#pentagram@irc.freenode.net logs for 13 Apr 2004 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage

[00:01:31] <-- sbx has left IRC ()
[01:37:08] --> Kirben has joined #pentagram
[01:37:08] --- ChanServ gives channel operator status to Kirben
[11:55:47] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[11:55:53] --> Kirben2 has joined #pentagram
[12:11:20] --> sbx has joined #pentagram
[12:28:15] <-- Kirben2 has left IRC (Read error: 104 (Connection reset by peer))
[12:28:21] --> Kirben has joined #pentagram
[12:28:21] --- ChanServ gives channel operator status to Kirben
[13:27:54] --> Colourless has joined #Pentagram
[13:27:54] --- ChanServ gives channel operator status to Colourless
[13:28:17] <Colourless> hi
[13:46:09] <-- sbx has left IRC (Read error: 113 (No route to host))
[14:46:02] --> exultbot has joined #pentagram
[14:46:02] --- Topic for #pentagram is: Spoon!
[14:46:02] --- Topic for #pentagram set by Darke at Sun Feb 15 13:15:28 2004
[17:39:52] <wjp> so, we've still got two 'core rewrite' projects that we should tackle sometime... :-)
[17:39:59] <wjp> config, filesystem
[17:40:20] <Colourless> :-)
[17:40:30] <wjp> Colourless: your 'flexes instead of directories' idea would work for me
[17:40:45] <Colourless> yeah, filesystem doesn't really need changing much
[17:41:30] <wjp> needs some extra wrapper classes and some merging logic, I guess
[17:41:39] <Colourless> yeah
[17:42:39] <Colourless> i would probably have the flex class itself act as the 'wrapper' class of sorts, with other abstract classes to read from the sources akin to the IDataSource class sort of
[17:42:59] <wjp> might as well
[17:43:12] <wjp> Flex is probably supposed to indicate something flexible anyway :-)
[17:43:36] <Colourless> flex stood for Flexi file
[17:43:48] <wjp> ?
[17:43:58] <Colourless> which probably does mean flexible
[17:44:18] <Colourless> somewhere i saw something that stated that Flx actually stood for Flexi file
[17:45:05] <wjp> hm, flexi doesn't mean anything to me other than being a prefix of flexible
[17:47:55] <Colourless> me neither
[17:49:26] <wjp> maybe there's a local filing cabinet brand named 'flexi' in Austin or something :-)
[17:50:19] <Colourless> :-)
[17:51:18] <Colourless> fixing up the config stuff is probably more urget then flex stuff.
[17:51:26] <wjp> yes
[17:51:41] <wjp> flex stuff is more for when we want to add our own data
[17:51:45] <Colourless> yeah
[17:51:55] <wjp> i18n is also an issue to consider
[17:52:16] <wjp> I already ran into the first hardcoded strings in u8 when doing the stats gump
[17:52:32] <Colourless> there are a few hardcoded strings.
[17:52:59] <Colourless> other things they have multiple shapes (in u8gumps.flx)
[17:53:31] <Colourless> having a nice robust i18n system would be good
[17:53:49] <wjp> ideally it would support both strings and shape/frame numbers
[17:54:15] <Colourless> yeah
[17:56:39] <Colourless> ideally, i would rather we didn't use indices from a header file to access the strings and shapes
[17:57:07] <wjp> might be worth looking at how exactly gnu gettext does things
[17:57:17] <Colourless> yeah i was thinking that for text
[17:58:07] <Colourless> for shapes we could put the functionality into the flex class
[17:58:33] <Colourless> or (Gumps)ShapeFlex class
[17:58:33] <wjp> use the english one internally and have it aliased to other ones?
[17:58:39] <Colourless> yeah
[17:58:50] <Colourless> helps for savegames too
[17:59:03] <wjp> yes
[17:59:11] <wjp> speaking of saving shapes... we don't do that yet :-)
[17:59:20] <Colourless> but copying savegames between different language version isn't going to be fun
[17:59:28] <Colourless> yeah i know
[17:59:40] <wjp> are usecode adresses different?
[17:59:54] <Colourless> well text string sizes are going to be differet
[18:00:28] <wjp> hm, yes, I can't remember any padding at the end to make strings in all languages the same length
[18:01:23] <Colourless> wrt to saving shape nums for gumps, what i've been attempting to think of is a way of saving an identifier for the flex the shape comes from
[18:01:46] <wjp> guids :-)
[18:01:53] <Colourless> :-)
[18:02:04] <Colourless> that isn't really what i had in mind
[18:02:20] <Colourless> it's hardly better than what we do in exult with an enum for the various flex files
[18:04:06] <Colourless> but in the end it's the most obvious solution
[18:04:19] <Colourless> (using an enum that is)
[18:04:41] <Colourless> it's not like we are going to support dynamic loading of flexes :-)
[18:07:13] <Colourless> pretty easy to have the game data class return a ShapeFlex file based on an index
[18:12:18] <wjp> btw, for translated shapes we may have to refer to another shape
[18:12:32] <wjp> um, another flex
[18:13:14] <wjp> or we could use the 'merging' functionality to have translated shapes appear to be in the same flex, I guess
[18:13:37] <Colourless> not really a problem if we have the support to address a shape with flex, shape, frame
[18:14:28] <wjp> scenario: we want to translate U8 into dutch
[18:14:48] <Colourless> yeah need all new shapes
[18:14:51] <wjp> we'd obviously have to translate the hardcoded strings, which wouldn't be a problem with gettext (or similar)
[18:15:14] <wjp> then there's usecode, which would most likely result in a totally new usecode file
[18:18:18] <wjp> not sure how you'd best handle the shapes
[18:18:30] <Colourless> i have no idea eaither
[18:18:38] <wjp> separate flex? (if so, how would you access it?)
[18:19:02] <wjp> we could always add a special 'l10n' flex to our enum if necessary
[18:19:29] <Colourless> yeah i was thinking that
[18:19:55] <Colourless> might want them 'game dependant' though
[18:20:04] <wjp> yeah
[18:20:22] <Colourless> but that doesn't really matter
[18:20:38] <wjp> as long as it won't cause headaches to implement it later on :-)
[18:20:40] <Colourless> we might want a global l10n flex and then game specific versions
[18:20:54] * wjp nods
[18:21:26] <Colourless> for simplicity sake we could also have the translated strings in the flexes too
[18:24:49] <wjp> sounds like this won't be much of a problem if done this way
[18:24:56] <wjp> so... config :-)
[18:27:25] <Colourless> hmm... yeah
[18:27:42] <Colourless> know of any libraries :-)
[18:28:02] <wjp> there's always scummvm :-)
[18:28:27] <Colourless> :-)
[18:35:17] <Colourless> you know we 'could' (even though i don't think it's a particularly good idea) attempt to pass the text being output by the usecode though the translator
[18:36:11] --> Dark-Star has joined #pentagram
[18:36:47] <Colourless> if anything it would probably make translating easier
[18:37:03] <Colourless> but some of the strings are incredibly long
[18:37:33] <wjp> and it wouldn't work well for concatenated strings
[18:38:20] <Colourless> no, it wouldn't work well
[18:39:34] --> Dominus has joined #pentagram
[18:39:46] --- ChanServ gives channel operator status to Dominus
[18:57:47] <-- Colourless has left IRC ("casts invisibility")
[21:01:55] --> sbx has joined #pentagram
[22:06:02] --> Kirben has joined #pentagram
[22:06:02] --- ChanServ gives channel operator status to Kirben
[23:54:26] <-- Dark-Star has left IRC ()