IRC logs for #opendune on OFTC at 2010-02-21
⏴ go to previous day
10:18:34 *** ChanServ sets mode: +v Xaroth
10:18:34 *** ChanServ sets mode: -o TrueBrain
10:18:34 *** ChanServ sets mode: +v DorpsGek
10:19:38 *** ChanServ sets mode: +o TrueBrain
10:41:47 <DorpsGek> SVN: truebrain (r929) -Add: C-ified Security_Check() (well, at least an initial draft)
10:53:40 <DorpsGek> SVN: truebrain (r930) -Fix: cleanup function_names.txt a bit
11:12:12 *** Yexo is now known as Guest5
11:12:17 *** ChanServ sets mode: +v Yexo
11:19:49 <DorpsGek> SVN: truebrain (r931) -Add: named a few variables and functions, based on dig work done for MrFlibble
11:33:14 *** ChanServ sets mode: +v Yexo
13:51:27 <TrueBrain> my script reports 2 emu_ functions of the MPU can be removed now :) Nice progress ;)
13:53:50 *** planetmaker is now known as tycoon
13:54:59 *** tycoon is now known as planetmaker
13:55:49 *** planetmaker is now known as Tycoon
13:56:29 *** Tycoon is now known as planetmaker
13:57:59 <TrueBrain> trunk$ sh unused.sh
13:58:00 <TrueBrain> src/emu_mt32mpu.c: emu_MPU_SetData no longer used
13:58:02 <TrueBrain> src/emu_mt32mpu.c: emu_MPU_Play no longer used
13:58:55 <glx> it was similar for GUI stuff ;)
13:59:10 <TrueBrain> point is, it is unneeded to call via the emu layer
13:59:54 <TrueBrain> k, then it makes sense my script fails ;)
14:00:09 <glx> maybe later they will be called without the FunctionCall wrapper
14:25:46 <glx> isn't the message in offtopic some spam?
14:31:00 *** ChanServ sets mode: +v Yexo
14:31:32 <TrueBrain> let me remove a bunch of users
14:32:51 <TrueBrain> if you join us just to leave your www address, I rather have you not at all :p
14:33:21 <TrueBrain> 17 users left ... 3 look fake, but they have at least once logged in ..
14:35:00 <TrueBrain> ah, yes, I was trying to come up with a plan in case because of some event the openttd webserver becomes unreachable :p
14:52:08 <TrueBrain> how does he keep on finding them :p
14:52:51 <TrueBrain> sadly ... it once again is corruption :(
14:52:56 <TrueBrain> how does that keep on happening ...
14:53:38 <TrueBrain> it really should be a jump to 22A6NNNN
14:54:24 <TrueBrain> if only I had a way to reproduce it :(
14:55:24 <glx> maybe you should try a music/sound enabled build ;)
14:55:31 <tneo> though I just keep on posting them, perhaps you can find a pattern to it
14:55:46 <glx> there are expected unresolved jumps/calls there
14:55:57 <TrueBrain> tneo: yeah, and I am happy you do :) I hope to find a common wrongness ...
14:56:14 <TrueBrain> tneo: src/config.c, change the = 0 to = 7 for sound and music
14:56:25 <TrueBrain> make sure you also update libemu
14:57:15 <tneo> and now it just stalls at the map after completing l1 again
14:57:46 <TrueBrain> I hope it is not due to my latest work :p Let me try ...
14:59:28 <tneo> which version of libemu do I need?
14:59:44 <TrueBrain> well, otherwise latest OpenDUNE doesn't compile, so .. :p
15:00:03 <tneo> then I have that already :-)
15:00:06 <TrueBrain> is that true .. hmm ..
15:00:14 <TrueBrain> no, my statement is wrong
15:00:24 <TrueBrain> damn .. we wrote this too modulair :)
15:00:54 <glx> old libemu just do nothing with mpu commands :)
15:00:58 <TrueBrain> make sure to install timidity, and launch it
15:01:08 <TrueBrain> but OpenDUNE will tell you on startup if you haven't :)
15:01:11 <glx> and very old tells you unsupported port
15:01:29 <tneo> i have timidity needed for openttd :-)
15:01:39 <TrueBrain> then you should hear music ;)
15:02:13 <TrueBrain> compile libemu with: make ALSA:=1
15:02:48 <TrueBrain> normally you compile libemu with 'make'
15:03:05 <TrueBrain> k, OpenDUNE hangs while displaying the map
15:03:10 <glx> I use ./compile.sh (because I need to set CFLAGS for -m32
15:03:18 <TrueBrain> I have make-i386.sh ;)
15:08:27 <TrueBrain> it hangs at the exact same point
15:09:23 <tneo> hmm i don't seem to be able to get libemu with opensuse or I miss something...
15:09:45 <TrueBrain> libemu is also a piece of software we wrote
15:10:15 <TrueBrain> svn://svn.opendune.org/extra/libemu
15:10:45 <TrueBrain> it gives the libemu.so you need
15:11:12 <tneo> and which svn command do I run to get libemu?
15:11:17 <TrueBrain> somehow something breaks in ShowMap .. weird ...
15:11:25 <TrueBrain> svn checkout svn://svn.opendune.org/extra/libemu
15:11:46 <TrueBrain> ln -s libemu.so /your/path/to/opendune
15:13:13 <tneo> > /your/path/to/opendune/trunk or the root ?
15:13:32 <glx> where is the opendune exe
15:13:58 <glx> so trunk if it's in trunk :)
15:14:39 <tneo> [EMU] [ INT21:3D ] Requesting mode '20' which is not completely supported.
15:14:39 <tneo> [EMU] [ INT21:3D ] Requesting mode '20' which is not completely supported.
15:14:39 <tneo> [EMU] [ INB:331 ] Not Yet Implemented
15:14:39 <tneo> Crash-dump in 'memory/crash.bin'
15:15:04 <glx> or wrong libemu version :)
15:15:06 <TrueBrain> glx: it appears sound is the reason the map fails
15:15:25 <tneo> Checked out revision 931.
15:15:55 <TrueBrain> tneo: in your opendune dir (trunk), remove libemu.so
15:15:59 <TrueBrain> and do the ln -s command again
15:16:24 <glx> tneo: that's an "very" old libemu then ;)
15:16:51 <TrueBrain> hmmm .. even with sound disabled it doesn't continnue
15:16:55 <TrueBrain> something smells fishy
15:17:11 <tneo> I just ran svn checkout svn://svn.opendune.org/extra/libemu glx
15:17:29 <glx> but the libemu.so in your opendune dir is the very old one
15:17:37 <tneo> make: stat: libemu.so: Too many levels of symbolic links
15:17:37 <tneo> make: *** No rule to make target `libemu.so', needed by `opendune'. Stop.
15:17:57 <TrueBrain> remove libemu.so again from trunk
15:18:00 <TrueBrain> make sure you are in the libemu dir
15:18:03 <TrueBrain> then execute the ln -s line
15:18:29 <glx> or from opendune dir: ln -s path/to/libemu.so
15:25:13 <TrueBrain> and every time to other places
15:25:20 <tneo> Program Termination: jumped to 65A5:A11A, which is not decompiled.
15:25:20 <tneo> The jump was triggered at decompiled/cs__B488.c:238
15:25:20 <tneo> The jump appears to originate from B488:017F.
15:25:29 <glx> wrong emu_cs somewhere I'd say
15:25:44 <TrueBrain> his for sure has a wrong cs and ip
15:25:59 <TrueBrain> the ones I had were all over the place, from input to MPU
15:26:07 <TrueBrain> (everything running in the timer in fact)
15:26:08 <glx> no the question is which emu_ wrapper is faulty
15:26:36 <TrueBrain> with soundDrv = 0, are sound things still called?
15:28:57 <glx> IIRC the CallFunction wrapper are still called, but the driver index is 0xFFFF
15:30:14 <TrueBrain> I keep on getting a score of 65 :)
15:30:24 <TrueBrain> (savegame is seconds before it finishes the mission :p)
15:31:43 <glx> but driver csip is 0 too, so nothing should be called
15:34:46 <TrueBrain> so mistake on my side
15:34:49 <TrueBrain> will trace it a bit later :)
15:35:04 <glx> probably too early CS:IP pop
15:37:33 <glx> no reason, all File_* don't call anything
15:43:37 <TrueBrain> so the question of all questions: what is wrong ;)
15:44:07 <glx> drivers wrapper are ok (pop in temp csip)
15:44:15 <Xaroth> TrueBrain: you use an IDE for python?
15:45:14 <TrueBrain> except memory usage: yes
15:45:34 <glx> Program Termination: jumped to 44AF:0C57, which is not decompiled.
15:45:35 <glx> The jump was triggered at decompiled/cs__AB00.c:1341
15:45:35 <glx> The jump appears to originate from AB00:0C57.
15:45:40 <TrueBrain> fuck, didn't fix the locking problem last time
15:45:54 <glx> hmm that's inside mt32mpu.adv
15:45:57 <TrueBrain> hmm .. no .. I did ..
15:46:02 <TrueBrain> glx: it gives problems all over the place :)
15:46:08 <TrueBrain> I think the memory is randomly changed
15:46:58 <TrueBrain> no, locking is correct and valid
15:48:17 <glx> hmm let's retry without music/sound (this unresolved jump is ok I think)
15:49:04 <glx> should be an emu_sp+=2 and something else
15:49:26 <TrueBrain> yeah, random memory corruption, that is for sure
15:49:37 <TrueBrain> "jumped to 368B:3B33"
15:50:24 <TrueBrain> a place that should call 22A6
15:50:29 <TrueBrain> it just randomly writes in the memory :)
15:51:43 <TrueBrain> it really is File stuf
15:52:04 <TrueBrain> and I can't test it function by function, because of the way it works
15:52:13 <TrueBrain> or rather, the way I rewrote it :p
15:53:23 <TrueBrain> can you all promise me not to look to the next commit?
15:53:34 <Xaroth> you shouldn't have said that
15:54:15 <TrueBrain> lets first see if it fixes this problem
15:54:30 <glx> without sound/music it hangs
15:54:43 <TrueBrain> glx: just wait a bit
15:54:54 <TrueBrain> it is the timer interrupt that is crashing
15:55:28 <TrueBrain> I say again: random places :)
15:55:34 <TrueBrain> well, with some kind of order
15:55:37 <TrueBrain> but it depends on when it happens
15:55:55 <glx> 2BD1 is the only custom handler added when there's no sound
15:55:55 <DorpsGek> SVN: truebrain (r932) -Fix (r878): (nameless)
15:56:15 <TrueBrain> tneo: update to latest, and you are good to go again :)
15:56:20 <TrueBrain> tnx for the bug report, it was useful after all ;)
15:57:28 <TrueBrain> Xaroth: as I said: memory usage :)
15:57:39 <TrueBrain> Xaroth: but it really is nice .. it does a lot .. not always correct, but most of the time
15:57:44 <TrueBrain> 'refactoring' mostly is VERY useful
15:57:55 <TrueBrain> CTRL+R on a variable or function allows you to change the name project-wide
15:58:06 <TrueBrain> When C-ifying, I first name things loc<number>
15:58:11 <TrueBrain> then CTRL+R, new name, enter, done
15:58:13 <Xaroth> M$ should get off their arses and make VC open source :P
15:58:27 <tneo> i still don't have sound though :-)
15:58:31 <TrueBrain> also CTRL+Click on a function name goes there
15:58:40 <TrueBrain> tneo: did you enable it ingame?
15:58:44 <TrueBrain> and did you modify src/config?
15:59:04 <TrueBrain> on startup, does it complain about timidity?
15:59:14 <TrueBrain> [EMU] No valid MIDI output ports.
15:59:16 <TrueBrain> Please install and start Timidity++ like: timidity -iA
16:00:04 <TrueBrain> when OpenDUNE runs, what does 'aconnect -o -i' tell you?
16:00:07 <tneo> Found MIDI output: Midi Through Port-0
16:00:29 <glx> not surprising you don't hear anything I think ;)
16:00:52 <TrueBrain> client 128: 'TiMidity' [type=user]
16:00:56 <TrueBrain> it should tell you something like that
16:01:04 <TrueBrain> and more important: client 129: 'OpenDUNE - Pre v0.3' [type=user]
16:01:32 <tneo> client 0: 'System' [type=kernel]
16:01:32 <tneo> client 14: 'Midi Through' [type=kernel]
16:01:34 <tneo> client 16: 'TerraTec DMX6Fire' [type=kernel]
16:01:36 <tneo> 0 'MIDI-Front DMX6fire 0'
16:01:38 <tneo> 32 'Wavetable DMX6fire 0'
16:01:40 <tneo> client 128: 'OpenDUNE - Pre v0.3' [type=user]
16:01:49 <Xaroth> 117mb total install size
16:02:07 <TrueBrain> hmm .. so it does work
16:02:07 <glx> real midi in your sound card ?
16:02:20 <TrueBrain> but ... it ends up at Midi Through
16:02:25 <TrueBrain> which is either not connected or what ever :p
16:03:39 <glx> you have not started timidity ;)
16:03:52 <TrueBrain> glx: it won't select it either way
16:03:56 <TrueBrain> it picks the first valid MIDI port
16:04:04 <TrueBrain> and client 14 says it is a valid MIDI out
16:04:27 <glx> in my VM it selects Midi through too
16:04:51 <TrueBrain> the algorithm is stupid: walk clients (from low to high), walk ports (from low to high), pick first which allows sending MIDI commands to
16:05:11 <TrueBrain> it should be made configurable I guess, but for that we need to move if from libemu to OpenDUNE, which means you need to be finished first ;)
16:05:42 <glx> not really, I can merge midi stuff in MPU
16:06:08 <glx> just inside MPU_WriteData and related
16:06:22 <glx> (the functions writing to MPU ports)
16:06:33 <TrueBrain> okay, let me rephrase: I was planning to wait for you to finish what you are doing now, before suggesting that :)
16:06:42 <TrueBrain> and before the 0.3 release
16:07:29 <TrueBrain> but not before we have some way to really do configuring of some kind :p
16:07:35 <TrueBrain> well .. plenty of time :)
16:07:51 <TrueBrain> I am happy I found r932 :)
16:08:23 <tneo> with timidity started still no sound ...
16:08:35 <TrueBrain> tneo: you somehow would need to drop client 14 and 16 :p
16:09:40 <TrueBrain> no, I have no simple suggestions :)
16:09:55 <TrueBrain> will put it on my todo list :)
16:10:23 <TrueBrain> glx: but really, just first try to remove the driver-layer; then we can start the next phase of making it pretty ;)
16:10:33 <TrueBrain> I also really have to redo scenario.c ... damn, it is ugly
16:10:55 <TrueBrain> slowly I try to remove more and more 'csip' stuff in the C-ify layer
16:11:05 <TrueBrain> sadly, it mostly means C-ifying stuff :)
16:11:48 <TrueBrain> and now I go indoor boarding, so I wish you all a very good evening!
16:12:00 *** planetmaker is now known as DevServer
16:12:40 *** DevServer is now known as planetmaker
16:19:34 *** planetmaker is now known as Tycoon
16:20:02 *** Tycoon is now known as planetmaker
continue to next day ⏵