IRC logs for #opendune on OFTC at 2013-01-14
⏴ go to previous day
14:15:40 <glx> i is not initialised and in my case it started at 8272
14:20:44 <glx> causing an almost immediate exit
14:53:07 <glx> anyway I think I found the problem about missing string
14:53:53 <TrueBrain> I am sure we wrote code to handle the situation, as we tested with it and everything
14:53:57 <TrueBrain> but I guess it never made trunk :D
14:54:41 <TrueBrain> (the US had the STR_LOAD_GAME empty, and a STR_LOAD_A_GAME as string)
14:55:23 <TrueBrain> its the only version that has STR_LOAD_A_GAME btw
14:57:29 <glx> no EU is the only that has STR_LOAD_GAME :)
14:58:11 <TrueBrain> what I dont get, is that Iam sure we wrote code for it
14:58:15 <TrueBrain> so .. no clue what happened to that :P
14:58:48 <glx> indeed we wrote code for it but there's something weird in the reading I think
14:59:34 <TrueBrain> yeah, we have code that patches it ..
15:00:24 <TrueBrain> but the code is a bit fishy tbh
15:01:57 <TrueBrain> ah, yes, dirty hack
15:02:17 <TrueBrain> string.c, line .. where does it tell my line ...
15:02:42 <TrueBrain> basically, if there are only 339 strings, it is not the EU, so it copies the STR_LOAD_A_GAME to STR_LOAD_GAME
15:02:48 <TrueBrain> so that should work
15:02:58 <TrueBrain> I am also pretty sure we tested this :P
15:05:09 <glx> hmm let's print some stuff via opendune
15:05:51 <glx> yes I know, but I modified langread to show the indexes as opendune should see them ;)
15:32:14 <glx> the problem seems to be skipping of empty lines
15:36:03 <glx> it's indeed not empty it contains a space
15:36:20 <glx> so it's not skipped and our fix doesn't trigger
15:40:13 <TrueBrain> so we should skip empty lines?
15:40:52 <glx> if (strlen(src) == 0 && s_strings[0] != NULL) {
15:41:00 <glx> but it fails for this one
15:41:06 <TrueBrain> with empty I mean trimmed empty ;)
15:41:35 <TrueBrain> or the hack could change in:
15:41:54 <TrueBrain> if (strcmp(s_trings[STR_LOAD_GAME], " ") == 0) ...
15:42:37 <glx> the hack already works for FRE and GER
15:43:35 <TrueBrain> or, ofc, if ((strlen(src) == 0 && s_strings[0] != NULL) || (strlen(src) == 1 && src[0] == ' ')) {
15:54:00 <TrueBrain> clean it up and commit :D
16:29:50 <TrueBrain> ctype still needed in ini.c?
17:34:40 <TrueBrain> lovely, comments totally unrelatd to the code itself
17:34:43 <TrueBrain> you got to love random trolls
17:35:30 <TrueBrain> not to mention that his comment is totally and utterly void
17:35:52 <TrueBrain> I guess he doesn't know #include <string.h> has nothing to do with #include "string.h"
17:36:05 <TrueBrain> he is just a troll; by lack of other words
17:36:37 <TrueBrain> I wonder if I should just remove the useless comments
17:37:49 <glx> we even have many files with the same name ;)
17:38:24 <TrueBrain> any project will; that is why compilers differ between <> and "" :P
17:40:49 <TrueBrain> think for once I am going to reply to him in a not-so-nice way
17:40:53 <TrueBrain> getting a bit tired of it
17:42:11 <TrueBrain> I did love my comment ... it just read: *facepalm*
17:42:19 <TrueBrain> but I decided not to post it :P
17:48:02 <Xaroth|Work> even I facepalmed at that comment
17:48:15 <Xaroth|Work> and with my general lack of C knowledge, that sais a lot
17:50:34 <TrueBrain> didnt know you spoke french?
17:51:34 <glx> totally unappropriate in the context ;)
17:52:11 <Xaroth|Work> like you never make spelling errors...
17:52:20 <TrueBrain> its was just funny!
17:52:32 <TrueBrain> and glx' comment makes it even more funny :D
17:57:24 <TrueBrain> there, told him it is the wrong place to post these things
17:57:28 <TrueBrain> told him I would remove his comment
17:57:37 <TrueBrain> told him to make a new issue if he is sure about his comment
17:57:41 <TrueBrain> removed the comments after that
17:59:39 <TrueBrain> felt like the cleanest solution
18:10:42 <TrueBrain> created bug about DSP/MPU deadlock
18:15:35 <TrueBrain> back to 12 open issues, of which most dsw has a pending fix
18:15:42 <TrueBrain> they only need to be validated and applied to our tip :)
18:15:47 <TrueBrain> not always trivial; but .. yeah
18:15:53 <TrueBrain> guess I can do that next weekend :)
18:16:03 <TrueBrain> will only leave the 2 deadlock issues (1 on Windows, 1 on Linux)
18:16:08 <TrueBrain> easiest is disabling MPU :P:P
19:28:01 <TrueBrain> owh, he really is such a wining bitch, omfg
20:31:14 <glx> about deadlocks, the code handling timer, mpu and dsp is a mess :)
20:50:17 * glx have an idea but that implies many changes
20:51:21 <glx> like create a thread that creates the window and do the polling
20:51:45 <glx> then dsp event are handled by the window, and the timer too
20:52:26 <glx> and the rest of opendune is in the main thread
20:53:11 <glx> well it's something I can try
21:49:18 <glx> "A handle to the window to be associated with the timer. This window must be owned by the calling thread." so managing timer from main thread is a problem if the window is in its own thread
21:49:56 <glx> "A handle to the window whose messages are to be retrieved. The window must belong to the current thread. " so dispatching event via another thread is a problem if window is in main thread
21:53:09 <glx> I need to think more about it
21:55:17 <glx> now I remember why window must be created in timer thread (Video_Tick) :)
continue to next day ⏵