#gemrb@irc.freenode.net logs for 23 Feb 2015 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:23:11] <-- edheldil_ has left IRC (Ping timeout: 256 seconds)
[00:58:09] <-- turtleman_ has left IRC (Ping timeout: 252 seconds)
[02:22:22] --> raevol has joined #gemrb
[06:22:28] <-- brada has left IRC (Quit: brada)
[06:48:31] <-- Lightkey has left IRC (Ping timeout: 250 seconds)
[07:01:26] --> Lightkey has joined #gemrb
[08:19:14] --> Eli2_ has joined #gemrb
[08:22:31] <-- Eli2 has left IRC (Ping timeout: 265 seconds)
[08:58:42] --> fireflame has joined #gemrb
[09:15:38] <-- raevol has left IRC (Quit: Leaving)
[12:23:17] <wjp> GOG has 50%/50%/25% discounts on BG/BG2/IWD enhanced today
[12:31:11] --> Mechanimal has joined #gemrb
[12:33:47] <lynxlynxlynx> too much work even without that
[12:53:52] <-- fireflame has left IRC (Ping timeout: 246 seconds)
[13:18:19] --> turtleman_ has joined #gemrb
[14:06:53] <-- turtleman_ has left IRC (Ping timeout: 240 seconds)
[14:28:02] --> turtleman_ has joined #gemrb
[14:48:18] --> fireflame has joined #gemrb
[14:48:24] <fireflame> hello
[14:48:58] <fireflame> i know this is not your specialty but i know many people here have some knowledge with old games, so i thought maybe you could help me
[14:49:48] <fireflame> i bought recently an old game called desperado. It has a known bug that makes the whole computer crashes, disabling the ctrl alt supp function, and forcing to make whole computer restart.
[14:50:51] <fireflame> I did some research, and found out this bug was related to the sound files. To make short, the solution is to uncompress all the wave files of the games with a 22050 hz frequency, which i did using winffmpeg. It reduced the freuqency of bugs, but they still do happen
[14:51:51] <fireflame> so i thought maybe you would have an idea or a clue about a possible solution, i feel i am close tyo fixing the problem
[14:54:28] <fireflame> one theory is that the fmod/dll file maybe flawed, i thought finding a slightly older version of this dll could help but i didnt find any, game uses a 3.30 version of fmod
[15:07:16] <lynxlynxlynx> it manages to crash the os? that's quite something
[15:07:57] <fireflame> actually
[15:08:18] <fireflame> also it often produces a sound that would almost seems to be coming from one of the computer component
[15:08:27] <fireflame> people call it "motorboat bug"
[15:08:32] <lynxlynxlynx> perhaps the audio bell
[15:08:55] <fireflame> http://www.gog.com/forum/desperados_series/possible_solution_to_crashes_maybe/post7
[15:08:57] <Pepelka> Possible solution to crashes (maybe), page 1 - Forum - GOG.com
[15:08:58] <Pepelka> »Download the best games on Windows & Mac. A vast selection of titles, DRM-free, with free goodies, customer love, and one fair price for all regions.«
[15:09:19] <lynxlynxlynx> i would try reenconding the wavs with different settings, but that's it
[15:09:35] <lynxlynxlynx> could easily be multiple bugs at play
[15:09:38] <fireflame> like different frequencies?
[15:10:26] <fireflame> welli sometimes had this bug while playing star wars battle front 2 for example, but id dint make the game crash and the bug faded after a while
[15:10:54] <fireflame> specific bug to old sounds libraries maybe
[15:20:56] <fireflame> ok something strange happened
[15:21:06] <fireflame> i changed the frequency of a single file
[15:21:30] <fireflame> this time it crashed but instead of having to restart computer game just crashed
[15:21:44] <fireflame> the frequency i added was random, like 18050 instead of 22050
[15:22:08] <fireflame> &and for some reason, the crash happened on a sound that came later, not on the modified sound
[15:22:19] <fireflame> what i modified was the sound of a watch
[15:22:39] <fireflame> what made the game crashed was the alert music that you hear when a guard is suspicious
[15:22:59] <fireflame> so how could the sound of the watch affect the music and make it loop
[15:23:36] <lynxlynxlynx> residual state
[15:24:17] --> brada has joined #gemrb
[15:25:17] <fireflame> i feel i am close to finding az solution
[15:25:44] <brada> if you say so…
[15:26:13] <fireflame> well now when it crashes its return to desktop better than restart computer
[15:35:53] <-- turtleman_ has left IRC (Ping timeout: 240 seconds)
[16:17:00] --> turtleman_ has joined #gemrb
[16:27:09] <-- turtleman_ has left IRC (Ping timeout: 246 seconds)
[16:50:04] --> fireflame_ has joined #gemrb
[16:50:17] <fireflame_> i crashed again:/
[16:50:28] <-- fireflame has left IRC (Ping timeout: 246 seconds)
[17:00:54] --> fizzle has joined #gemrb
[17:16:40] --> turtleman_ has joined #gemrb
[17:18:36] --> kpedersen has joined #gemrb
[17:21:11] <-- kpederse1 has left IRC (Ping timeout: 252 seconds)
[17:29:14] <-- brada has left IRC (Quit: brada)
[17:32:39] <fizzle> hey, where did he go?
[17:35:56] <-- turtleman_ has left IRC (Ping timeout: 265 seconds)
[17:48:28] <fireflame_> who?
[17:51:38] <lynxlynxlynx> probably brad
[17:59:13] <fizzle> yeah
[18:28:21] --> kpederse1 has joined #gemrb
[18:29:14] --> brada has joined #gemrb
[18:29:24] <brada> fizzle: im back. whats up?
[18:29:38] <fizzle> hey
[18:29:48] <fizzle> I'm still looking at UpdateScrollbar
[18:30:07] <-- kpedersen has left IRC (Ping timeout: 252 seconds)
[18:30:20] <fizzle> in that code, shouldn't textHeight always be >= pageH?
[18:30:31] <brada> lemme see
[18:31:09] <brada> no i dont think so
[18:31:35] <fizzle> how come?
[18:32:35] <brada> the check is for pageH < frame.h
[18:32:38] <brada> is that what you mean?
[18:32:45] <fizzle> no
[18:32:55] <fizzle> I mean textHeight and pageH
[18:33:22] <brada> so then yes i suppose, im just confused why you are asking since that isnt being compared
[18:33:33] <fizzle> because that's the bug
[18:33:42] <fizzle> I'm in a fresh text area
[18:33:48] <brada> if you say… it worked before your scrollbar change tho
[18:33:55] <fizzle> pageH is 99
[18:34:03] <fizzle> but textHeight is only 88
[18:35:21] <fizzle> I don't really get how the various container heights are updated, though, so I'm not sure what isn't updating correctly here
[18:35:29] <brada> wait
[18:36:11] <brada> pageH is just the dialog select options + the height of the last node + 2x line height
[18:36:51] <fizzle> that far I can follow
[18:37:41] <fizzle> I'm assuming that textHeight should include this plus possible previous conent
[18:37:45] <fizzle> content
[18:43:50] <brada> so pageH < textHeight could be true if the expected newlines arent present
[18:44:29] <brada> except i image the TA is already likely populated...
[18:44:38] <fizzle> the on-screen layout *looks* like they are
[18:45:40] <brada> thats all that matters since its concerned with height and not the actual tet
[18:46:07] <fizzle> but it's using textHeight to fill up the page
[18:46:09] <brada> i mean it does seem that textHeight should be pageH + whatever else
[18:46:18] <fizzle> and that's missing 11 px then
[18:46:28] <brada> but again if thats the case why did it work before?
[18:47:00] <brada> i wrote this an unfortunately long time ago so i dont have 100% recollection
[18:47:09] <fizzle> because by rounding to lines (and rounding up) you didn't notice you had 11 px missing,I suppose
[18:49:18] <brada> maybe, but im not sure *how* textHeight could be off by a line. seems like a miscalculation there would cause all sorts of other bugs
[18:57:08] <brada> in my test textHeight is 15 lines and turning on the debug flag and counting the lines i get 15…
[19:00:00] <fizzle> maybe it's the other way around and nodeBounds or selectFrame is wrong
[19:00:17] <fizzle> the code isn't exactly easy to follow, unfortunately
[19:02:17] <brada> yes, that will get cleaned up a bit in the future after this subview stuff im working on is complete
[19:03:46] --> turtleman_ has joined #gemrb
[19:04:16] <brada> if its always off by one line you can jsut take the *2 off i suppose
[19:05:04] <brada> maybe nodeBounds already includes one of the newlines?
[19:05:18] <brada> if so then, yes, removing the *2 is the correct fix
[19:06:21] <brada> unfortunately, looking at the debug output that doesnt appear to be the case...
[19:06:45] <brada> oh, but perhapps the last newline is part of the select options section?
[19:08:02] <brada> yes, it appears that it probably is
[19:08:12] <brada> so again the removal of the *2 is appropriate
[19:09:46] <brada> added on TextArea.cpp line 629
[19:10:01] <brada> and measuring the size in UpdateScrollbar agrees
[19:10:10] <fizzle> we're still 1px off but it already looks better
[19:10:52] <brada> iirc there is a fixme about a 1px rounding error somewhere in either SB or TA
[19:16:09] <fizzle> selectFrame's h = 51 looks suspicious
[19:17:01] <fireflame_> i would like to sk something
[19:17:07] <fireflame_> this is femrb 0.8
[19:17:25] <fireflame_> so does it mean hopefully we can expect a 1.0 version this year?
[19:18:10] <fireflame_> i am curious to see if people will create good games with gemrb
[19:20:17] <fizzle> better ask lynx about release plans, but if 1.0 means support for Planescape Torment, it's highly unlikely
[19:21:53] <lynxlynxlynx> this year? highly unlikely
[19:22:17] <lynxlynxlynx> we're at 0.8.2, so this year will see one or two more releases
[19:22:57] <lynxlynxlynx> 0.9 is reserved for something big, so it could easily be that we're still at 0.8.x at the end of the year
[19:23:40] <lynxlynxlynx> but you can make it happen faster, either by contributing or getting others to do so ;)
[19:24:11] <lynxlynxlynx> there's not much stopping you from creating games with gemrb right now
[19:24:25] <lynxlynxlynx> at least two studios showed interest so far
[19:24:37] <fireflame_> nice
[19:26:53] <fireflame_> right now if you want a rpg either you look on the mainstream river and have to pick action rpgs, either you look at indie games, which can be more creative
[19:27:30] <fireflame_> i enjoyed dragon age or mass effect, but not as much as fallout or baldur series
[19:27:59] <fireflame_> this year there are good games that may come(or next)
[19:28:09] <fireflame_> under rail, darkest dungeon
[19:42:59] --> raevol has joined #gemrb
[19:46:30] <brada> fizzle: if the off by 1px isnt in fact a rounding error, then its probably due to content containers not collapsing their borders
[19:47:39] <brada> so you could get away with subtracting one from selectFrame height, but put a comment about why if you do
[19:50:35] <lynxlynxlynx> fireflame_: also the two kickstarter favourites come out (unsure about numenera) and perhaps the bg1-bg2 midgame
[19:50:51] <fireflame_> yes indeed
[19:50:52] <-- turtleman_ has left IRC (Ping timeout: 264 seconds)
[19:51:00] <lynxlynxlynx> and the witcher 3 :)
[19:51:11] <fireflame_> i will have to change my rig for this one^^
[19:55:01] <fizzle> brada: I'm not yet convinced it's a rounding error
[19:55:47] <brada> so then what i just said :p
[20:00:36] <-- raevol has left IRC (Quit: Leaving)
[20:09:06] <fizzle> brada: TextContainer.cpp:505
[20:09:42] <fizzle> that's where the extra px comes from but removing the +1 unfortunately causes an infinite loop instead :/
[20:13:10] <brada> fizzle: then it is exactly what i said
[20:13:23] <brada> that +1 is certainly required
[20:14:00] <brada> the very top of TextSpan::LayoutForPointInRegion shows how text spans deal with that by collapsing their borders
[20:23:09] --> edheldil_ has joined #gemrb
[20:28:18] <fizzle> why is the +1 needed?
[20:28:43] <brada> because rectangles are considered intersecting if their borders overlap
[20:29:59] --> raevol has joined #gemrb
[20:30:14] <fizzle> what are we talking about if we say border here?
[20:31:08] <brada> the px between contents
[20:31:25] <brada> jsut adding width and height of the previous one you still get intersecting regions
[20:31:32] <brada> so you need + 1 to both
[20:32:08] <fizzle> what is that intersection needed for?
[20:32:18] <brada> i suppose you could add a parameter to the intersection method if that would make things more clear
[20:32:29] <brada> for determining layout...
[20:32:58] <fizzle> I mean typically (0,0, 10, 10) and (10, 10, 10, 10) would not be considered intersecting
[20:33:33] <fizzle> it seems to just complicate matters
[20:33:51] <brada> i left the intersection algoritim as is because it is used elsewhere
[20:36:26] <fizzle> hm, looking at the other uses I'm not sure that was intentional...
[20:37:57] <brada> maybe not, but i felt safer :p
[20:38:20] <brada> so of course feel free to investigate if you feel it is worth your time… i didnt :)
[20:41:01] <fizzle> I'll give that a try
[20:41:47] <fizzle> much cleaner that way, and as you noticed, you'll forget an increment or decrement sooner or later
[20:42:51] <fizzle> but not today
[20:42:54] <fizzle> gn
[20:42:59] <brada> night
[20:43:08] <-- fizzle has left #gemrb
[20:53:14] <lynxlynxlynx> and if you want to powergame irl: http://imgur.com/gallery/GqLQG
[20:53:15] <Pepelka> Hey Imgur! Want to get stuff done and slay monsters? - Album on Imgur
[20:53:16] <Pepelka> »The Internet's visual storytelling community. Explore, share, and discuss the best visual stories the Internet has to offer.«
[20:55:07] <DrMcCoy> lynxlynxlynx: ...if you happen to have a mobile phone
[20:57:05] <DrMcCoy> fireflame_: If you like the later BioWare games, let me plug my own project: https://xoreos.org/about/index.html ;)
[21:00:06] <lynxlynxlynx> not sure, you can login with a browser
[21:01:38] <fireflame_> the screenshot reminds me about kotor
[21:02:24] <DrMcCoy> You mean the screenshot at the about page? There's also a dedicated screenshots page with actual KotOR screenshots: https://xoreos.org/screenshots/index.html
[21:03:04] <fireflame_> yes the screenshot about th epage
[21:03:16] <fireflame_> the interface is in a way similar to kotor, different but also similar
[21:26:39] <-- raevol has left IRC (Quit: Leaving)
[23:28:45] <-- edheldil_ has left IRC (Ping timeout: 252 seconds)
[23:47:06] <Pepelka> [commit] bradallred: TextArea: fix previous dialog nodes bleeding into view https://github.com/gemrb/gemrb/commit/6d9a16ded6d723f1274d10352636ed5f920e2c2d
[23:53:45] <lynxlynxlynx> oh, i forget if i mentioned it
[23:53:57] <lynxlynxlynx> we don't draw spaces sometimes
[23:54:12] <lynxlynxlynx> it's now reproducible in iwd2 chargen, no ttf involved
[23:55:16] <lynxlynxlynx> Gender:Female etc is drawn, while the ta content is correct