#uwadv@irc.freenode.net logs for 20 Oct 2002 (GMT)

Archive Today Yesterday Tomorrow
Underworld Adventures homepage


[01:04:57] <-- QQtis has left IRC (Read error: 104 (Connection reset by peer))
[17:37:21] --> Dark-Star has joined #uwadv
[17:37:30] <-- Dark-Star has left #uwadv ()
[18:31:35] --> phlask has joined #uwadv
[18:31:46] <-- phlask has left #uwadv ()
[19:32:32] --> wjp has joined #uwadv
[19:32:32] --- ChanServ gives channel operator status to wjp
[19:36:17] --> vividos has joined #uwadv
[19:36:17] --- ChanServ gives channel operator status to vividos
[19:36:25] <vividos> good evening
[19:40:05] <wjp> hi
[19:40:14] <wjp> why do you always enter this channel minutes after me? :-)
[19:40:28] <vividos> precognition :)
[19:41:22] <wjp> :-)
[19:41:23] * vividos just recently watched minority report
[19:42:25] <vividos> the new arrangement from QQtis is rather good
[19:43:30] <wjp> ack.. forgot to download that
[19:43:58] <wjp> I've been away most of friday, saturday and today, so I'm starting to forget to do things :-)
[19:44:17] <vividos> :)
[19:45:09] <vividos> did you try out the %uahome% stuff yet?
[19:45:15] <vividos> probably not :)
[19:45:45] <wjp> not really, but I did rebuild it
[19:45:49] <wjp> no compile errors
[19:46:38] <vividos> ok
[19:48:24] <vividos> somehow I have the feeling that interest in such projects as exult or uwadv has ceased over the year ... I mean, the channel logs of #exult were longer in the summer months
[19:49:12] <wjp> hm, could be. I know I have less time for them now than in the summer
[19:50:02] <vividos> I more meant the fans and supporters of the projects
[19:50:09] <wjp> ah
[19:50:18] <vividos> don't know about exult's forum, though
[19:50:32] <wjp> might be interesting to get some statistics
[19:50:56] <wjp> pity that the current forum DB doesn't go back all that far
[19:51:55] <vividos> why was it cleared? a new phorum version?
[19:53:23] <wjp> um.. well.. there was an.. 'accident' :-)
[19:54:04] <wjp> small SF glitch coupled with small phorum glitch resulted in phorum eating our DB
[19:55:29] <wjp> I'm making automated backups of it bi-daily now :-)
[19:57:41] <vividos> of the database?
[19:58:36] <wjp> yeah
[20:01:00] * vividos wonders if uwadv would also need a forum
[20:02:10] <wjp> I'd wait until it's playable a bit
[20:03:47] <vividos> ok
[20:04:09] <vividos> I wondered if it would be possible to do a search engine for the channel logs
[20:04:21] * wjp hmms
[20:04:33] <wjp> does google index them yet? :-)
[20:04:39] <vividos> #exult is rather large by now
[20:04:47] <vividos> parts of it
[20:05:10] <wjp> I might do it sometime
[20:05:21] <vividos> http://www.google.com/search?q=+site:www.math.leidenuniv.nl+%23exult+pentagram
[20:05:49] <wjp> just pass the search terms on to grep, and produce links from the output, or something like that
[20:06:14] <vividos> would work, I guess
[20:06:37] <wjp> anyway, time to check that %uahome%
[20:06:44] <wjp> what do I do to check that? :-)
[20:07:38] <wjp> hm, compile error
[20:07:55] <wjp> prototype for `void ua_critter::load (const char *, unsigned char)' does not match any in class `ua_critter'
[20:07:59] <vividos> just look at the config files that are parsed (and the uahome == ~/.uwadv/ is mkdir'd
[20:08:00] <vividos> hmm
[20:09:19] <wjp> maybe you forgot to commit critter.cpp?
[20:09:39] <vividos> that was it :)
[20:10:51] <vividos> btw, that critter loading is strange. using msvc, in debug mode, it needs 48 seconds (!) to load all images, and in release mode (with inlining and such) it only needs 4 seconds
[20:11:11] <wjp> 48 seconds?! painful
[20:11:24] <vividos> and the culprit for that is std::vector<ua_image>::push_back()
[20:11:39] <wjp> hm, did you change the config file list?
[20:11:45] <vividos> yes
[20:12:13] <vividos> the %uahome%/uwadv.cfg is now loaded
[20:12:25] <wjp> and ./uwadv.cfg isn't anymore, it seems :-)
[20:13:23] <wjp> one minor thing: %uahome%/savegame is translated to /home/wjp/.uwadv//savegame
[20:13:24] <vividos> should I add it again?
[20:13:27] <wjp> (note the double //)
[20:13:29] <wjp> no, it's ok
[20:13:44] <vividos> the uwadv.cfg.in should have the second / removed
[20:14:06] <wjp> ?
[20:14:41] <vividos> well, the setting in uwadv.cfg.in is:
[20:14:42] <vividos> savegame-folder %uahome%uasave/
[20:14:43] <wjp> hm, I don't think %uahome%savegame is a good idea
[20:15:12] <vividos> in uwadv I assume that all path names end with a slash or backslash
[20:15:51] <wjp> yeah, but %uahome%savegame looks really weird
[20:15:54] <vividos> well, we could add the slash again, or check for a slash after an expanded keyword
[20:16:28] <vividos> the /home/wjp/.uwadv//savegame should work, too
[20:16:31] <wjp> true
[20:19:06] <vividos> ngrr, just forgot about the 48 seconds ...
[20:19:44] <vividos> (critloader.cpp is committed, btw)
[20:19:51] <wjp> I know; it compiles now
[20:20:21] <wjp> hm, when you push_back() to a vector, did you resize the vector first to ensure it has enough capacity for all images?
[20:20:32] <vividos> yes, but that didn't help much
[20:20:59] <wjp> you push_back images or image pointers?
[20:21:08] <wjp> s/you/do you/
[20:21:13] <vividos> images
[20:21:28] <vividos> they're stored in std::vector<Uint8> members
[20:21:39] <vividos> and the images aren't that large
[20:22:08] <wjp> when exactly does it load these things, btw?
[20:22:27] <vividos> when calling init_core()
[20:22:31] * wjp doesn't really notice any slowdown
[20:22:52] <vividos> the console should print the time it needed to load the critters
[20:22:54] <wjp> or maybe half a second after "using SDL_mixer version 1.2.4"
[20:22:59] <wjp> ah, right
[20:22:59] <vividos> yes
[20:23:05] <wjp> hm, that was 1382ms?
[20:23:09] <vividos> for all images?
[20:23:11] <wjp> seemed much less
[20:23:17] <wjp> loading all critter animations ... done, needed 1382 ms
[20:23:37] <vividos> wow :)
[20:23:49] <vividos> but that only depends on cpu speed
[20:25:39] <wjp> strangely using -O2 doesn't seem to speed it up
[20:26:20] <vividos> the copy ctor of ua_image already seems to be built with max. speed
[20:26:53] <wjp> oh, in which source file is that?
[20:27:00] <wjp> I just rebuilt critloader with -O2
[20:27:10] <wjp> image.cpp I guess :-)
[20:27:15] <vividos> it's auto-created by the compiler
[20:27:48] <wjp> right, 894ms now
[20:28:01] <vividos> with -O2?
[20:28:05] <wjp> yeah
[20:28:15] <wjp> (critloader.o and image.o with -O2, the rest without)
[20:28:36] <vividos> we should recommend your machine specs for playing uwadv :)
[20:29:03] <wjp> athlon xp 1800, btw
[20:29:09] <vividos> I know
[20:29:23] <wjp> you do? I guess I mentioned that sometime before, then :-)
[20:29:26] <vividos> yes
[20:29:53] <vividos> http://www.salikon.dk/quillframe.html
[20:30:08] <vividos> playing Nim with Dr. Cat is nice :)
[20:30:57] <wjp> who takes the last bead wins? hmm...
[20:31:25] <vividos> the game is from ultima 5 and Dr. Cat (from Paws) always started, taking 2 beads
[20:31:32] <vividos> guess who won all the time
[20:31:48] <wjp> strategy isn't too hard :-)
[20:31:57] <wjp> just get it 0 mod 4, and keep it that way :-)
[20:33:01] <vividos> oh, it was ultima 6, I guess
[20:34:02] <vividos> hehe, there was a u7 book called "ONE HUNDRED AND ONE WAYS TO CHEAT AT NIM"
[20:34:24] <wjp> heh :-)
[20:34:57] <wjp> there was also a book about usecode, I think :-)
[20:35:07] <vividos> http://home.hiwaay.net/~rgregg/ultima/literal/Books_U7.html
[20:35:10] <vividos> cool :)
[20:35:31] <vividos> but the best is "hubert's hair-rising adventure" :)
[20:35:36] <wjp> :-)
[20:35:53] <vividos> lol: EVERYTHING AN AVATAR SHOULD KNOW ABOUT SEX: (all pages in this book are blank!)
[20:36:04] <wjp> yes :-)
[20:36:55] <vividos> the avatar probably already knows all about that topic *g*
[20:37:45] <vividos> phil's guide to unfinished usecode: An enlightening discourse on the enigma of usecode. Beginning with the heretofore unresolved mysteries of global variables...
[20:37:46] <wjp> "Recent experiments have proven my theory that these rays, known hereafter as "Mempto Rays", are lethal to all non-living matter. In fact, Mempto rays have demonstrated their ability numerous times, once killing an entire boulder in a matter of a few hours."
[20:38:35] <vividos> :)
[20:39:23] <vividos> u7 was much more fun for the devs than uw1, it seems
[20:40:07] <wjp> there should also be a book in one of these games explaining why you should never build a house with doors in only the north and west walls :-)
[20:40:22] <vividos> yes :)
[20:42:12] <vividos> btw, rest of Quill Dragon's site is also interesting
[20:51:39] <vividos> yesterday was a chat session on #twistedlabs (the other uw1 remake people) and it seems nobody was there :)
[20:52:19] <wjp> ouch :-)
[20:52:35] <vividos> (at least from the /who query)
[20:52:46] <vividos> http://fabrice.bellard.free.fr/TinyGL/
[20:52:59] <vividos> seems that we *could* see uwadv on handhelds
[20:53:43] <wjp> interesting
[20:57:23] <vividos> it's at least more portable than that one: http://www.dsbox.com/minigl.html
[21:08:57] <wjp> that page hasn't been updated for ages
[21:12:32] <vividos> I still wonder if there are OpenGL implementations for handhelds
[21:13:31] <vividos> well, when someone donates a zaurus, I probably consider porting it using TinyGL :)
[21:17:28] <vividos> ah, found another: http://pierrel5.free.fr/ThemePGL.htm
[21:24:19] <wjp> one of the sysops at my uni recently got a Zaurus
[21:24:35] <vividos> talking about critloader.o and using -O2 again ... it didn't help to just compile that one with -O2?
[21:24:43] <wjp> no
[21:24:52] <vividos> ok
[21:25:31] * vividos only knows that a Zaurus is some sort of pda with linux on it, right?
[21:26:49] <vividos> there's even a zaurus.de page :)
[21:30:52] <vividos> only 399 euro, pretty cheap ;)
[21:32:14] <vividos> what footprint does current exult have?
[21:33:03] <vividos> or better: what footprint does uwadv have?
[21:38:22] <wjp> pretty big
[21:38:47] <wjp> 97Mb ingame, it seems
[21:38:59] <vividos> for exult?
[21:39:03] <wjp> uwadv
[21:39:24] <wjp> exult is around 10Mb on startup
[21:39:29] <wjp> (start of game, I mean)
[21:39:41] <wjp> but it grows when you visit more of the map
[21:40:18] <vividos> wintop says (debug version using msvc): 10 MB code, 23 MB data
[21:40:45] <vividos> why does it grow?
[21:40:53] <wjp> it doesn't load all map data at the start
[21:40:57] <vividos> ah ok
[21:41:28] <wjp> there are some plans to swap out map data again when you're far away from a location
[21:41:34] * vividos starts a mingw32 version of uwadv
[21:42:33] <vividos> huh, this one has 30 MB data
[21:43:28] <wjp> top gives a size of 90Mb here, with a RSS of 22Mb
[21:43:40] <vividos> RSS?
[21:43:45] <wjp> resident set size
[21:44:00] <wjp> the amount of virtual memory actually in physical memory
[21:44:09] <vividos> ah :)
[21:44:10] <wjp> um, no, that's not right
[21:44:32] * wjp thinks
[21:45:07] <wjp> I'm not using any swap currently, and still RSS < total size
[21:45:39] <vividos> maybe it swaps out unused pages
[21:47:23] <wjp> it may be the amount of memory that is actually being actively used, or something
[21:47:29] <wjp> (where 'actively' is open for interpretation)
[21:58:17] <vividos> hmm, cannot enable optimization in debug builds in msvc :/
[22:01:09] <vividos> at least inlining optimization cut down critters loading time to 38 secs
[22:02:11] <wjp> how can it possibly take that long?
[22:02:49] <wjp> I don't think I've ever seen a factor 10 between debug and release builds
[22:07:08] <wjp> I should go catch up on some sleep
[22:07:15] <wjp> g'night
[22:07:30] <-- wjp has left IRC ("Zzzz...")
[22:29:03] <vividos> bye
[22:35:42] <-- vividos has left IRC ("Leaving")