#pentagram@irc.freenode.net logs for 21 Mar 2005 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage

[00:13:50] --> Kirben has joined #pentagram
[00:13:50] --- ChanServ gives channel operator status to Kirben
[02:51:14] --> Ember has joined #pentagram
[03:09:11] <-- Darke has left IRC (Read error: 110 (Connection timed out))
[06:07:06] <-- Lightkey has left IRC (Read error: 60 (Operation timed out))
[06:14:46] --> Lightkey has joined #Pentagram
[07:01:54] --- Ember is now known as Darke
[07:21:06] --> Colourless has joined #Pentagram
[07:21:06] --- ChanServ gives channel operator status to Colourless
[08:59:25] --- Colourless is now known as Cless|Away
[12:37:14] --- Cless|Away is now known as Colourless
[12:37:15] <Colourless> hmm
[12:37:26] <Colourless> we now have 'some' speech
[13:05:06] <-- Kirben has left IRC ("System Meltdown")
[14:40:05] --> trin has joined #pentagram
[15:55:01] <-- trin has left IRC (orwell.freenode.net irc.freenode.net)
[16:03:46] --> trin has joined #pentagram
[16:06:11] --> Chetic has joined #pentagram
[17:25:32] <wjp> hmm... speech; nice :-)
[17:27:42] <wjp> say, how far had our plans to create a pentagram binary that doesn't need a data dir advanced?
[17:31:12] <Colourless> well, for windows it's reasonably trivial to do it
[17:31:29] <Colourless> i should still be able to build pentagram with the data embedded
[17:31:52] <wjp> if we use a small data->header conversion tool it'll be the same on all platforms
[17:31:56] <Colourless> yeah
[17:32:58] <wjp> need to rework filesystem's support of it a bit, I think; I'd probably prefer if 'real' data files would override the static versions in the binary
[17:33:20] <Colourless> yeah might be an idea
[17:33:38] <Colourless> the code i've put in there at the moment
[17:33:55] <wjp> let's see... what's the easiest to reach place with speech? :-)
[17:34:05] <Colourless> the gate into tenebrae
[17:34:12] <wjp> ah, Guardian :-)
[17:34:30] <wjp> works :-)
[17:34:42] <Colourless> continuing what i was saying before
[17:35:45] <Colourless> as it currently is, filesystem will use the file in memory, and ignore what's on the disk
[17:36:21] <wjp> yes
[17:40:02] <wjp> I'm getting a segfault in ~GameData when deleting the SpeechFlexes
[17:40:34] <Colourless> can you be more specific
[17:41:07] <wjp> not yet :-)
[17:41:24] <wjp> 0x0807d3c4 in ~GameData (this=0x8356a38) at games/GameData.cpp:90
[17:41:24] <wjp> 90 if (s) delete [] *s;
[17:41:44] <Colourless> uh
[17:41:51] <Colourless> should be delete *s;
[17:42:17] <Colourless> no idea why i used the array delete
[17:43:04] <wjp> why the double pointer?
[17:43:28] <Colourless> so i don't access the filesystem every time a bark is done
[17:43:45] <Colourless> if the file doesn't exist, the pointer is set to 0
[17:43:53] <wjp> ah
[17:44:54] <wjp> we could consider caching all available speech at startup
[17:45:02] <Colourless> yeah we could
[17:45:02] <wjp> maybe hash all strings as well or somethin
[17:45:04] <wjp> g
[17:45:20] <wjp> bbl, dinner
[17:48:39] <Colourless> one interesting problem
[17:49:13] <Colourless> there can be multiple voice samples per 'barked' string.
[17:50:02] <Colourless> each sample contains only part of the entire 'bark'
[17:50:19] <Colourless> so that complicates things
[17:52:41] <Colourless> currently do handle that at all so not all the speech plays
[17:52:47] <Colourless> in particular stratos
[18:03:50] <wjp> hm, how does that splitting work exactly?
[18:04:16] <wjp> if Stratos says AB there's one speech for A and one for B?
[18:06:16] <-- Lightkey has left IRC (Read error: 60 (Operation timed out))
[18:06:46] <Colourless> yes
[18:07:07] <wjp> hm, so we'd have to do some form of prefix matching?
[18:07:11] <Colourless> yeah
[18:07:35] <Colourless> it's probably going to be all moved into a new types of 'bark gump' or something
[18:07:56] <Colourless> probably a SpeechBarkGump or something
[18:08:15] <wjp> did the original hide the 'subtitles' for spoken text? I seem to remember it did
[18:08:16] <Colourless> main reason is for better timing
[18:08:21] <Colourless> yes it did
[18:08:30] <wjp> I vote for a toggle :-)
[18:08:39] <Colourless> i want to put in option. No Speech, Only Speech, Both
[18:09:03] * wjp nods
[18:13:36] * trin nods
[18:14:07] <wjp> hm, interesting; one of the "its/it's" typos has been duplicated in Stratos' speech
[18:14:25] <wjp> something to watch out for should we ever patch usecode :-)
[18:14:49] --> Lightkey has joined #Pentagram
[18:31:19] <wjp> I've started working on the credits/quotes screens a bit, by the way
[18:34:59] <Colourless> yay
[18:40:41] <wjp> messy stuff...
[18:41:13] <wjp> I started out with the assumption that the credits and quotes were the same but just with a different data file; I'm not so sure anymore now :-)
[18:41:24] <Colourless> they should be the same
[18:41:40] <Colourless> just run the ol' decoded on them and compare
[18:41:42] <wjp> well, default font colours and default alignment might be different
[18:41:56] <wjp> but I'll figure it out :-)
[18:47:20] <wjp> hm, I think I have an interpretation of the symbols now that'll work
[18:47:44] <wjp> but I have to go; bye
[18:48:46] <Colourless> bye
[20:50:45] <-- Colourless has left IRC ("casts improved invisibility")
[21:06:15] <-- oaQ^ has left IRC ("Reconnecting")
[21:06:16] --> oaQ^ has joined #pentagram
[22:28:37] --> Chetic_ has joined #pentagram
[22:30:13] <-- Chetic has left IRC (orwell.freenode.net irc.freenode.net)
[22:30:13] <-- trin has left IRC (orwell.freenode.net irc.freenode.net)
[22:31:55] --> trin has joined #pentagram
[23:05:15] --> Kirben has joined #pentagram
[23:05:15] --- ChanServ gives channel operator status to Kirben