The Allegro Wiki is migrating to github at https://github.com/liballeg/allegro_wiki/wiki

Allegro Hack Day/2007 November 3

From Allegro Wiki
Jump to: navigation, search
Nov 03 13:12:54 -->	tjaden (n=tjaden@92.71.96.58.exetel.com.au) has joined #allegro-dev
Nov 03 13:12:54 ---	ChanServ gives voice to tjaden
Nov 03 13:13:12 <elias>	<trentg> Shall we start?
Nov 03 13:13:12 <elias>	<leverton> I'm here
Nov 03 13:13:12 <elias>	<trentg> I don't have much
Nov 03 13:13:12 <elias>	<elias> oh, i'm late
Nov 03 13:13:12 <elias>	<trentg> We should merge newgfx to 4.9 now
Nov 03 13:13:12 <elias>	<elias> i concur
Nov 03 13:13:18 <elias>	tjaden: that's all you missed so far :)
Nov 03 13:13:32 <tjaden>	hi, thanks
Nov 03 13:15:28 <trentg>	No idea how to do it myself
Nov 03 13:16:08 <tjaden>	elias, i tried the newgfx branch on my machine
Nov 03 13:16:13 <tjaden>	it doesn't abort on the glx stuff anymore
Nov 03 13:16:20 <tjaden>	but the graphics don't show properly
Nov 03 13:16:39 <elias>	not at all?
Nov 03 13:17:17 <tjaden>	not really
Nov 03 13:17:30 <tjaden>	let me just run them again
Nov 03 13:17:40 <elias>	could still be something to do with GLX
Nov 03 13:17:51 <tjaden>	exnew_bitmap is just plain white
Nov 03 13:18:02 <elias>	i only have glx 1.4 here.. and only tried to make it 1.2 compatible according to the docs
Nov 03 13:18:26 <tjaden>	ok. i haven't tried to fix it yet
Nov 03 13:18:37 <trentg>	It mostly worked for me, except resizing
Nov 03 13:19:36 <elias>	here both exnew_bitmap and exnew_resize work
Nov 03 13:19:49 <elias>	but yes, the driver needs a lot of work
Nov 03 13:20:07 <elias>	so far, it is only OpenGL 1.0 without extension, and only supports ABGR8888 format
Nov 03 13:20:11 -->	peterhull (n=peterhul@cpc1-wilm1-0-0-cust604.bagu.cable.ntl.com) has joined #allegro-dev
Nov 03 13:20:11 ---	ChanServ gives voice to peterhull
Nov 03 13:20:29 <elias>	hi peterhull 
Nov 03 13:20:34 <peterhull>	hi!
Nov 03 13:20:38 <peterhull>	Is it today?
Nov 03 13:20:52 <elias>	yes
Nov 03 13:21:12 <peterhull>	have I missed much, sorry I'mlate
Nov 03 13:21:16 <tjaden>	exnewapi i get a black square on white bg instead of red on black
Nov 03 13:21:16 <tjaden>	any ideas?
Nov 03 13:21:37 <elias>	i get black on white as well
Nov 03 13:22:00 <elias>	because i still haven't implemented blending for anything but draw_*_image
Nov 03 13:22:10 <elias>	(or so I thought, haven't checked in detail yet)
Nov 03 13:22:17 <tjaden>	oh, ok
Nov 03 13:22:25 <tjaden>	hi
Nov 03 13:23:37 <elias>	eventually, all of AllegroGL should somehow go into this.. but I have no idea about the best way to do that
Nov 03 13:24:02 <peterhull>	Is exnewapi the best example to get up and running?
Nov 03 13:24:24 <trentg>	It covers more than the others
Nov 03 13:24:35 <peterhull>	I've compiled allegro 4.9-newgfx for OSX, but I need to test it
Nov 03 13:24:44 <elias>	all the exnew*.c are good tests
Nov 03 13:24:52 <peterhull>	exnew_events seems to be a different API
Nov 03 13:24:52 <elias>	and we really need to make more like those
Nov 03 13:25:09 <trentg>	exnew_events doesn't work
Nov 03 13:25:16 <peterhull>	hmm I'm out of date
Nov 03 13:26:01 <tjaden>	it might be better to merge 4.9 into 4.9-newgfx instead of the other way
Nov 03 13:26:02 <elias>	OSX also needs a graphics driver
Nov 03 13:26:34 <peterhull>	elias: I'm working on it, ever so slowly.
Nov 03 13:26:59 <peterhull>	I would like to see Allegro continue on OS X but I haven't a lot of time
Nov 03 13:27:02 <leverton>	Thomas Harte made a comment recently about an OS X driver
Nov 03 13:27:10 <leverton>	might want to see if he'll work with you
Nov 03 13:27:13 <elias>	ah, good. eventually, part of the OpenGL drivers likely can be shared on linux/windows/osx
Nov 03 13:27:45 <elias>	but the main thing in the GLX driver is all the X11 and GLX stuff
Nov 03 13:27:58 <peterhull>	I did have a version that would get an OpenGL context up and running, and deliver events, but that might have "bit-rotted" now
Nov 03 13:28:22 <peterhull>	If I can do that, then the actual drawing code can be shared with Linux as you say, elias
Nov 03 13:28:50 <peterhull>	matthew, have you got a reference for the TH comment?
Nov 03 13:29:15 <leverton>	in his leopard thread
Nov 03 13:29:16 <leverton>	just today
Nov 03 13:29:35 <peterhull>	Should exnew_events be ignored or is it worth updating?
Nov 03 13:29:37 <leverton>	http://www.allegro.cc/forums/thread/593856/707995#target
Nov 03 13:29:53 <elias>	exnew_events should be updated eventually
Nov 03 13:30:06 <elias>	just wasn't our focus in the gfx branch
Nov 03 13:31:24 <elias>	so "CoreGraphics" can provide an OpenGL context?
Nov 03 13:31:50 <peterhull>	Not really. CoreGraphics isn't really the right thing to be using. 
Nov 03 13:32:06 <peterhull>	The terminology is a bit confusing (at least to me)
Nov 03 13:32:32 <peterhull>	CG is also called Quartz, it's more like windows GDI
Nov 03 13:33:06 <peterhull>	I'd use Quartz more for high quality (e.g. antialiased lines, beziers, that kind of thing)
Nov 03 13:33:35 <peterhull>	OpenGL is for performance
Nov 03 13:33:58 <peterhull>	But then there's Quartz Extreme which is Quarz on the OpenGL pipeline
Nov 03 13:34:22 <tjaden>	how good is EGL support on various platforms, does anyone know?
Nov 03 13:35:02 <peterhull>	As in OpenGL ES?
Nov 03 13:35:11 <tjaden>	yes
Nov 03 13:35:24 <tjaden>	er, well, i think yes :)
Nov 03 13:36:34 <peterhull>	AFAIK it's mostly a subset of the 'proper' OpenGL
Nov 03 13:37:13 <peterhull>	but I'm not sure how much is extra, fixed point maths is one thing in ES that's not in OpenGL
Nov 03 13:37:33 <tjaden>	apparently it's a platform independent GLX/WGL/AGL
Nov 03 13:38:03 <peterhull>	Oh, so not what I'm on about at all, then!
Nov 03 13:38:08 <elias>	got a link?
Nov 03 13:38:28 <tjaden>	http://www.freedesktop.org/wiki/EGL
Nov 03 13:38:42 <tjaden>	i guess it's very new, so unusable
Nov 03 13:39:00 <elias>	oh, looks nice though
Nov 03 13:39:29 <elias>	especially that you can draw to off-screen bitmaps
Nov 03 13:39:43 <elias>	opengl 1.0 doesn't allow that at all, which makes my current glx driver a bit stupid
Nov 03 13:41:12 <elias>	oh, or is that just another extension to EGL on that site
Nov 03 13:41:41 <tjaden>	looks like it
Nov 03 13:42:57 <tjaden>	so, 4.9 into newgfx or the other way around?
Nov 03 13:43:22 <elias>	hm
Nov 03 13:43:37 *	elias does an svn merge both ways and looks which produces more output..
Nov 03 13:46:25 <elias>	hm, svn is slow for this
Nov 03 13:46:42 <tjaden>	peterhull, while you're here. the last time i tried to merge 4.2 changes into 4.9, the thing getting in the way was the osx port
Nov 03 13:47:04 <tjaden>	i don't really know which changes to keep and whatnot
Nov 03 13:47:23 <tjaden>	and i can't test if the thing still compiles after the merge
Nov 03 13:48:29 <tjaden>	mind if i just let it break?
Nov 03 13:48:50 <elias>	hm.. "Skipped missing target" what does that mean again
Nov 03 13:49:00 <tjaden>	renamed file
Nov 03 13:49:10 <tjaden>	or deleted
Nov 03 13:50:21 <elias>	i wonder if i swapped parameters to merge
Nov 03 13:52:22 <--	peterhull has quit ()
Nov 03 13:53:50 <elias>	ok, so i guess merging from 4.9 to 4.9-newgfx is better
Nov 03 13:54:37 <elias>	should i try doing it? or does your svk copy make it easier somehow?
Nov 03 13:54:54 <tjaden>	lemme check
Nov 03 13:56:24 <tjaden>	yup, easy here. only 3 conflicts
Nov 03 13:57:04 -->	peterhull (n=peterhul@cpc1-wilm1-0-0-cust604.bagu.cable.ntl.com) has joined #allegro-dev
Nov 03 13:57:04 ---	ChanServ gives voice to peterhull
Nov 03 13:57:21 <peterhull>	tjaden: yes, just break it and I'll submit a patch
Nov 03 13:57:33 <tjaden>	ok
Nov 03 13:57:47 <peterhull>	trentg: I have a question for you on tls.c
Nov 03 13:57:52 <trentg>	ok
Nov 03 13:58:28 <peterhull>	There's no native support on OS X (using __thread) so I will use the pthread implementation
Nov 03 13:58:47 <trentg>	ok
Nov 03 13:59:01 <juvinious>	3's a charm
Nov 03 13:59:16 <peterhull>	so in tls.c you have (I think) the same thing twice, one for mingw and one for linux and msvc
Nov 03 13:59:34 <tjaden>	should i just do the merge now?
Nov 03 13:59:41 <trentg>	Well, one for old versions of MinGW and one for everything else
Nov 03 14:00:17 <peterhull>	ok, right
Nov 03 14:00:41 <elias>	tjaden: i'd say yes
Nov 03 14:01:01 <tjaden>	ok. i'll just give it a quick compile
Nov 03 14:01:27 <elias>	i managed to ruin both my checkouts due to wrong merge parameters :P
Nov 03 14:01:56 <peterhull>	trentg: sorry, I was working up to the question but I have to go now (baby stuff!) so can I contact you later
Nov 03 14:01:58 <peterhull>	:(
Nov 03 14:02:20 <trentg>	yes, you can pm me on allegro.cc or email me or write to the list
Nov 03 14:02:58 <peterhull>	And in general is is OK for me to work on just getting the OpenGL context set up, then I can borrow the actual drawing code from Linux?
Nov 03 14:03:26 <peterhull>	bye for now
Nov 03 14:03:29 <juvinious>	later
Nov 03 14:03:32 <--	peterhull has quit ()
Nov 03 14:04:38 <--	juvinious has quit ("Error! Abort!")
Nov 03 14:06:11 <trentg>	So, should we change AL_ prefixed things to ALLEGRO_?
Nov 03 14:06:36 <trentg>	From the input and events systems
Nov 03 14:07:07 <tjaden>	might as well
Nov 03 14:09:17 <tjaden>	some of those names are going to be unwieldly with a long prefix
Nov 03 14:09:24 <tjaden>	if they are not already :)
Nov 03 14:10:16 <trentg>	Yeah.. but remember last time when we tried to agree on a prefix? :)
Nov 03 14:10:39 <elias>	we sort of set to use ALLEGRO_ "for now" I think
Nov 03 14:10:41 <elias>	*said
Nov 03 14:10:56 <trentg>	Yeah, we could always change it
Nov 03 14:11:20 <elias>	for the font addon, you use a different prefix, right?
Nov 03 14:11:31 <trentg>	Yes, A5FONT_
Nov 03 14:11:34 <tjaden>	btw, why's the font addon separate?
Nov 03 14:11:59 <trentg>	Not sure, we just decided to separate it.. maybe we shouldn't though.
Nov 03 14:12:06 <tjaden>	is it just the a4 font api?
Nov 03 14:12:14 <trentg>	Yes
Nov 03 14:12:20 <trentg>	More or less
Nov 03 14:12:34 <tjaden>	i don't mind, just curious
Nov 03 14:13:18 -->	juvinious (n=juviniou@c-65-34-141-193.hsd1.fl.comcast.net) has joined #allegro-dev
Nov 03 14:13:18 ---	ChanServ gives voice to juvinious
Nov 03 14:13:30 <tjaden>	maybe it's the first step towards the modular structure that some people wanted
Nov 03 14:14:26 <elias>	it should at least make it easy to just drop in/remove additional addons
Nov 03 14:14:38 <elias>	even if we won't distribute them separately
Nov 03 14:15:14 <tjaden>	memblit.c takes forever to compile on my machine
Nov 03 14:15:21 <elias>	same here :(
Nov 03 14:15:34 <tjaden>	do we know why?
Nov 03 14:15:44 <trentg>	I tried to make it faster, but only got about 1s difference
Nov 03 14:15:44 <elias>	yes, it compiles like 10000 functions
Nov 03 14:16:02 <elias>	i think we might want a version which is slightly slower on runtime, but uses less functions
Nov 03 14:16:05 <tjaden>	maybe we can drop the optimisation settings
Nov 03 14:16:14 <tjaden>	at least for development
Nov 03 14:17:01 <trentg>	There's already a slow version in there, just have to comment out the big messy part and change a couple lines
Nov 03 14:17:16 <elias>	maybe depending on #ifdef DEBUG
Nov 03 14:17:24 <elias>	so the debug version would compile fast
Nov 03 14:20:45 <trentg>	Yeah, will do that
Nov 03 14:24:12 <tjaden>	commiting the merge now
Nov 03 14:24:40 <tjaden>	i might not have resolved the conflicts in scons files correctly
Nov 03 14:27:21 <elias>	the scons files in newgfx should have been the up-to-date ones (also kazzmir worked on those lately)
Nov 03 14:27:44 ---	#allegro-dev :You need to be a channel operator to do that
Nov 03 14:27:48 <tjaden>	newgfx didn't have my changes to rename all the HAVE_ constants
Nov 03 14:27:57 <elias>	ah
Nov 03 14:28:00 <tjaden>	it's finally committed
Nov 03 14:28:05 *	elias kicks CIA-27 
Nov 03 14:28:06 <CIA-27>	ow
Nov 03 14:28:11 <elias>	hm, useless both
Nov 03 14:28:11 <juvinious>	shouldn't cia be reporting that :P
Nov 03 14:28:15 <elias>	*bot
Nov 03 14:28:17 <elias>	yes, it shoudl
Nov 03 14:28:37 <juvinious>	it seems like it's been flaky lately
Nov 03 14:28:44 <elias>	yeah
Nov 03 14:28:53 <elias>	in wesnoth, it also reports only about every other commit
Nov 03 14:29:05 <elias>	here it doesn't report at all though
Nov 03 14:29:32 <tjaden>	you work on wesnoth?
Nov 03 14:29:50 <elias>	only some very minor things, mainly connected to python scripting
Nov 03 14:30:04 <tjaden>	i downloaded it but never got around to playing
Nov 03 14:30:09 <juvinious>	is tat a game?
Nov 03 14:30:36 <juvinious>	oh interesting looks like a tactics type game
Nov 03 14:31:09 <tjaden>	i saw chris robinson's name mentioned a few times in wine weekly news
Nov 03 14:31:39 <elias>	wesnoth is just interesting somehow.. Rusty Russel (linux kernel programmer) and Eric S. Raymond (some open source guy) are both major contributors
Nov 03 14:31:55 <juvinious>	neat
Nov 03 14:32:06 <elias>	yes, chris seems to be a wine dev now.. instead of doing our sound API :P
Nov 03 14:32:21 <juvinious>	chris? kitty?
Nov 03 14:32:26 <elias>	yes..
Nov 03 14:32:40 <tjaden>	and ages ago i saw eric botcazou's name on some GNAT papers
Nov 03 14:32:52 <elias>	GNAT?
Nov 03 14:33:00 <tjaden>	gnu ada compiler
Nov 03 14:33:03 <elias>	ahh
Nov 03 14:33:23 <elias>	i saw him on something gcc related a few times, when looking for googling for gcc bugs
Nov 03 14:34:07 <tjaden>	stalking allegro devs :)
Nov 03 14:34:17 <elias>	heh
Nov 03 14:34:29 <juvinious>	scary :P
Nov 03 14:34:48 <elias>	angelo mottola also was here a few weeks ago, he said he would work on the 4.9 OSX port if he would have any free time :)
Nov 03 14:35:12 <elias>	btw, what is the merged 4.9 called now?
Nov 03 14:35:21 <tjaden>	it's still newgfx
Nov 03 14:35:36 <elias>	ah, so i guess, should remove the current 4.9 then move newgfx to 4.9
Nov 03 14:35:53 <elias>	i think i'll try to get it as svk instead of svn
Nov 03 14:38:09 <tjaden>	what needs to be done for 4.2.3?
Nov 03 14:39:08 <elias>	"Retrieving log information from 1 to 9545".. wonder how long it will take
Nov 03 14:39:36 <tjaden>	it's probably a good idea to download the tarball of the repository...
Nov 03 14:42:00 <elias>	i didn't really follow the 4.2.3 issue
Nov 03 14:42:07 <elias>	Evert and mimix know about it I guess
Nov 03 14:42:37 <elias>	it was windows only in any case
Nov 03 14:46:31 <tjaden>	i know about the calling conventions but i think people wanted to sneak in more changes again
Nov 03 14:47:15 <tjaden>	ugh, i wrote the wrong revision numbers in my commit log and now i can't edit it
Nov 03 14:47:36 <tjaden>	svn: DAV request failed; it's possible that the repository's pre-revprop-change hook either failed or is non-existent
Nov 03 14:47:52 <elias>	i think SVN doesn't allow changing them
Nov 03 14:47:55 <elias>	er, SF
Nov 03 14:48:40 <tjaden>	bugger
Nov 03 15:02:38 <trentg>	Do we need to change internal stuff like AL_FUNC, AL_VAR etc?
Nov 03 15:03:03 <tjaden>	i hope not
Nov 03 15:03:12 <tjaden>	they are ugly enough as it is
Nov 03 15:04:27 <trentg>	hehe
Nov 03 15:07:05 <tjaden>	i think they are overkill as well
Nov 03 15:07:27 <tjaden>	don't change them for now
Nov 03 15:07:44 <tjaden>	it will make merging between branches impossible
Nov 03 15:08:12 <trentg>	Do you want me to wait to change the others as well?
Nov 03 15:08:52 <tjaden>	as long as it's restricted to a5 stuff, it's ok
Nov 03 15:13:11 <tjaden>	in debug mode: /home/tjaden/ahack/4.9-newgfx2/src/xglx/xbitmap.c:33: undefined reference to `_al_draw_bitmap_region_memory_fast
Nov 03 15:13:43 <trentg>	Ah, didn't see that..
Nov 03 15:14:18 <trentg>	You can replace it with _al_draw_bitmap_region_memory
Nov 03 15:17:20 <tjaden>	done
Nov 03 15:17:32 <elias>	i get an error about a missing icon.xpm
Nov 03 15:17:40 <elias>	i'll just add it from 4.3
Nov 03 15:17:47 <tjaden>	uh
Nov 03 15:17:53 <tjaden>	didn't you delete it?
Nov 03 15:17:57 <elias>	i did?
Nov 03 15:18:16 <tjaden>	oh, renamed
Nov 03 15:18:25 <elias>	yes, from alex to icon
Nov 03 15:18:35 <elias>	no idea why it's not in the 4.9 src/x folder..
Nov 03 15:18:45 <tjaden>	it is for me
Nov 03 15:18:50 <elias>	oh
Nov 03 15:18:57 <tjaden>	no, it's not in version control
Nov 03 15:19:00 <tjaden>	hmm
Nov 03 15:19:17 <elias>	hm, does cmake copy it there from somewhere else?
Nov 03 15:19:33 <tjaden>	not that i know
Nov 03 15:19:55 <elias>	find | grep xpm
Nov 03 15:20:02 <tjaden>	it got deleted in the merge but not added
Nov 03 15:20:04 <elias>	ohh, already found one big advantage of svk
Nov 03 15:20:14 <elias>	no more double seatch results inside .svn folders :)
Nov 03 15:20:18 <tjaden>	please add it back
Nov 03 15:20:21 <elias>	ok
Nov 03 15:20:38 <tjaden>	svn's handling of renames probably
Nov 03 15:21:44 <elias>	ah, so "svk ci" also resyncs everywthing with the mirrow
Nov 03 15:22:18 <tjaden>	yes, if you're committing to a mirrored path
Nov 03 15:23:21 <elias>	http://cia.vc/account/bots/3135/
Nov 03 15:23:28 <elias>	can someone set the bot to inactive?
Nov 03 15:23:35 <elias>	if i try, i only get Error 500
Nov 03 15:23:54 <elias>	(my hope is if we can get it to leave and rejoin, it will print log messages again)
Nov 03 15:25:32 <elias>	ah, wait, i guess it only works with an account
Nov 03 15:29:00 -->	dacap (n=dacap@190.7.30.219) has joined #allegro-dev
Nov 03 15:29:00 ---	ChanServ gives voice to dacap
Nov 03 15:29:54 <tjaden>	hi
Nov 03 15:30:10 <dacap>	hello everybody, there was a hack day today?
Nov 03 15:30:19 <dacap>	hi peter
Nov 03 15:30:24 <elias>	could say it is still on-going :)
Nov 03 15:30:36 <dacap>	:)
Nov 03 15:30:48 <dacap>	what are you talking about?
Nov 03 15:31:02 <tjaden>	that reminds me, do your key_shift changes need to be in 4.2.3?
Nov 03 15:31:30 >chanserv<	#allegro-dev op elias
Nov 03 15:31:55 >chanserv<	op #allegro-dev elias
Nov 03 15:31:56 ---	ChanServ gives channel operator status to elias
Nov 03 15:32:00 <--	elias has kicked CIA-27 from #allegro-dev (elias)
Nov 03 15:32:01 -->	CIA-27 (i=cia@208.69.182.149) has joined #allegro-dev
Nov 03 15:32:01 ---	ChanServ gives voice to CIA-27
Nov 03 15:32:12 >chanserv<	deop #allegro-dev elias
Nov 03 15:32:20 >chanserv<	help op
Nov 03 15:32:30 >chanserv<	op #allegro-dev -elias
Nov 03 15:32:30 ---	ChanServ removes channel operator status from elias
Nov 03 15:33:08 <dacap>	tjaden: yes
Nov 03 15:33:33 <dacap>	tjaden: e.g.: ASE must be compiled with Allegro 4.2.2 SVN right now for Windows
Nov 03 15:34:49 <tjaden>	the update_shifts call, and something else?
Nov 03 15:35:37 <dacap>	the old GetAsyncKeyState patch
Nov 03 15:36:05 <dacap>	I think that change is in SVN right now
Nov 03 15:36:45 <tjaden>	yes, i see it
Nov 03 15:37:25 <dacap>	So the "update_shifts" is the only one
Nov 03 15:37:39 <dacap>	that need to be checkin
Nov 03 15:37:54 <tjaden>	i'll do it now
Nov 03 15:38:31 <dacap>	ok, thanks peter
Nov 03 15:40:04 <dacap>	here comes my brother... ok, I'll see the log of the hack day tonight... good bye
Nov 03 15:40:09 <--	dacap (n=dacap@190.7.30.219) has left #allegro-dev
Nov 03 15:44:29 <elias>	so, can i "svn rm 4.9" and "svn mv 4.9-newgfx 4.9" ?
Nov 03 15:45:02 <trentg>	Can you wait a couple minutes?
Nov 03 15:45:18 <elias>	sure