The Allegro Wiki is undergoing maintenance, please be patient.

Allegro Hack Day

From Allegro Wiki
Jump to: navigation, search

Next IRC meeting

Name Can attend
trentg yes
elias yes
Matthew Leverton  ???
mimix yes
Peter Hull no (away, sorry)
tjaden maybe (bit early)
goalieca  ???


Topics

  • 4.9.4 release
  • 4.3 branch release

Log

Sep 06 14:20:13 -->	tjaden (n=tjaden@92.71.96.58.exetel.com.au) has joined #allegro-dev
Sep 06 14:20:14 ---	ChanServ gives voice to tjaden
Sep 06 14:22:18 <elias>	good morning tjaden
Sep 06 14:22:36 <tjaden>	eugh, morning
Sep 06 14:23:30 <tjaden>	gonna get some breakfast first, back later
Sep 06 14:23:59 <trentg>	Maybe we should have future hackdays later, like 22:00
Sep 06 14:24:34 <elias>	will get late here then :P
Sep 06 14:24:49 <elias>	but i wouldn't mind
Sep 06 14:27:43 <elias>	kazzmir_: just in wesnoth, you can only build either release or debug there, not both at once
Sep 06 14:27:46 <elias>	*just checked
Sep 06 14:27:58 <kazzmir_>	well.. sucks for them
Sep 06 14:28:53 <elias>	yeah, but now maybe we know why :P
Sep 06 14:36:17 <kazzmir_>	somehow -Laddons/font/lib is getting in there
Sep 06 14:36:26 <tjaden>	ok back
Sep 06 14:38:20 <elias>	kazzmir_: but did you fix the 4 times config?
Sep 06 14:38:26 <kazzmir_>	not yet
Sep 06 14:39:05 <elias>	> grep font/lib SConstruct scons/*
Sep 06 14:39:05 <elias>	scons/all.scons:    demoEnv.Append(LIBPATH = ['addons/font/lib'])
Sep 06 14:39:05 <elias>	scons/all.scons:    exEnv.Append(LIBPATH = ['addons/font/lib'])
Sep 06 14:39:16 <kazzmir_>	yea but that ends up being scons_build/release/addons/font/lib
Sep 06 14:39:27 <kazzmir_>	if you remove those lines you see the -Lscons_build/... doesnt show up
Sep 06 14:39:51 <elias>	ah, must be duplicate=0
Sep 06 14:39:56 <elias>	try adding that to the SConscript in unix.scons
Sep 06 14:40:02 <kazzmir_>	ohh yknow its because of settings.cfg
Sep 06 14:40:05 <elias>	oh
Sep 06 14:40:07 <elias>	ok
Sep 06 14:40:14 <kazzmir_>	i made the env's clone but didnt erase settings.cfg
Sep 06 14:41:24 <elias>	what happened to milan?
Sep 06 14:42:09 <kazzmir_>	does that 911 really mean something?
Sep 06 14:42:24 <elias>	i assume some kind of emergency came up..
Sep 06 14:45:31 <CIA-31>	alleg: kazzmir * r10722 /allegro/branches/4.9/ (5 files in 5 dirs): clone addons envs
Sep 06 14:47:36 <tjaden>	did ex_fs_resize restore the video mode properly before, for anyone?
Sep 06 14:47:46 <trentg>	It did for me
Sep 06 14:47:48 <mattymatt2>	bro's bike broke
Sep 06 14:47:58 <mattymatt2>	^milan's 911
Sep 06 14:48:37 <elias>	for me as well
Sep 06 14:48:44 <kazzmir_>	elias, do you mind if i remove the duplicate=0 things?
Sep 06 14:48:56 <elias>	kazzmir_: i don't like the duplicating
Sep 06 14:49:02 <kazzmir_>	it causes extra -I flags to be included
Sep 06 14:49:06 <elias>	yeah
Sep 06 14:49:12 <elias>	i mean
Sep 06 14:49:16 <kazzmir_>	fine.. you want duplicate=0 ?
Sep 06 14:49:18 <elias>	i assume, -I into the original
Sep 06 14:49:33 <elias>	i'd prefer it, if possible
Sep 06 14:49:38 <elias>	else grep has messed up results
Sep 06 14:49:45 <elias>	everything 5 times from each build variant
Sep 06 14:50:20 <kazzmir_>	ok
Sep 06 14:50:45 <tjaden>	does the scons build support out-of-source?
Sep 06 14:51:18 <kazzmir_>	what do you mean
Sep 06 14:51:33 <tjaden>	er, out of source builds
Sep 06 14:51:50 <tjaden>	keeping all the generated files in a separate directory
Sep 06 14:51:53 <kazzmir_>	yes
Sep 06 14:52:02 <kazzmir_>	everything generated by scons goes into scons_build
Sep 06 14:52:15 <kazzmir_>	and .sconf_temp/ for the configure stuff
Sep 06 14:52:18 <tjaden>	what about the alplatf.h, etc. files?
Sep 06 14:52:24 <kazzmir_>	those go into scons_build too
Sep 06 14:52:33 <tjaden>	great
Sep 06 14:52:40 <elias>	i guess we could make scons_build an option
Sep 06 14:52:46 <elias>	then i could specify a parent dir
Sep 06 14:52:51 <elias>	but still, duplicate=0 should work
Sep 06 14:52:52 <kazzmir_>	yea, that should work
Sep 06 14:53:07 <elias>	it's just bad design to copy files i think, no idea why they made it default
Sep 06 14:53:08 <mattymatt2>	why not the existing build dir?
Sep 06 14:53:21 <mattymatt2>	the same one the makefile uses
Sep 06 14:53:30 <elias>	we have no more makefiles as of today :)
Sep 06 14:53:39 <tjaden>	it still dumps files everywhere
Sep 06 14:53:41 <elias>	so i guess we can remove all the obj/ and lib/ folders
Sep 06 14:54:13 <elias>	tjaden: it shouldn't
Sep 06 14:54:22 <kazzmir_>	elias, ok i fixed the configure thing by using a cheap hack, all the alunixac.h files are always generated. that way scons wont erase them. i did it by doing Alias('all', alunixac)
Sep 06 14:54:27 <tjaden>	i meant the old makefiles would
Sep 06 14:54:35 <elias>	ah, ok
Sep 06 14:55:12 <kazzmir_>	but you still see 'using old settings' 4 times, which is slightly annoying
Sep 06 14:55:15 <CIA-31>	alleg: kazzmir * r10723 /allegro/branches/4.9/scons/ (all.scons unix.scons): make alunixac and alplatf always be generated to avoid multiple configure sessions
Sep 06 14:55:19 <mattymatt2>	but still, shouldn't scons_build and cmake use the same obj dir?
Sep 06 14:55:21 <kazzmir_>	and the addons keep reconfiguring themselves.. ill fix that
Sep 06 14:55:46 <elias>	mattymatt2: with cmake you specify whatever you want
Sep 06 14:56:07 <kazzmir_>	mattymatt2, mm, different build systems do things differently. besides, what does it gain us?
Sep 06 14:56:10 <elias>	and with scons we can easily get the same, it's just hardcoded at the top of the SConsctruct right now
Sep 06 14:57:04 <elias>	kazzmir_: yeah, i guess a non-hackish solution would be to only have the config checks for the builds it actually does
Sep 06 14:57:07 <mattymatt2>	gain is a simpler, more regular dir structure, whichever build system is used
Sep 06 14:57:30 <elias>	since it's out-of-source builds, doesn't matter
Sep 06 14:57:52 <mattymatt2>	completely outside the allegro dir?
Sep 06 14:58:00 <elias>	yeah
Sep 06 14:58:12 <elias>	allegro dir isn't touched
Sep 06 14:58:25 <mattymatt2>	ah cool. so it can build straight from CD or SVN pristine
Sep 06 14:58:25 <elias>	that is, i always make a sub-folder
Sep 06 14:58:35 <elias>	yeah
Sep 06 14:58:49 <tjaden>	kazzmir_ , i get AttributeError: SConsEnvironment instance has no __call__ method:
Sep 06 14:58:54 <mattymatt2>	I'm placated and millified :)
Sep 06 14:59:02 <mattymatt2>	^mollified
Sep 06 14:59:27 <kazzmir_>	tjaden, whats the stack trace
Sep 06 14:59:39 <tjaden>	  File "/home/tjaden/ahack/ws_4.9/SConstruct", line 469:
Sep 06 14:59:39 <tjaden>	    buildNormal()
Sep 06 14:59:39 <tjaden>	  File "/home/tjaden/ahack/ws_4.9/SConstruct", line 458:
Sep 06 14:59:39 <tjaden>	    doBuild(AllegroContextNormal())
Sep 06 14:59:53 <kazzmir_>	the error is at the bottom instead of the top
Sep 06 15:00:20 <kazzmir_>	i mean whats at the bottom of the stack trace
Sep 06 15:00:39 <tjaden>	i'll just make sure i've got the latest
Sep 06 15:02:08 <kazzmir_>	oh yea i got the same error just now
Sep 06 15:02:09 <elias>	it still runs config 4 times here after svn up :/
Sep 06 15:02:29 <kazzmir_>	the first time it will
Sep 06 15:02:31 <tjaden>	http://pastebin.com/m36f83426
Sep 06 15:02:34 <elias>	hm
Sep 06 15:02:34 <kazzmir_>	but after that it should only do i tonce
Sep 06 15:02:53 <elias>	well, i think eventually it should get a proper fix
Sep 06 15:02:57 <kazzmir_>	yea..
Sep 06 15:03:01 <kazzmir_>	hold on.. phone call
Sep 06 15:03:06 <elias>	the env.Config should only be run if the corresponding build is going to happen
Sep 06 15:04:20 <elias>	anyway, for 4.9.4, i guess it's good enough - can recommend cmake for now
Sep 06 15:05:21 <CIA-31>	alleg: tjaden * r10724 /allegro/branches/4.9/stamp-h.in: Delete stamp-h.in, used for autotools.
Sep 06 15:05:26 <tjaden>	4.9 looks in pretty good shape
Sep 06 15:08:37 <tjaden>	we don't seem to call allegro_exit any more
Sep 06 15:10:25 <tjaden>	because al_init doesn't atexit(allegro_exit)
Sep 06 15:14:41 <elias>	guess it should
Sep 06 15:14:48 <elias>	and allegro_exit should be renamed al_exit
Sep 06 15:15:04 <tjaden>	i'm not sure about atexit
Sep 06 15:15:17 <tjaden>	we had to do funny things with that in A4, probably for windows
Sep 06 15:15:22 <elias>	hm
Sep 06 15:16:33 <elias>	i guess the only critical thing would be fullscreen mode
Sep 06 15:17:17 <elias>	maybe we make a convenience macro again
Sep 06 15:17:36 <elias>	#define al_init() real_al_init; atexit(al_exit);
Sep 06 15:17:41 <elias>	then it should work everywhere
Sep 06 15:18:05 <tjaden>	yeah, something like that
Sep 06 15:18:39 <elias>	i think al_init should get a macro again anyway, so we can do that compile-time vs link-time version checks
Sep 06 15:18:52 <tjaden>	yes
Sep 06 15:19:03 <tjaden>	although hopefully there is less need for that
Sep 06 15:19:54 <elias>	i think so, yeah
Sep 06 15:20:11 <elias>	only 5.0.0 and 5.0.1 will be compatible, but not 5.0.0 and 5.1.0
Sep 06 15:20:34 <tjaden>	well that's the same as A4
Sep 06 15:20:41 <elias>	hm, true
Sep 06 15:29:45 <tjaden>	i'm going through the changelog again, and only a third of the way through. this one is big
Sep 06 15:31:22 <kazzmir_>	tjaden, I fixed that issue with scons.. will commit in a sec
Sep 06 15:41:00 <CIA-31>	alleg: kazzmir * r10725 /allegro/branches/4.9/ (3 files in 3 dirs): fix some scons build issues
Sep 06 15:42:46 <kazzmir_>	do we still need to build the x11 modules?
Sep 06 15:43:40 <elias>	no, right now all modules are for A4
Sep 06 15:43:53 <kazzmir_>	ok
Sep 06 15:44:05 ---	milan_911 is now known as mimix
Sep 06 15:46:10 <tjaden>	does ex_resize2 work fine on windows?
Sep 06 15:46:11 <CIA-31>	alleg: kazzmir * r10726 /allegro/branches/4.9/scons/all.scons: dont duplicate library source in scons build
Sep 06 15:46:46 <mimix>	hello
Sep 06 15:46:47 <trentg>	tjaden, I think so
Sep 06 15:47:25 <tjaden>	in x, we didn't resize the backbuffer on window resize
Sep 06 15:47:35 <trentg>	D3D always did that
Sep 06 15:47:41 <tjaden>	what was the emergency, mimix?
Sep 06 15:47:59 <tjaden>	ok, good. wgl?
Sep 06 15:48:09 <elias>	kazzmir_: still does 4 configs here, and every time
Sep 06 15:48:10 <mimix>	my cousin's bike broke... in another country :)
Sep 06 15:48:14 <kazzmir_>	elias, yea..
Sep 06 15:48:27 <kazzmir_>	im getting there ;)
Sep 06 15:48:49 <elias>	ah, ok, great :)
Sep 06 15:48:58 <mimix>	resizing a wgl window doesn't work right
Sep 06 15:50:17 <mimix>	i didn't try recently, but i think i didn't implement it right
Sep 06 15:51:16 <kazzmir_>	afk, be back soon..
Sep 06 15:51:45 <trentg>	It seems to work with wgl except the color is wrong
Sep 06 15:51:58 <trentg>	Look b&w almost
Sep 06 15:56:11 <trentg>	Seems to be when using a memory bitmap with opengl the colors are off
Sep 06 15:56:14 <trentg>	*wgl
Sep 06 15:56:36 <trentg>	e.g. ex_bitmap works as is but if I make mysha a memory bitmap the colors are wrong
Sep 06 16:00:06 <mimix>	i can't try or look into it now
Sep 06 16:00:28 <trentg>	I think one of the components is swapped... maybe red<->green
Sep 06 16:13:18 <mimix>	hm i did change something
Sep 06 16:13:49 <CIA-31>	alleg: tjaden * r10727 /allegro/branches/4.9/docs/src/changes-4.9.txt: Extend changelog for 4.9.4 a bit. Reorder some things.
Sep 06 16:15:10 <mimix>	can we have a 4.4.0rc1 one of these days?
Sep 06 16:15:38 <tjaden>	sure
Sep 06 16:16:01 <mimix>	macosx-universal is almost there
Sep 06 16:16:15 <tjaden>	we have to decide on djgpp
Sep 06 16:16:28 <tjaden>	i assume we drop it
Sep 06 16:16:40 <mimix>	we leave it as it is
Sep 06 16:16:46 <tjaden>	does it work?
Sep 06 16:16:58 <mimix>	dunno, did anyone try?
Sep 06 16:17:10 <mimix>	if not, i assume it works :)
Sep 06 16:17:40 <tjaden>	there's that no-asm issue
Sep 06 16:17:55 <tjaden>	elias and i got bored trying to compile it
Sep 06 16:18:07 <tjaden>	in qemu
Sep 06 16:20:06 <elias>	it doesn't work
Sep 06 16:20:09 <mimix>	the guy on a.cc was trying to make it work on WinXP
Sep 06 16:20:25 <elias>	the C only version doesn't seem to be fully implemented for djgpp
Sep 06 16:20:48 <elias>	seems some kind of "bank switcher" has no C implementation.. or at least that was my suspicion before I gave up with qemu
Sep 06 16:21:03 <elias>	and the asm version also isn't fully there any longer
Sep 06 16:21:13 <mimix>	people recommend using 4.0.3 for DOS anyway
Sep 06 16:21:24 <elias>	ever since we did the USE_C and NO_ASM unification
Sep 06 16:21:35 <elias>	they actually had some slightly different meaning in the djgpp version it seems :P
Sep 06 16:22:05 <elias>	yeah
Sep 06 16:23:37 <--	jonrafkind has quit (Connection timed out)
Sep 06 16:23:51 <tjaden>	ok, so we just say it's broken and waiting for someone to fix it
Sep 06 16:24:24 -->	jonrafkind (n=jon@c-67-186-250-110.hsd1.co.comcast.net) has joined #allegro-dev
Sep 06 16:24:25 ---	ChanServ gives voice to jonrafkind
Sep 06 16:25:18 <elias>	one of the examples using a 320x200 mode even works with C-only in qemu
Sep 06 16:25:43 <elias>	(likely it needs no "bank switching")
Sep 06 16:25:49 <tjaden>	yup
Sep 06 16:26:07 <trentg>	R<->B are swapped in wgl
Sep 06 16:26:13 <trentg>	With the default format
Sep 06 16:27:31 <trentg>	So it looks like the endian issue again
Sep 06 16:27:46 <elias>	endian issue?
Sep 06 16:28:05 <trentg>	Where allegro uses 0xAARRGGBB and opengl uses 0xAABBGGRR
Sep 06 16:28:39 <elias>	ah, yes, different pixel formats
Sep 06 16:28:50 <elias>	but it should do conversion
Sep 06 16:29:21 <elias>	hm, and 0xAABBGGRR is called RGBA in OpenGL :P
Sep 06 16:30:03 <trentg>	Ya, so I think maybe the wgl driver is using RGBA where it should be using BGRA
Sep 06 16:30:10 <CIA-31>	alleg: tjaden * r10728 /allegro/branches/4.9/readme_a5.txt: Update readme a bit.
Sep 06 16:30:54 <trentg>	You wouldn't notice unless using memory bitmaps because they would both still be the same
Sep 06 16:31:56 <elias>	does allegro.log tell which format WGL is using?
Sep 06 16:32:10 <jonrafkind>	elias, acodec does getOption('ogg', False), shouldnt that be True ?
Sep 06 16:32:21 <jonrafkind>	the default should be use whats available I would think
Sep 06 16:32:31 <trentg>	elias, it says r8 g8 b8 a0
Sep 06 16:32:52 <elias>	jonrafkind: yes, i guess
Sep 06 16:32:56 <jonrafkind>	ok
Sep 06 16:32:57 <trentg>	er.. actually it goes on with a bunch
Sep 06 16:33:28 <trentg>	ya, that's the one it choses
Sep 06 16:33:32 <elias>	i wonder if WGL really uses 24-bit formats, i always thought they are inefficient
Sep 06 16:33:49 <elias>	might be a general conversion error in the opengl code then
Sep 06 16:33:57 <elias>	i doubt anything in X11 ever uses 24-bit formats
Sep 06 16:34:32 <elias>	so a mistake there would have gone undetected
Sep 06 16:34:54 <elias>	need to make an ex_conversion, testing all possible memory<->memory and memory<->video conversions
Sep 06 16:36:26 <trentg>	Anyway, I need to get some supper ready.. is the scons build ready for release?
Sep 06 16:36:36 <elias>	what does WGL say for color size?
Sep 06 16:36:38 <elias>	24 or 32?
Sep 06 16:36:46 <elias>	seems to be a different field from the component sizes
Sep 06 16:36:50 <trentg>	32
Sep 06 16:36:53 <elias>	ah, ok
Sep 06 16:37:05 <elias>	so it uses ALLEGRO_PIXEL_FORMAT_XBGR_8888
Sep 06 16:39:05 <trentg>	Hm, it seems that conversion is off in convert.h
Sep 06 16:40:00 <trentg>	It didn't change anything though
Sep 06 16:40:13 <trentg>	Maybe I got the memory bitmap format wrong
Sep 06 16:41:39 <trentg>	That one looks right
Sep 06 16:44:44 <jonrafkind>	elias, things work if I use a global settings.cfg
Sep 06 16:44:53 <jonrafkind>	right now I have it in root/ but I can put it somewhere
Sep 06 16:44:55 <jonrafkind>	probably scons_build
Sep 06 16:45:09 <jonrafkind>	"work" == doesnt reconfigure 4 times
Sep 06 16:46:08 <elias>	#define ALLEGRO_CONVERT_RGB_888_TO_XBGR_8888(p) \
Sep 06 16:46:08 <elias>		ALLEGRO_SHIFT_CONVERT(p, \
Sep 06 16:46:08 <elias>			0xFF000000, 0, 0, \
Sep 06 16:46:08 <elias>			0xFF0000, 0, 8, \
Sep 06 16:46:08 <elias>			0x00FF00, 8, 0, \
Sep 06 16:46:09 <elias>			0x0000FF, 24, 0)
Sep 06 16:46:32 <elias>	is that right?
Sep 06 16:46:58 <elias>	and it's assuming memory bitmaps would use RGB_888 thought, which i guess they wouldn't with an XBGR_8888 mode...
Sep 06 16:47:25 <trentg>	Looks wrong, but memory bitmaps would use XRGB_8888
Sep 06 16:47:33 <elias>	yeah
Sep 06 16:47:56 <elias>	jonrafkind: sounds good
Sep 06 16:52:56 <jonrafkind>	ok things look good, now I need to memoize the results so it doesnt read the .cfg files a billion times
Sep 06 16:57:05 <mimix>	i made wgl recognize more formats so it might choose something different than it did before
Sep 06 16:58:51 <CIA-31>	alleg: kazzmir * r10729 /allegro/branches/4.9/ (5 files in 5 dirs): reuse previous configured headers
Sep 06 17:00:45 <mimix>	what about al_create_display() failing because of the default display color format?
Sep 06 17:01:14 <mimix>	default is ANY_NO_ALPHA i think
Sep 06 17:01:51 <mimix>	i know that adding ALLEGRO_COLOR_FORMAT_ANY would be a pain
Sep 06 17:03:45 <--	mimix has quit ("need sleep")
Sep 06 17:08:52 <jonrafkind>	whats icodec?
Sep 06 17:10:47 <elias>	alternative to iio
Sep 06 17:10:59 <elias>	it has some more formats, but also more dependencies
Sep 06 17:11:04 <elias>	i.e. uses imagemagick
Sep 06 17:11:54 <jonrafkind>	ok, scons doesnt handle it right now
Sep 06 17:11:57 <jonrafkind>	is there an example for it/
Sep 06 17:12:05 <CIA-31>	alleg: kazzmir * r10730 /allegro/branches/4.9/ (6 files in 5 dirs): memoize results of settings in scons. dont print message about re-using old settings
Sep 06 17:12:58 <--	SiegeLord has quit (Remote closed the connection)
Sep 06 17:13:04 <tjaden>	so what about a schedule?
Sep 06 17:14:40 -->	SiegeLord (i=sl@cpe-69-202-73-187.twcny.res.rr.com) has joined #allegro-dev
Sep 06 17:14:40 ---	ChanServ gives voice to SiegeLord
Sep 06 17:14:57 <tjaden>	i guess there's just this wgl resize and scons build fixes, and they can be fixed before next weekend?
Sep 06 17:15:27 <trentg>	The wgl fix should be easy enough I think... but I don't know my way around the wgl driver yet
Sep 06 17:15:55 <trentg>	Next weekend should be enough time
Sep 06 17:15:58 <jonrafkind>	scons is almost done
Sep 06 17:16:11 <tjaden>	so maybe even mid-week
Sep 06 17:16:25 <tjaden>	and i think osx is relatively up to date this time
Sep 06 17:16:41 <jonrafkind>	i thought peter said a bunch of things dont work in osx
Sep 06 17:16:53 <tjaden>	joystick and fullscreen, i think
Sep 06 17:17:23 <trentg>	Better than last release though
Sep 06 17:17:37 <trentg>	And it is a wip release still
Sep 06 17:18:13 <trentg>	But who knows, maybe peterhull can have one or both of those working by next weekend
Sep 06 17:18:27 <trentg>	Fullscreen is nearly working
Sep 06 17:18:27 <jonrafkind>	isnt he gone for a week?
Sep 06 17:18:33 <trentg>	Oh, I don't know
Sep 06 17:20:32 <CIA-31>	alleg: kazzmir * r10731 /allegro/branches/4.9/ (SConstruct scons/unix.scons): scons: keep unix settings in a global directory
Sep 06 17:21:38 <jonrafkind>	so i wonder how should installing work
Sep 06 17:22:43 <elias>	well, we need 4 targets as well
Sep 06 17:23:04 <jonrafkind>	but they should all go in /usr/include/allegro
Sep 06 17:23:15 <jonrafkind>	i mean there wont be a seperate directory for the static header files, right?
Sep 06 17:23:15 <elias>	there is only one version of includes
Sep 06 17:23:20 <elias>	no
Sep 06 17:23:25 <elias>	they should not differ 
Sep 06 17:23:43 <elias>	so only need it for libraries
Sep 06 17:23:47 <elias>	install-headers
Sep 06 17:23:54 <elias>	and install-static-debug
Sep 06 17:24:07 <elias>	and they all depend on install-headers?
Sep 06 17:24:10 <jonrafkind>	hm...
Sep 06 17:24:14 <jonrafkind>	but what about the addons
Sep 06 17:24:32 <elias>	they should also have 4 versions i guess
Sep 06 17:24:36 <jonrafkind>	i mean the headers
Sep 06 17:24:37 <elias>	not sure how it's done in cmake right now
Sep 06 17:24:44 <jonrafkind>	the addons should specify to install their own headers
Sep 06 17:24:52 <elias>	hm
Sep 06 17:24:55 <elias>	yeah
Sep 06 17:25:03 <jonrafkind>	so you cantj ust have a global install-headers for everything
Sep 06 17:25:19 <jonrafkind>	it would be nice for unix.scons to specify the install-headers target on unix
Sep 06 17:25:26 <elias>	can't that addon callback we have add them to the list of headers to install with isntall-headers?
Sep 06 17:26:14 <jonrafkind>	yea I think we should have some structured object that each SConscript returns
Sep 06 17:26:22 <jonrafkind>	which could contain a number of things, but also the installed headers
Sep 06 17:26:29 <jonrafkind>	and the main SConstruct will filter out duplicates
Sep 06 17:26:58 <jonrafkind>	what is snd_pcm_drop ?
Sep 06 17:27:34 <jonrafkind>	oh alsa
Sep 06 17:31:06 <jonrafkind>	brb
Sep 06 17:31:25 <CIA-31>	alleg: kazzmir * r10732 /allegro/branches/4.9/addons/ (acodec/SConscript iio/SConscript kcm_audio/SConscript): scons: add extra libraries
Sep 06 17:31:34 <tjaden>	in ex_threads2, in release builds, do you get all blue-green palettes?
Sep 06 17:31:47 <tjaden>	i think the red channel is missing then
Sep 06 17:32:54 <elias>	maybe it uses the one color format we discovered a wrong conversion for earlier
Sep 06 17:32:58 <trentg>	I get a red channel
Sep 06 17:33:02 <elias>	it was some 24->32 conversion
Sep 06 17:33:27 <tjaden>	yes, could be
Sep 06 17:34:04 <elias>	hm, scons still configures 4 times here
Sep 06 17:34:39 <CIA-31>	alleg: trentg * r10733 /allegro/branches/4.9/include/allegro5/convert.h: Fixed some wrong color conversions
Sep 06 17:36:03 <elias>	and then fails to build
Sep 06 17:36:24 <elias>	addons/kcm_audio/openal.c:15:16: error: al.h: No such file or directory
Sep 06 17:36:37 <elias>	guess something wrong with the openal configure check now
Sep 06 17:38:02 <tjaden>	hmm, still blue-green. back later
Sep 06 17:39:28 <elias>	sleep-time for me
Sep 06 17:39:32 <--	elias has quit ("Client exiting")
Sep 06 17:39:58 <CIA-31>	alleg: trentg * r10734 /allegro/branches/4.9/examples/ (10 files): Removed redundant al_show_mouse_cursor calls
Sep 06 17:40:39 <CIA-31>	alleg: trentg * r10735 /allegro/branches/4.9/CMakeLists.txt: Fixed install prefix for MinGW
Sep 06 17:41:05 <CIA-31>	alleg: trentg * r10736 /allegro/branches/4.9/cmake/FileList.cmake: Install tls.h
Sep 06 17:46:53 -->	jon_ (n=jon@wireless576.wireless.utah.edu) has joined #allegro-dev
Sep 06 17:46:53 ---	ChanServ gives voice to jon_
Sep 06 17:47:58 <--	jonrafkind has quit (Connection timed out)
Sep 06 17:52:51 ---	jon_ is now known as jonrafkind
Sep 06 17:55:36 <--	SiegeLord has quit (Read error: 104 (Connection reset by peer))
Sep 06 17:58:29 -->	SiegeLord (n=sl@cpe-69-202-73-87.twcny.res.rr.com) has joined #allegro-dev
Sep 06 17:58:30 ---	ChanServ gives voice to SiegeLord
Sep 06 18:12:25 <CIA-31>	alleg: kazzmir * r10737 /allegro/branches/4.9/SConstruct: scons: bump version to 4.9.4 and clean up dead code
Sep 06 18:32:07 <goalieca>	hmm. i guess i missed all the excitement
Sep 06 18:33:45 <goalieca>	and my logs only go back to 1:42pm
Sep 06 18:33:49 <goalieca>	stupid apple irc client
Sep 06 18:34:18 <trentg>	I can post my logs, peterhull wanted them anyway
Sep 06 18:34:47 <goalieca>	cool. 

Previous "hack days"