouldn't affect the sound at all (although it of course hints at a bug anyway). I.e constant "0" will be silence, but any other constant data sequence is silence, too
[01:27:14] <Fingolfin> mostly, this reduces the audio range, that should be all bad about it... "should"
[01:27:50] <Fingolfin> but as I stated last night, since when I implemented that code, I went a long way in the audio/resampling area, and now will say that using a cubic interpolator isn't such a bright idea after all =)
[01:28:28] <Fingolfin> SDL_sound has a really nice well written high quality resampler built-in, BTW. Might be an idea to use it for Exult
[01:28:43] <Fingolfin> after all, SDL_sound also makes it very easy to load & play lots of sound formats...
[01:28:57] <Fingolfin> sadly we can't use it in Exult, because, as the name implies it requires SDL
[01:29:21] <Fingolfin> so far I haven't found an efficient, portable resampler w/o deps on a some kind of toolkit like SDL...
[01:29:45] <Fingolfin> sox has nice resampling code, but using floats/doubles (not efficient); other code has other limitations...
[01:30:05] <Fingolfin> but I was thinking about "borrowing" the SDL_sound code for ScummVM anyway (adapting it to our needs there)... ah well, we'll see <sigh>
[01:43:17] --> Kirben has joined #exult
[01:43:17] --- ChanServ gives channel operator status to Kirben
[01:51:47] <-- Rixa has left IRC (Read error: 104 (Connection reset by peer))
[01:52:00] <-- wjp has left IRC (Read error: 104 (Connection reset by peer))
[01:52:31] --> wjp has joined #exult
[02:05:53] <-- Fingolfin has left IRC ("42")
[02:53:59] <-- SorcererX has left IRC (Read error: 60 (Operation timed out))
[02:55:47] <-- Andersen has left IRC ("FUTUE TE ET IPSUM CABALLUM")
[04:49:53] --> DrCode has joined #exult
[04:50:37] <DrCode> Hi! First time I've connected when someone else was here in quite a while.
[04:51:49] <DrCode> Then again, maybe everyone is asleep:-)
[04:52:39] * Darke waves. Hi!
[04:53:17] <DrCode> Hi. Must be really late or really early there.
[04:53:29] <Darke> Looks like everyone but me maybe. Then again, it'll often take half an hour for people to respond to a 'hi', so... *grin*
[04:53:40] <Darke> Only around 3pm. Midafternoon actually.
[04:55:21] <DrCode> Guess it's summertime, too.
[05:02:21] <Darke> Yup.
[05:02:39] <Darke> AKA, far far far too hot. *grin*
[05:11:09] <DrCode> Too rainy here. But I'm starting to enjoy the snow in the mountains.
[05:20:39] <DrCode> Since all the Europeans are asleep, I guess they can't object to my idea of making a branch to work on the multiple maps.
[05:21:40] <Darke> Nope. I can't see a problem with that either, myself.
[05:22:13] <Kirben> Has anymore been found out about how original games set initial schedules for npcs ?
[05:25:13] <DrCode> I think we've know that for a while; it's a flag in npc.dat. I guess ES doesn't have the option of setting it, though.
[05:26:09] <Kirben> So would it be possible to allow finnigan to be out run at start of BG now ?
[05:26:49] <DrCode> I think he's set in motion by the usecode that's executed at the very start.
[05:27:12] <DrCode> He's set to "talk", and that's supposed to make him approach you.
[05:28:33] <DrCode> ES has a limitation: You can only start it up to edit NPC's after you've started Exult, which is after the game's already started.
[05:29:03] <DrCode> Map-editing really should take place before the game's started at all.
[05:36:17] <Kirben> hmm maybe original game checked how close an NPC was, before allowing them to talk to Avatar ?
[05:37:32] <Kirben> Switching between full screen and window seems to be disabled during conversations at the moment.
[05:39:13] <DrCode> We're probably not doing it exactly right, although we do check distance, I think.
[05:39:29] <DrCode> And we're probably not checking all possible keys when in conversation mode.
[05:46:57] <DrCode> Back to coding... See ya.
[05:46:59] <-- DrCode has left IRC ("using sirc version 2.211+KSIRC/1.2.4")
[07:39:31] <-- Matt_O has left IRC (Read error: 104 (Connection reset by peer))
[07:39:32] --> Matt_O has joined #exult
[10:35:29] --> Rixa has joined #exult
[10:36:57] --> Colourless has joined #Exult
[10:36:57] --- ChanServ gives channel operator status to Colourless
[10:37:21] <Colourless> hi
[10:47:38] <-- servus has left IRC ("Client exiting")
[11:07:59] <Darke> Hiya.
[11:17:11] <Colourless> Fingolfin: the bias introduced on sound resampling can cause clicking when a voice starts to play and then stops
[11:48:17] <Kirben> Mouse pointer is been shown too early in games again (Before Avatar appears), looks like an old regression caused by too much code cleanup. Looks like Filter_intro_events() was used to solve problem long ago (In exult.cc) but it got removed for some reason...
[11:51:04] <Kirben> and some of the functions used by Filter_intro_events() no longer exist either.
[11:52:17] <Kirben> http://cvs.sourceforge.net/viewcvs.py/exult/exult/exult.cc?r1=1.330&r2=1.331
[12:00:42] <Colourless> the code may have been removed in that revision, but it had already been #ifdef'd out
[12:00:55] <Colourless> s/#ifdef'd/#if 0'd/
[12:03:02] <Kirben> Yes but at least that shows where the function was originally used.
[12:04:20] <Kirben> http://cvs.sourceforge.net/viewcvs.py/exult/exult/exult.cc?r1=1.169&r2=1.170 was the original fix for the problem, long ago. If that is any help.
[12:05:30] --> SorcererX has joined #exult
[12:05:31] <Colourless> is this actually a problem though?
[12:07:22] <Kirben> Well it is only a cosmetic issue, but looks bad when mouse pointer can be moved around too early in BG/SI introductions.
[12:09:04] <Colourless> i personally don't care. In fact it was probably me who changed the code
[12:09:16] <Colourless> part of it anyway
[12:11:39] <Colourless> showing the mouse cursor is also a 'nice' idea for people playing windowed. It's nice to always know where the mouse cursor is
[12:14:35] <Colourless> iirc one of the changes that was made is that whenever the dont_move flag is set the mouse cursor will always be shown as a hand (how it works in SI sequences). As such it automatically makes the cursor the hand and you can't move. As such the hacky Filter_intro_events() code wasn't needed anymore, and IMO that is a good thing
[12:15:37] <Colourless> remove an 'as such' or 2 from that to make more sense :-)
[12:36:39] --- ChanServ gives channel operator status to wjp
[12:37:19] <wjp> hi
[12:38:34] <Kirben> Hi wjp
[12:45:59] <Colourless> hi
[12:47:57] <wjp> hey, a multimap branch, cool :-)
[13:03:55] <SorcererX> hmm... I should ask the Gentoo developers to update the Exult port in Portage (it's still at 1.00)
[13:15:19] <wjp> or maybe wait until 1.2
[13:33:31] <wjp> heh, amazing... gentoo has a NWN 1.61 package :-)
[13:34:45] <wjp> of course I didn't notice when I actually could've used it :-)
[13:35:26] <Colourless> and that is amazing how?
[13:35:33] <Colourless> it's not like you don't have NWN 1.61 yourself
[13:35:46] <wjp> ok, maybe not amazing... more like surprising
[13:36:24] <Colourless> and when did they get it?
[13:36:35] <Colourless> remember you got 1.61 on like the day it was out :-)
[13:37:53] <Colourless> interesting indeed though, since you need to register with Bioware to download the linux patch
[13:38:53] <wjp> package contains a download link to the '129to161eng' update
[13:39:22] <wjp> (automatically downloading it)
[13:46:20] <Colourless> were linux to become a popular desktop operating system, it might become unfeasible for companies like Gentoo to actually have a downloading system like that. Bandwith costs, and storage costs (assuming locally stored files) would be too expensive
[13:58:13] <Darke> Not really. Most of the files you're downloading will come from one of multiple mirror sites. Last time I checked gentoo had it's ebuild tree coming from about 30 different mirrors, and most of the other files are grabbed from the remote websites, most of the *.sf.net packages get grabbed from a random server from the sf.net download system for instance.
[13:59:09] <Colourless> the thing is, what number of people are downloading the files now?
[14:00:00] <Colourless> I'm talking a 'massive' increase in userbase, along with many more files. Assumption being if linux were a popular desktop os, then more software would be ported to/made for it
[14:01:23] <Darke> I dunno. How many people are downloading from sf.net? Or downloading patches from bioware? gentoo doesn't really have a central download spot for anything other then the portage tree itself, which is realtively small, and only really needs to contain minor patches to the tarballs and occasional binary files downloaded from elsewhere.
[14:07:02] <Darke> That admittedly is one benefit of gentoo over redhat, etc, it grabs the large files from the site that distributes it, itself, and only supplies minor patches, if that, to get it to fit nicely into the gentoo environment. Whereas redhat/suse/etc have to support their own large package mirror and distribution system.
[14:08:56] <Colourless> linux 'really' needs a unified system where the operating system itself will configure a generic package
[14:10:03] <Colourless> and of course needs everyone to support it
[14:11:31] <Darke> That's part of what the whole UnitedLinux fuss was about a while back. Have a nice standardized system of $thing goes $here so you could basicly install something and it would Just Work(tm).
[14:12:25] <Colourless> Linux needs less linux 'fans' involved with it's development :-)
[14:13:17] <Darke> As it is, that's what the whole autoconf/configure/makefile system is for. Unfortunately that only works well enough for source packages, which is why the ebuild system just puts simple wrappers around them to make things even easier. *grin*
[14:13:25] <Darke> Heh.
[14:14:09] <Darke> On a completely off-topic tangent, this: http://www.business2.com/b2/web/dumbest/1 "The 101 Dumbest Moments in Business" is a fun read. *Grin*
[14:14:19] <Colourless> yeah, i read that a little earlier
[14:16:23] <Colourless> so, what is the reason why different linux distro's are all so slightly incompatible
[14:18:21] <Darke> Umm... because they can be, basicly. Different groups of people started them, which had different ideas of how things 'should' be layed out, and the 'proper' way to do things.
[14:18:55] <Colourless> now, i'm curious what the 'first' linux version was like, and how things have deviated from it
[14:19:59] <Darke> Even closed source unixes are guilty of that. An IRIX system has things located in different places to an AIX system, to a AT&T derivitive (SCO's UnixWare), etc. Even MS is guilty of this with their placement of different parameter twiddles in different versions of windows. *grin*
[14:20:34] <Darke> Umm... 'first' could be interesting, and could mean many different things. *grin*
[14:22:47] <Colourless> yes MS doing it is quite obvious moving from Win9x to WinNT. Espectially since directory names changed. WinNT places almost all dlls for example in System32, with a few in System from compatibilty with ancient Win16 program. While Win9x places eveyrthing in System, except for WDM files that go in System32. And then the registry layout is slightly different between Win9x, Win2k and WinXP
[14:23:41] <Colourless> badly written win9x apps of course 'assume' things work the Win9x way and start putting all sort of weird things in the wrong places
[14:25:00] <Darke> Yup. Then there's the fiddling-for-fiddling-sake they do with the gui. Like shuffling around where they put all the nity-grity networking stuff in every new OS release.
[14:25:54] <Colourless> Whistler will be different again :-)
[14:26:17] <Darke> Yup.
[14:29:14] <Darke> In the end, there's really no difference between swapping between linux distros and between windows versions. The problem with linux is in the automated tools, there's little in the way of standard installer converters, like installshield under windows that automagicly sort things out for your minor varient. *grin*
[14:31:22] <Colourless> that i believe is one of the reasons why MS created the WindowsInstaller so (they wouldn't have to pay liscencing fees for InstallShield for MSVC) devs who couldn't afford InstallShield could still have a properly functioning installer that would install things in the right places
[14:32:38] <Colourless> InstallShield for MSVC is the 'lite' version that came with MSVC6
[14:35:23] <Darke> Makes sense really.
[14:40:29] <Colourless> it would really be nice if it was possibly for operating systems on the same platform to have binary compatible applications
[14:40:38] <Colourless> s/possibly/possible/
[14:41:07] <Colourless> something like wine gets there part of the way
[14:43:28] <Colourless> actually wine could probably do most of what i would be thinking (imports from the application can be linked to os dependant libraries)
[14:43:46] <Colourless> compilers and libraries would need to be set up just right though
[14:44:35] <Darke> Yup.
[14:44:47] <Colourless> You'd use something like SDL and a C Library that imports all library functions from a dynamically linked library
[14:47:44] <Colourless> each platform would then have different libraries that contain all the system dependant code
[14:48:28] <Colourless> the PITA thing is of course different executable filetypes
[14:49:12] <Darke> Ick. Yeah.
[14:50:56] <Colourless> that's why we need USECODE ;-)
[14:53:55] * Darke arughs! and runs away screaming!
[14:54:32] <Colourless> could be worse... usecode couldn't have been compiled to native x86 code :-)
[14:56:02] <Darke> Hrm... u8's could have been, u7's probably not. I can kinda spit out C++ with u8's, you'd just need a fancy base class for a few things that just wouldn't work, and you'd have something to complie native. *grin*
[14:56:29] <Colourless> i was thinking of u8's
[14:56:49] <Colourless> only a few of the opcodes don't fit with what an x86 opcode could do
[14:57:03] <Colourless> they would have been translated into function calls
[14:58:13] --> Cahaan has joined #exult
[14:59:03] <Darke> Yup. My logic also. Though I'd just translate to C, then let it do the dirty work for me. *grin*
[14:59:15] <Colourless> IIRC on the webpage one of the u8 programmers (ely perhaps) there was C++ code that would allow for process style classes that supported breaking
[15:00:10] <Darke> Hrm? I don't think I've seen that. Might have been Zack's also.
[15:00:27] <Colourless> it save instruction pointers and all sorts of lovely things :-)
[15:04:06] --> Dark-Star has joined #exult
[15:05:18] <Darke> Hrm... eeeeevil!
[15:23:39] <Colourless> hmm, elys 'stuff' is gone
[15:25:14] <Darke> Bit annoying. Was it just in a zip file or something? Or an actual webpage.
[15:25:39] <Colourless> on the webpage
[15:25:44] <Colourless> IIRC
[15:26:17] <Colourless> i've still got some of his stuff
[15:26:40] <Colourless> such as the RPGItemSource which looks like almost carbon copies of the U8 item code
[15:26:56] <Colourless> you know things like this:
[15:26:56] <Colourless> // Make sure we are valid and in the void
[15:26:56] <Colourless> if( IsValid() && GetFlag( IFLAG_ETHEREAL ) )
[15:26:56] <Colourless> {
[15:27:13] <Darke> Ah, yeah. I think I've got that too.
[15:28:07] <Colourless> void Pop( Item _container ); // Pop into the _container item. If _container is a container
[15:28:14] <Colourless> // then this object will be placed into its container list.
[15:28:16] <Colourless> // If _container is a COLLECTABLE item and this item is the same
[15:28:18] <Colourless> // type, they will be merged adding the Q.quantity field.
[15:28:20] <Colourless> // This item class will then be set to the object that was merged
[15:28:21] <Colourless> // with.
[15:28:27] <Colourless> we don't do that in Pentagram
[15:28:40] <Darke> Gee... y'know, that doesn't look *anything* like it could have come from u8's source, really! *grin*
[15:29:11] <Colourless> actually it doesn't :-)
[15:29:23] <Colourless> the item data is actually stored in the class itself
[15:29:34] <Colourless> rather than u8's magical system of a struct of arrays
[15:29:43] <Darke> *snicker*
[15:38:38] <Colourless> ok, there is a reason why i think jason ely's stuff is now gone. The stuff was on the dransik site. However it looks like Asylemsoft was acquired by TKO so Ely's personal stuff is now gone
[15:40:03] <Colourless> or something along those lines anyway
[15:40:20] <Darke> Ah.
[15:42:30] <Colourless> http://www.santacruzsentinel.com/archive/2003/December/17/biz/stories/01biz.htm
[15:46:22] <Darke> Y'know, the whole computer game/mmorg/role-playing thing seems so surreal when translated into bland biz-speak. *grin*
[15:47:08] <Colourless> i find it interesting that the metioned online nick's in the article
[15:47:29] <Colourless> s/nick's/nicks/
[15:48:07] <Darke> I've often seen nicknames of people mentioned in news articles, so it really wouldn't be much difference.
[15:50:41] <Colourless> i guess half of it is all 'filler' making news out of something that really isn't all that special
[15:50:51] <Darke> Pretty much.
[15:52:15] <Colourless> the head line pretty much sums it all up for most people
[15:53:55] <Colourless> overall, fairly standard journalism.
[15:54:59] <Darke> Yup. Definately a niche interest news item.
[16:04:32] <-- Kirben has left IRC ("System Meltdown")
[16:48:29] --> Fingolfin has joined #exult
[16:48:29] --- ChanServ gives channel operator status to Fingolfin
[16:49:13] <Fingolfin> yo
[16:49:18] <Colourless> hi
[16:49:26] <Darke> Greetings.
[16:55:34] <wjp> hi
[17:47:49] <-- Darke has left IRC (Read error: 110 (Connection timed out))
[17:55:28] --> Darke has joined #exult
[19:51:39] <-- Cahaan has left IRC (Read error: 60 (Operation timed out))
[20:42:49] <Fingolfin> hmm
[20:43:04] <Fingolfin> "make exult" doesn't work. "make" does. That's a bit odd, isn't it?
[20:43:25] <wjp> blame it on automake :-)
[20:43:38] <wjp> make -j 5 doesn't work either on a 4 cpu machine
[20:43:45] <Fingolfin> hm
[20:43:49] <Fingolfin> tsk tsk =)
[20:44:03] <wjp> (the make -j 5 issue is a inter-directory dependency issue)
[20:44:22] <Fingolfin> guessed as much :-)
[20:44:31] <Fingolfin> yet another reason for a "flat" make system
[20:44:42] <Fingolfin> besides being much faster to do a "make" in an up-to-date dir
[20:44:48] <wjp> indeed
[20:45:01] <wjp> is the 'make exult' issue a directory issue too?
[20:45:11] <wjp> I would guess the flex headers?
[20:47:00] <Fingolfin> exactly :-)
[20:47:33] <wjp> automake is just so broken. Makes you wonder if anybody ever actually thought about it
[20:48:38] <wjp> (somebody with the required theoretical background, that is)
[20:51:26] <Fingolfin> well
[20:51:28] <Fingolfin> in its defense
[20:51:38] <Fingolfin> it simply grew far beyond what it was meant for =)
[20:52:02] <wjp> yeah, I guess :-)
[20:52:03] <Fingolfin> There are actually several modern alternatives to it. E.g. cook, or ant, and many others..
[20:52:30] <Fingolfin> the great thing about automake is, though, that you only have to have make on your system to use an automake package
[20:52:41] <Fingolfin> while for most others (ant, cook, ...) you need that tool
[20:52:51] <Fingolfin> which makes it hard for people to adopt it
[20:53:01] <Fingolfin> at least for people who want to reach a broad audience
[20:53:59] <wjp> yeah
[20:54:22] <wjp> gotta love the tenacity of that guy on #scummvm :-)
[20:55:52] <wjp> lol
[20:55:56] <wjp> "what is source code"? :-)
[20:59:22] <wjp> rofl
[20:59:24] <Fingolfin> <sigh>
[21:21:04] --> Dominus has joined #exult
[21:21:07] --- ChanServ gives channel operator status to Dominus
[21:21:33] <Dominus> hey ho
[21:31:43] <-- Colourless has left IRC ("casts invisibility")
[21:34:26] <wjp> hi
[21:43:13] <wjp> hm, freenode server upgrade soon
[22:03:20] <wjp> Dominus: btw, feel like having a 'documentation update session' sometime soonish before 1.2?
[22:03:46] <wjp> (reading through all of the docs with a couple of people and updating any outdated sections)
[22:03:53] <Dominus> sure
[22:03:59] <wjp> too much docs to do it alone :-)
[22:04:19] <Dominus> I read them so often I do miss obvious stuff now
[22:04:30] <Dominus> like all the things you corrected yesterday
[22:05:47] <Dominus> or those references in the Studio docs to the time we had the two branches snapshots
[22:06:49] <wjp> you mean the installation section?
[22:07:37] <Dominus> yep
[22:07:45] <wjp> what's wrong with that?
[22:08:24] <Dominus> file:///E:/User/Exult/xml/exult_studio.html#run_steps
[22:08:36] <Dominus> oops, wrong url :-)
[22:08:39] <wjp> hehe :-)
[22:08:43] <wjp> I don't think that one'll work here :-)
[22:08:43] <Dominus> in 2.3. Step-by-step instructions to run Exult Studio
[22:08:59] <Dominus> nsure that it says 1.10 CVS during Exult's start-up screen.
[22:09:07] <wjp> 1.10cvs? we never had that
[22:09:31] <Dominus> It could be that the first 1.1.x branches looked like that
[22:09:47] <wjp> I hope not :-)
[22:10:33] <Dominus> anyway I need to make that more generic, the CVS part should be left in along with an explanation why only CVS and not release
[22:11:43] <Dominus> Vex' problem in the forum is very likely from playing with an official release
[22:12:04] <wjp> which is why I asked for the top of his stdout
[22:12:22] <wjp> ah, he just posted that right now :-)
[22:12:30] <wjp> yes, official release
[22:12:32] <wjp> no exult studio support
[22:13:45] <Dominus> A warning gump in exult would be nice that when you "toggle_map_editor" and Exult is not compiled with Studio support
[22:14:11] <Dominus> funny, I had just looked at the post again, probably moments before he posted that
[22:14:32] <wjp> yes, me too :-)
[22:14:54] <wjp> (or rather there wasn't a "new" on the front page, but when I clicked on the topic anyway there was a new post)
[22:26:24] <Dominus> I have to go again. Something I really should do with the docs/faq is to add more pictures.... People need to see what I'm talking about :-)
[22:26:33] <Dominus> well, I'm off
[22:26:36] <Dominus> see you
[22:26:51] <-- Dominus has left IRC ("a pooka invited me to Charlie's")
[23:27:56] <wjp> Fingolfin: thanks :-)
[23:28:08] <wjp> (you can remove the 1.1beta3 from the older releases, btw)