#pentagram@irc.freenode.net logs for 31 Oct 2011 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage


[00:30:36] <-- Dominus1 has left IRC (Quit: Leaving.)
[09:00:13] --> Dominus has joined #pentagram
[09:00:13] --- ChanServ gives channel operator status to Dominus
[09:03:35] <Dominus> thanks Kirben for the commit
[09:07:27] <Dominus> web docs updated
[09:13:13] <Dominus> now how to make keybindings dependent on an ini setting? My idea would be to add a cheater section in GuiApp.cpp (?) and there assign another console command to the cheat commands (e.g. con.AddConsoleCommand("Cheat::heal_key",Cheat::heal)) in a function that checks whether cheats are enabled and in u8bindings.ini assign the shortcut to Cheat::heal_key
[09:13:34] <Dominus> would that work? Done too complicated?
[09:19:53] <Colourless> probably easier to flag commands as cheats
[10:39:22] <Dominus> I don't quite get what you meant, but it made me think of a way less hackish way: change the layout of the *bindings.ini to categorize shortcuts, e.g. [gameplay] [cheats] [debug] and only read those keys when cheats or debug (which I think needs added since most of our non gameplay commands are not even cheats :)) are enabled in pentagram.ini
[10:39:32] <Dominus> no idea how to do that though :)
[10:40:28] <Dominus> so... anyone up for that? I'm sure you all have nothing better to do than coding on Pentagram :)
[10:41:30] <wjp> Dominus: that patch looks roughly ok, but I_true is used for more than use AvatarCanCheat
[10:41:36] <wjp> s/use/just/
[10:41:43] <Dominus> oh
[10:42:05] <Dominus> that is not god then
[10:43:00] <wjp> hm, the section idea sounds like it could work
[10:45:53] <Dominus> yeah, I think that's a clean approach
[10:46:44] <wjp> what's the state of the keybinding gui?
[10:47:14] <Dominus> I wasn't able to assign keys on OS X when I tried it yesterday
[10:47:37] <Dominus> it seems a bit unfinished
[10:52:47] <Dominus> adding sections to the ini might complicate it further
[11:14:35] <Dominus> the keybinding gui is strange. every binding try gives an error, not to mention that pressing ctrl, alt, cmd,shift already closes the gui...
[11:19:11] <Dominus> and I can't figure out how to make a key+modifier work in the bindings. for example Left-Alt+x
[11:24:29] <wjp> it might end up being easier to add a cheat flag to AddConsoleCommand
[11:25:16] <wjp> then "cheat" will be a property of the specific command, not of the binding
[11:37:19] <Dominus> my thinking was to keep the ability to enter the commands in the console and really just disable the shortcut keys
[13:06:08] <Dominus> watt: if you read this, is it possible to assign a key+modifier? for example ctrl+g?
[13:23:33] <wjp> even if you make 'cheat' a property of the specific command, it's still possible to keep running them from the console enabbled
[13:26:07] <Dominus> yes, I kind of figured out that one could still use them. I hate that code is mostly gibberish to me. I would like to do sooooo many things :)
[13:26:14] <-- Kirben has left IRC ()
[13:35:15] <Dominus> I tried to look at adding a hq3x scaler from comparing exult's hq2x and hq3x and then going to pentagram... I can see the similarities but then... :)
[13:38:00] <wjp> hm, I think pentagram does things quite differently
[13:40:29] <Dominus> mostly one could take things over from the hq2x scaler of pentagram but when it goes to #define PIXEL stuff it's way different
[13:41:37] <Dominus> hmm, actually not that much
[13:51:13] <wjp> note that the 2 in ScaleBits at the bottom is the scale factor
[14:17:22] <Dominus> thanks that will help :) I'll see what I can do there this night :)
[14:37:52] --> spupuser2 has joined #pentagram
[14:45:25] <-- spupuser2 has left IRC (Quit: Leaving)
[15:55:55] <Dominus> me vs my son. me trying to bring him to bed for his late afternoon nap...
[15:56:40] <Dominus> I fell almost asleep, moved, he woke up and now I'm all pfff from being half asleep... and he's fit :)
[17:36:53] * Dominus thinks he's gonna dream of sizeof(uintX) tonight...
[18:20:46] <Dominus> hmm, compiling...
[18:45:22] <Dominus> hmmmm
[18:45:46] <Dominus> now that I look at it, there is something wrong with pentagram
[18:46:16] <Dominus> I can *not* see any difference between *ANY* of the scalers, except for the bilinear scaler
[18:46:59] <Dominus> pentagram says it is using them but it seems to fal back to point without giving notice
[18:49:13] --> Dominus1 has joined #pentagram
[18:50:27] <Dominus1> wjp, colourless, could it be that scalers are broken?
[18:51:13] <-- Dominus has left IRC (Ping timeout: 252 seconds)
[18:54:11] <wjp> they work here
[18:55:20] <Dominus1> do you see a difference between point and scale2x, 2xSaI, hq2x? I opened pentagram with point and then opened new instances with a different scaler
[18:55:55] <Dominus1> not seeing any difference between them on OS X and Windows
[18:58:01] <Dominus1> bbl
[18:58:38] <wjp> point and 2xsai are noticably different
[20:23:49] <SugarCube> Make sure your window resolution is bigger than 320x200? :)
[21:10:37] <Dominus1> yeh, my fault for not setting width/height to 2x the scalex/y setting. If it is not exactly 2x the scalers won't work. But I found I really need to set both width/height and scalex/y
[21:10:50] <Dominus1> if I only set one pair, no scaling as well
[21:13:58] <Dominus1> seems only bilinear does that nicely
[21:15:10] <wjp> I'm afraid I kind of forgot the exact rules for those
[21:17:33] <Dominus1> :)
[21:17:55] <Dominus1> of course my modified hq2x scaler that is now supposed to be a hq3x scaler does nothing :)
[21:53:20] <Dominus1> hmm, kind of got it but it's unstable
[22:03:34] <Dominus1> wjp or colourless, could either one take a look at this diff http://pastebin.com/vsvwmB5z ?
[22:03:49] <Dominus1> it's making the hq2x scaler a hq3x one
[22:04:22] <Dominus1> for now I just modified the existing hq2x scaler to not having to get the configure stuff right as well for testing
[22:05:58] <Dominus1> the one problem this has on compiling is that it complains a couple of times about "comparison between signed and unsigned integer expressions"
[22:06:17] <Dominus1> but that is already an issue for the hq2x scaler
[22:52:02] <wjp> Dominus1: hm, it crashed here, but due to a bug that was seemingly already in the hq2x code
[22:53:12] <wjp> in if (j==0 || !clipY_Begin) , replace j==0 by j>0, I think
[22:53:26] <Dominus1> yes, I had some strange crashes there with the hq2x thing as well, during testing when it actually works :)
[22:54:37] <wjp> it works here with that change
[22:55:01] <wjp> and at first glance the patch looks good. Well done :-)
[22:55:32] <wjp> it's awfully slow though
[22:57:09] <Dominus1> hmm, not here :) (but I have fast one)
[22:57:48] <wjp> hm, in theory this is a pretty fast PC (i7 920)
[22:58:16] <Dominus1> hmm, yes that sounds like a fast one
[23:01:38] <Dominus1> I'll see about adding the scaller correctly, tomorrow
[23:01:57] <Dominus1> thanks for looking it over
[23:02:19] <Dominus1> now I'm thinking of adding the hq4x scaler as well :)
[23:05:23] <wjp> I wonder why it's so slow here
[23:06:31] <wjp> I think I'm going to blame these amd/ati drivers for now. I should really switch back to my nvidia card...
[23:06:49] --> Kirben has joined #pentagram
[23:06:49] --- ChanServ gives channel operator status to Kirben
[23:07:08] <Dominus1> do you have any idea on how to "fix' the signed/unsigned warnings?
[23:08:36] <wjp> add (uint32) before the abs() calls on the lines it points out
[23:08:40] <wjp> if you have the same ones I get
[23:09:20] <wjp> or (unsigned int) is probably more correct
[23:15:08] <Dominus1> hmm, how should that look for example in line 269 (with my patch)?
[23:15:13] <Dominus1> ( abs((YUV1 & Vmask) - (YUV2 & Vmask)) > trV ) );
[23:16:51] <Dominus1> phew, need to go to sleep...
[23:16:52] <wjp> ( (unsigned int)abs(...
[23:17:34] <Dominus1> hmm, gonna try that tomoorw, right now it didn't give me a good result
[23:17:45] <Dominus1> thanks again... and good night
[23:17:49] <wjp> good night
[23:18:04] <-- Dominus1 has left IRC (Quit: Leaving.)
[23:27:56] <wjp> I committed that j==0/j>0 thing, by the way