#pentagram@irc.freenode.net logs for 10 Dec 2003 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage


[00:12:50] <-- wjp has left IRC ("Zzzz....")
[01:02:26] <-- Fingolfin has left IRC ("42")
[01:50:44] --> Cashman has joined #pentagram
[01:53:36] <Cashman> helllo p[eople
[01:54:04] <Cashman> action action 123
[01:56:07] <Cashman> anyone around/
[01:56:10] <Cashman> anyone around?
[01:59:31] <-- Cashman has left IRC ()
[02:11:14] <watt> I've been thinking more on keybinding config files.. The xml configuration can easily be used. I tring to think if the keys should be "bindings/<some binding>/firstkey" which would give a key to match
[02:11:52] <watt> or... "bindings/<some key>" which would give the binding to match.
[02:13:24] <watt> I'm begining to think "binding/<some key>" would be easier to code and allow N keys to bind to the same action.
[02:14:34] <watt> <F8>QuickLoad</F8> .... hmmm.
[02:15:08] <watt> this would also make all key binding potentially singletons.
[03:34:05] <watt> nah.. just a vector of them.. I thinks almost working
[03:44:10] <watt> say.. configuration keys aren't case insensitive are they?
[03:46:42] <watt> HAHA victory!
[03:55:59] <watt> woah.. avatar doesn't walk after QuickLoading ??
[04:52:58] <watt> patch-o-rama.. I need to learn how to make patches efficiently.
[05:13:10] --> Cashman has joined #pentagram
[05:13:23] <Cashman> I think u gonna net a merry christmas watt, well good luck :-)
[05:13:25] <Cashman> action action
[05:22:39] <watt> oh no.. I some how messed up quick loading! ARG!
[05:23:02] <Cashman> bugger
[05:23:24] <Cashman> I remember wjp had a heck of a time implementing loading/save which I believe is not complete
[05:23:35] <Cashman> there are still problems, maybe oneday
[05:26:13] <watt> well.. the funny part is that my patch doesn't touch loading process.. it simply calls it in a different place. Which I swear was working.
[05:26:51] <Cashman> good luck, do u have admin access or have u just altered ur local cvs on hdd
[05:27:22] <watt> I'm anony... thus the patch making.
[05:28:24] <Cashman> oh u mean make a patch and submit it?? then admin takes a look and has the right to apply it?? I dont know much about cvs
[05:28:40] <Cashman> I just use to download cvs using a cvs engine - script and compile, that was it
[05:29:38] <watt> right.. but since I'm new to making patches (and somewhat new to cvs) it takes a long time to get everything right.
[05:30:55] <Cashman> :-) good luck
[05:31:08] <Cashman> it took me ages to get cvs working.. when I tried well like 6 months ago
[05:33:54] <watt> cvs isn't too difficult.. its all the details that slow me down.. you usually have to worry about permissions when commiting for one
[05:34:50] <watt> I like subversion..
[05:35:11] <watt> everything there just tends to work and be easy.
[05:35:45] <watt> no actual users on the box, just some apache config files.
[05:36:28] <Cashman> nice
[05:36:50] <-- Cashman has left IRC ()
[05:46:55] <watt> oh.. the avatar was just refusing to carry out animations until moved ?!?!?!!?&^*%^& but after the first quickload does this, the rest do not.
[05:48:20] <watt> well.. other than that my patch is working quite well.. I'll post it on sf.. hopefully someone will have some time to look over it a tell me what needs to be changed/completed
[06:13:50] <watt> oh.. my keybinding flaw (in my comments for the patch) could potentially be big.. binding a mover on down and overriding up would be a problem.. but then again the old style bindings will ultimately be removed if the patch is good.
[06:15:20] <watt> it really helps to type out my thoughts :)
[08:27:35] <watt> oh.. someone might want to close the old combat patch.. I had submitted that one as anony (kind of, it didn't have anything attached).
[09:44:18] <watt> sleeping pills goooooooooooooooooooooooooooood..........
[09:44:38] <watt> awake baaaaaaaaad.
[11:36:11] --> Colourless has joined #Pentagram
[11:36:11] --- ChanServ gives channel operator status to Colourless
[11:38:28] <Colourless> hi
[12:18:24] --> Fingolfin has joined #pentagram
[12:18:24] --- ChanServ gives channel operator status to Fingolfin
[12:51:13] <-- Fingolfin has left IRC ("42")
[14:11:14] <-- Kirben has left IRC ("System Meltdown")
[15:27:16] <Colourless> it would have been nice if SDL supported full colour hardware cursors
[15:28:47] <Colourless> then we wouldn't have to worry about having to blit them ourselves
[15:28:55] <Colourless> s/blit/draw/
[15:29:21] <Colourless> and also the framerate of pentagram wouldn't effect the cursor movement speed
[16:49:13] --> wjp has joined #pentagram
[16:49:13] --- ChanServ gives channel operator status to wjp
[16:49:24] <wjp> hi
[16:49:42] <wjp> first of all...
[16:49:53] <wjp> Cashman: save/load code is complete and working properly, AFAICT
[16:50:02] <Colourless> hi
[16:50:14] <wjp> second...
[16:50:23] <wjp> I took a brief look at the input patch
[16:50:42] <wjp> I think it's too java-ish, at first glance
[16:51:39] <wjp> by that I mean that I don't like the fact that every single binding has to be a new class
[16:54:10] <wjp> s/binding/bindable function/, I guess
[16:54:58] <wjp> watt: I wouldn't worry too much about 'old' and 'new' keyhandling systems clashing
[16:55:30] <wjp> the current only exists to let us do 'things' even though there wasn't any real keyhandling yet
[16:55:36] <wjp> s/current/current one/
[16:57:39] <Colourless> current stuff is a hack really
[16:57:56] <wjp> yeah, that's what I was saying (or trying to say), really :-)
[16:58:33] <Colourless> it was originally included so we could quit the program wasn't it?
[16:58:38] <wjp> yes :-)
[16:58:45] <Colourless> and then extra debugging functions got added
[16:58:56] <wjp> and then i/z/etc... to test container gumps
[16:58:57] <Colourless> and then some odd functions that allowed you to move around
[16:59:29] <Colourless> the AvatarMoverProcess -> Grade A Hack there :-)
[16:59:59] <wjp> :-)
[17:00:52] <Colourless> i'm allowed to say that though, I made it :-)
[17:15:41] <wjp> using our dynamic classtype code for this is interesting, though
[17:15:54] <watt> hehe..
[17:16:37] <wjp> search complexity is linear in the number of bindings times the number of bindable functions, but I doubt that will be a problem :-)
[17:16:48] <wjp> s/search/load/
[17:18:51] <Colourless> well, you know, how expensive is our save game loading routine wrt searching for the class loading function ptr?
[17:19:05] <wjp> logarithmic I think
[17:19:27] <Colourless> that's... pretty awful :-)
[17:19:36] <wjp> (logarithmic in the number of object classes; linear in the number of saved objects)
[17:19:59] <wjp> a lot better than the quadratic algo it was at first ;-)
[17:20:17] <Colourless> thing about strcmp, it has early exit, which can mean it's faster than it might appear
[17:20:34] <wjp> I counted strcmp as constant time
[17:22:23] <Colourless> how are things different with the bindable funcs?
[17:22:51] <wjp> it does a linear search through all the bindings instead of using a map :-)
[17:23:27] <wjp> (as I said, not likely to be a problem ;-) )
[17:24:09] <Colourless> i think if you search the logs for the last week, i think i suggested using a map
[17:24:35] <wjp> I'll believe you without searching the logs :-)
[17:26:27] <Colourless> well, it was actually more than a week ago :-)
[17:26:37] <Colourless> closer to 2 weeks ago actually :-)
[17:30:33] <Colourless> Javaish indeed
[17:30:52] <Colourless> i don't think we need so many files for such obviously simple classes
[17:31:07] <wjp> or so many directories
[17:31:40] <Colourless> especially since the directories don't fit the GMP (Grand Master Plan)
[17:32:36] * Colourless attempts to hide in shame
[17:32:55] * Colourless did actually suggest an input directory at some stage according to devel mailing list
[17:33:11] <Colourless> from the email:
[17:33:11] <Colourless> input/ -- For the input handling classes (maybe). This just depends on how we actually handle input. Ideally I think the should only be a single event handling function in the engine which then dispatches the input to the component that currently has focus. Another dir for this may not be needed.
[17:33:20] <wjp> I'm not really opposed to an input directory, although I don't think we really need one
[17:33:41] <Colourless> of course that comment totally ignored bindings
[17:35:10] <Colourless> ok linear lookup time indeed :-)
[17:35:59] <Colourless> of course adding a 'break' into the offending code would make is log
[17:36:03] <Colourless> s/is/it/
[17:36:07] <wjp> no, still linear
[17:36:40] <wjp> anyway, I should go home about now (or about 1 hour ago, actually, but anyway :-) )
[17:36:43] <Colourless> hmm
[17:36:49] <wjp> I'll bbl
[17:37:00] <-- wjp has left IRC ("bbl")
[17:37:06] <Colourless> code could be more efficent for sure
[17:38:35] <Colourless> adding a 'break' wouldn't actually work, since it would end up breaking things
[17:38:52] <Colourless> yes, i punned :-P
[17:40:14] <Colourless> understanding is the key
[17:40:33] <Colourless> the way it works is.... not how i would have done it
[17:41:34] <Colourless> does it matter? Maybe, maybe not.
[17:44:54] <Colourless> looking that the binding cfg file traversing the XML tree would be ideal at cutting down the time it takes to parse the file
[17:45:32] <Colourless> which is supported
[17:47:24] <Colourless> Configuration::getSubkeys() being the func to use
[17:52:29] --> wjp has joined #pentagram
[17:52:30] --- ChanServ gives channel operator status to wjp
[17:52:36] <watt> as for the linear lookup.. I considered using a logarithmic lookup instead.. but then I thought it wasn't really necessary since that function is only used once at startup and then never again
[17:52:57] <wjp> back
[17:54:42] <Colourless> wb
[17:57:13] <watt> And yes very Javaish.. Java is the only language I've had professional experience with. Thus all of my code tends to lean toward that thought mode.
[17:57:44] <watt> And to me, it's more a pattern thing than a java thing.
[17:59:07] <watt> it tends to give extra little benifits.. using the class_name for binding was one of them
[18:07:01] <Colourless> I'm warning you watt. No one, i repeat no one will take my crown as worst typer and speller
[18:07:15] <wjp> ;-)
[18:17:01] * watt hides in fear.
[18:22:43] <watt> well.. another exciting day of classes.. or.. run away! run away! run away!
[20:48:36] --> ragzter has joined #pentagram
[22:07:31] <-- Colourless has left IRC ("casts invisibility")
[23:22:03] --> Knight has joined #pentagram
[23:22:21] <Knight> lol
[23:22:21] <Knight> nice topic
[23:24:00] <Knight> hi?
[23:24:16] <wjp> hi :-)
[23:25:13] <Knight> how r ya
[23:26:05] <wjp> I'm ok
[23:26:55] <Knight> hope you are not a bot..
[23:26:56] <Knight> :p
[23:27:04] <wjp> how rude :-)
[23:27:15] <Knight> ;)
[23:33:05] --> Kirben has joined #pentagram
[23:33:05] --- ChanServ gives channel operator status to Kirben
[23:33:58] <Knight> i didn't know this channel was on this server..
[23:33:58] <Knight> lol
[23:34:07] <Knight> if not i would have enter sometime before
[23:36:21] <Knight> alone again..
[23:36:22] <Knight> :(
[23:36:30] <Knight> sniff..
[23:36:50] <Knight> the log of your page lies.. it always shows so much talk
[23:37:03] <Knight> lol
[23:38:55] <wjp> sorry, I'm a bit busy with something else currently :-)
[23:41:00] <Knight> oh :(
[23:41:01] <Knight> mmm i really like this pentagram thing
[23:41:01] <Knight> real fan of u8 here
[23:41:01] <Knight> :d
[23:41:01] <Knight> :D
[23:48:47] <wjp> did you already try pentagram or are you waiting for it to mature a bit more?