#exult@irc.freenode.net logs for 17 Jan 2011 (GMT)

Archive Today Yesterday Tomorrow
Exult homepage


[00:08:30] <Malignant_Manor> They can barely fit side by side if I go full width without a check mark.
[00:12:40] <Dominus> definitely not good side by side
[00:29:02] --> Rottingbeef has joined #exult
[00:56:27] <-- Darrenor64 has left IRC ()
[01:13:46] --> Colourless has joined #exult
[01:13:46] --- ChanServ gives channel operator status to Colourless
[02:02:05] <-- SiENcE has left IRC (Read error: Connection reset by peer)
[02:03:12] <-- Malignant_Manor has left IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
[02:12:36] <-- Rottingbeef has left IRC ()
[02:39:58] <-- Kirben has left IRC (Ping timeout: 260 seconds)
[02:49:27] --> Kirben has joined #exult
[02:49:27] --- ChanServ gives channel operator status to Kirben
[02:58:36] <-- Dominus has left IRC (Quit: Leaving.)
[03:13:53] --> Rottingbeef has joined #exult
[03:55:29] <-- Rottingbeef has left IRC ()
[05:57:31] <-- Sevalecpp has left IRC (Read error: Connection reset by peer)
[06:00:33] <-- eviltar has left #exult
[06:01:12] --> eviltar has joined #exult
[07:18:23] <-- Colourless has left IRC (Ping timeout: 255 seconds)
[07:44:53] --> Rottingbeef has joined #exult
[07:58:24] <-- Kirben has left IRC ()
[08:02:32] --> Colourless has joined #exult
[08:02:32] --- ChanServ gives channel operator status to Colourless
[08:15:09] --> Kirben has joined #exult
[08:15:09] --- ChanServ gives channel operator status to Kirben
[11:16:58] --> Dominus has joined #exult
[11:16:59] --- ChanServ gives channel operator status to Dominus
[11:52:36] --> Sevalecpp has joined #exult
[11:52:36] <-- Sevalecpp has left IRC (Changing host)
[11:52:36] --> Sevalecpp has joined #exult
[11:56:46] --> SiENcE has joined #exult
[12:51:31] <-- Rottingbeef has left IRC (Ping timeout: 276 seconds)
[13:05:31] <-- Kirben has left IRC ()
[13:22:39] --> Malignant_Manor has joined #exult
[13:24:01] <Malignant_Manor> Dominus: , now you can do your documentation.
[13:24:15] <Dominus> he he
[13:24:30] <Malignant_Manor> I decided on the simpler menu.
[13:24:43] <Malignant_Manor> It just rebuilds if you toggle fullscreen
[13:24:54] --> Rottingbeef has joined #exult
[13:29:03] <Dominus> let me test my full-automatic-snapshot creation and upload script...
[13:29:23] <Dominus> and I'll be gone for an hour or a bit more and will then be able to comment on it
[13:29:38] <Dominus> but simple approach sounds good
[13:31:30] <Malignant_Manor> It's certainly easier to code, read the code, and update.
[13:32:21] <Malignant_Manor> I wouldn't have had room for the Exult HD. (if/when that would get added.)
[13:56:33] <Malignant_Manor> Looks like I need to update the fullscreen change. It doesn't revert right.
[14:55:06] <Dominus> hmm, thinking of making snapshots almost totally automated. When receiving svn trunk update message, start script...
[14:56:33] <Malignant_Manor> That would be nice for you.
[14:58:08] <Dominus> and actually easy to do. but I rather not.
[14:58:33] <Dominus> It could be disturbing if suddenly a terminal pops up and disturbs whatever else I do
[14:58:49] <Dominus> better not :)
[15:00:02] <Malignant_Manor> I have no idea why canceling fullscreen toggle is screwing up window positioning.
[15:29:05] <Malignant_Manor> Colourless: , do you have any idea why the screen doesn't get re centered on video gump revert from fullscreen to window with my new settings?
[15:31:10] <Dominus> oh wow, revision 6746 crashed Exult on toggling fullscreen
[15:37:34] <Malignant_Manor> Colourless: , do you have any idea why the screen doesn't get re centered on video gump revert from fullscreen to window with my new settings?
[15:37:39] <Malignant_Manor> oops sorry
[15:37:45] <Malignant_Manor> wrong window
[15:45:44] <Colourless> no idea. haven't been paying any attention to whats been going on
[15:46:26] <Malignant_Manor> It works right if the fullscreen had the same resolution as the window it is reverting to.
[15:48:15] <Colourless> well, sounds like a variable isn't being reset, but i'm going off now, so, sorry can't help out at the moment
[15:48:59] <Colourless> asking me about an issue at 2 in the morning my time... not the best :-)
[15:49:11] <Malignant_Manor> 6747 should have that fixed.
[15:50:02] <Malignant_Manor> That was uploaded about 25 minutes ago.
[15:51:03] <Malignant_Manor> unless it is something unrelated to what I forgot
[15:52:50] <Dominus> nope, doesn't fix that
[15:55:53] <-- Colourless has left IRC (Ping timeout: 255 seconds)
[15:56:07] <-- Malignant_Manor has left IRC (Ping timeout: 276 seconds)
[15:58:33] --> Malignant_Manor has joined #exult
[16:01:43] <Malignant_Manor> Dominus: , can you post your video section of the cfg at pastebin?
[16:03:17] <Dominus> http://pastebin.com/ZuKXZVwh
[16:06:47] <Malignant_Manor> The crash isn't happening for me (but this build isn't optimized)
[16:07:02] <Malignant_Manor> How are you reproducing this?
[16:08:10] <Dominus> I start Exult -> setup -> video -> click on fullscreen
[16:08:13] <Dominus> crash
[16:08:57] <Dominus> I'll make a debug built
[16:09:42] <Malignant_Manor> I wish it didn't take forever for me to build anything.
[16:10:12] <Malignant_Manor> I'm building an optimized version since it should crash.
[16:10:16] <Dominus> you need a faster machine
[16:10:46] <Dominus> actually I had no optimization set
[16:10:59] <Malignant_Manor> It should default to 02
[16:11:04] <Malignant_Manor> o2
[16:11:12] <Dominus> oh, ok
[16:12:05] <Malignant_Manor> It should crash in debug though if it is trying to go to a nonexistent text string.
[16:13:03] <Dominus> in Gump_ToggleTextButton::activate (this=0x100e77b10, button=1) at Gump_ToggleButton.cc:84
[16:13:47] <Dominus> http://pastebin.com/zEYW65R1
[16:19:51] <Malignant_Manor> can you comment out line 129 in VideoOptions_gump.cc and see if it still crashes?
[16:20:06] <Malignant_Manor> It won't crash for me.
[16:20:25] <Malignant_Manor> I need to disable the parachute and build again.
[16:35:49] <eviltar> not to chime in unwarranted-ly, but i have to build exult with optimzations off or i get unexpected crashes :\
[16:36:17] <eviltar> \Od
[16:36:27] <Malignant_Manor> What operating system and when?
[16:36:41] <eviltar> eh, something kind of like NT, on teh 360
[16:36:44] <Malignant_Manor> Are you talking about the 360 version?
[16:36:46] * wjp wonders if those wouldn't be expected crashes then ;-)
[16:37:01] <eviltar> ^ maybe :)
[16:37:14] <eviltar> its mostly like winAPI tho
[16:38:18] <eviltar> but.. it isnt running slow with them off so no tears shed here
[16:41:10] <Malignant_Manor> If Macs are so stable, why is it crashing instead of Windows?
[16:41:55] <eviltar> my build? or are you guys working on a ppc build again?
[16:42:17] <Malignant_Manor> No, Dominus's crash
[16:42:31] <eviltar> b'c big endian is a pita
[16:42:47] <eviltar> ?
[16:43:13] <Malignant_Manor> Exult should automatically take endianess into account on compile.
[16:43:36] <eviltar> is there a defenition it uses? or all operations are bit by bit?
[16:43:54] <wjp> that question doesn't make sense
[16:44:15] <Malignant_Manor> The target you setup when you compile should take care of it
[16:44:28] <eviltar> i see what you mean
[16:44:51] <eviltar> the MAC sections are already accounting for big endian?
[16:45:42] <wjp> byte swaps are done when loading/manipulating data where appropriate
[16:46:00] <Malignant_Manor> some sections of the code have this, #if SDL_BYTEORDER == SDL_LIL_ENDIAN
[16:47:06] <eviltar> ok, im not trying to be a smarty pants, im reletively a newb at c++
[16:47:41] <wjp> handling LE/BE isn't c++-specific at all :-)
[16:47:48] <eviltar> but the compiler in VS, does some funky things building for PPC asm if that's what he's using
[16:48:13] <wjp> it's mostly just a matter of making sure that you read data from disk in the right way
[16:48:28] <eviltar> i dig it,
[16:48:43] <eviltar> i never had to account for endianess doign PHP/ASP
[16:49:13] <eviltar> its been a new concept
[16:49:43] <wjp> ah, I see
[16:50:56] <Malignant_Manor> Dominus: , are you there?
[16:54:28] <-- RadoS has left IRC (Remote host closed the connection)
[16:54:47] <eviltar> but for instance, in gamewin.cc, i had to do change the try, catch b'c the way it compiled for PPC was not the same as for x86
[16:54:52] <eviltar> http://pastebin.com/Y9N6vGku
[16:55:22] <eviltar> in Game_window::read_gwin
[16:55:44] <Malignant_Manor> It's easier to see changes with a diff
[16:56:00] <eviltar> everything moved inside the TRY{}
[17:00:57] <Malignant_Manor> wjp: , here's the diff http://pastebin.com/GhSe7anE
[17:01:11] <Malignant_Manor> I never messed with endianess so I am no help.
[17:01:54] <eviltar> that error was strictly to do with the asm produced by the compiler
[17:02:07] <eviltar> i dont think that one was an endian issue
[17:03:14] <eviltar> the compiler wasn't generating a proper stack frame to handle the try/catch
[17:27:33] <wjp> that's a weird patch
[17:27:43] <wjp> what is it trying to address?
[17:28:47] <wjp> note that that function itself throws a file_read_exception
[17:30:08] <Dominus> Malignant_Manor: sorry, baby bed arrived and I had to build it...
[17:30:33] <-- SiENcE has left IRC (Quit: @all: cya)
[17:32:51] <Dominus> Malignant_Manor: with line 129 commented out, no more crash
[17:34:19] <Dominus> if I replace rebuild_buttons with rebuild_dynamic_buttons it works as well, no idea if that helps though :)
[17:34:53] <Malignant_Manor> It deletes itself.
[17:35:08] <Malignant_Manor> But it shouldn't affect it.
[17:35:47] <Malignant_Manor> I can try to separate it from the other buttons.
[17:36:56] <Dominus> might be that it is just stricter on Mac :)
[17:37:23] <Dominus> I like how it works with the settings for fullscreen
[17:37:29] <wjp> now that I'm actually at home, what is the crash? :-)
[17:38:00] <Dominus> eviltar or me?
[17:38:53] <eviltar> that patch was to address an unhandled exception trying to start a new fresh game
[17:39:25] <wjp> Dominus: you :-)
[17:40:11] <Dominus> with current svn in the video settings gump, clicking on fullscreen. gdb http://pastebin.com/zEYW65R1
[17:40:13] <wjp> eviltar: a much easier way of doing exactly the same thing is not throwing that exception in there in the first place.
[17:40:21] <eviltar> honestly idk why it works like that
[17:40:38] <eviltar> see, i wondered that
[17:40:59] <eviltar> but i thought maybe detecting the excpetions was somethign i should keep if other parts of code were dependant on it
[17:41:56] <Malignant_Manor> Dominus:, can you try this? http://pastebin.com/tkUFKAaT
[17:42:04] <wjp> ==17764== Invalid read of size 8
[17:42:04] <wjp> ==17764== at 0x6D93E3: Gump_ToggleTextButton::activate(int) (Gump_ToggleButton.cc:84)
[17:42:10] <wjp> ==17764== Address 0xe52a6d0 is 0 bytes inside a block of size 88 free'd
[17:42:10] <wjp> ==17764== at 0x4C2516E: operator delete(void*) (vg_replace_malloc.c:346)
[17:42:12] <wjp> ==17764== by 0x6D4FBA: VideoTextToggle::~VideoTextToggle() (VideoOptions_gump.cc:90)
[17:43:04] <wjp> Malignant_Manor: that's a no-op
[17:43:14] <wjp> that patch does nothing
[17:43:18] <Malignant_Manor> It deletes itself so I probably should make sure it doesn't
[17:43:32] <Malignant_Manor> I can try separating it from rebuild.
[17:46:26] <Dominus> and yes, wjp is of course right - crashes as well
[17:48:17] <wjp> it does look like the issue is that the button gets deleted in the middle of its activate() method
[17:49:16] <Malignant_Manor> http://pastebin.com/Lcs99cXw
[17:49:44] * Dominus reverted snapshot to last revision working and set up his build system to automatically update svn/build snapshot/upload it when I don't react to the prompt after three minutes
[17:52:08] <Dominus> Malignant_Manor: still crashes http://pastebin.com/H4Le8isV
[17:52:26] <Dominus> just under gdb it didn't crash right away, only on 2nd try
[17:56:11] <Dominus> unfortunately I have to go now. going to see green hornet with friends, so it's gonna be some hours later when I can take a look at a patch again
[17:59:22] <Malignant_Manor> wjp: , do you have any idea?
[18:05:49] <wjp> hm
[18:06:24] <wjp> valgrind now already shows weird things happening in VideoOptions_gump::VideoOptions_gump()
[18:06:34] <Malignant_Manor> My other patch was hasty and didn't initialize it. http://pastebin.com/sdh5J9aW
[18:07:03] <Malignant_Manor> I had changed the menu to do separate settings for window and fullscreen.
[18:08:04] <Malignant_Manor> I tried to have it reload settings and rebuild after you toggle the fullscreen button.
[18:08:26] <wjp> valgrind doesn't complain now
[18:09:10] <Malignant_Manor> There was an uninitialized variable after my quick move.
[18:12:25] <Malignant_Manor> does that mean it doesn't crash anymore?
[18:14:38] <wjp> yes
[18:16:15] <Malignant_Manor> Do you know how the game window gets centered?
[18:16:45] <Malignant_Manor> With these settings changes, it doesn't want to recenter properly when
[18:17:27] <Malignant_Manor> you cancel out of a fullscreen that is of a different resolution than your window resolution you are switching to.
[18:17:36] <Malignant_Manor> It worked before my changes.
[18:21:12] <Malignant_Manor> With these video settings, open the video options menu, toggle fullscreen, and hit apple. Cancel and the window will not be centered. http://pastebin.com/W2KuCV28
[18:35:53] --> RadoS has joined #exult
[18:37:09] <-- RadoS has left IRC (Client Quit)
[18:37:39] <Malignant_Manor> Thanks (almost forgot to say that)
[18:37:41] --> RadoS has joined #exult
[19:14:07] <-- ParuNexus has left IRC (Ping timeout: 272 seconds)
[19:29:02] --- servus_ is now known as servus
[19:39:00] <-- Malignant_Manor has left IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
[19:41:36] --> ParuNexus has joined #exult
[22:02:36] <-- Rottingbeef has left IRC ()
[22:04:30] <Dominus> thanks for figuring this stuff out. Unfortunately my automatic building didn't yet trigger, let's see how it behaves next time...
[22:05:09] <Dominus> Green Hornet is great. If you can avoid it it's not needed to be viewed in 3D (as so many other movies, too)
[22:09:12] <-- Dominus has left IRC (Quit: Leaving.)
[22:23:10] --> Kirben has joined #exult
[22:23:10] --- ChanServ gives channel operator status to Kirben
[22:51:57] --> Rottingbeef has joined #exult
[22:55:14] --> Colourless has joined #exult
[22:55:14] --- ChanServ gives channel operator status to Colourless
[23:23:11] <-- Kirben has left IRC (Read error: Connection reset by peer)
[23:28:41] --> Marzo has joined #exult
[23:34:33] --> Kirben has joined #exult
[23:34:33] --- ChanServ gives channel operator status to Kirben
[23:36:15] <-- Rottingbeef has left IRC ()