--- Log opened Wed Sep 29 00:00:00 2010
20100929 00:00:05< Ivanovic> (as long as the speed for zooming out is reasonable in software mode)
20100929 00:00:17< Ivanovic> screw the storing stuff on disc part
20100929 00:00:30< Ivanovic> with the lowmem mode we directly *remove* about 95% of all images
20100929 00:00:52< Ivanovic> for units we just keep the base image
20100929 00:01:04< Ivanovic> for terrains we would only keep the base tile for most images
20100929 00:01:54< Ivanovic> beside this: the stuff is not in the nand for most devices, it is on sd cards and those are reasonably cheap that saving memory on disc does not concern me this much (especially if you compare the space used for images in trunk to the space used for music and sounds *on disc*)
20100929 00:03:14-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Quit: Leaving.]
20100929 00:04:14-!- grzywacz [~grzywacz@wesnoth/developer/grzywacz] has quit [Remote host closed the connection]
20100929 00:04:52< Ivanovic> GNUtoo|laptop: what do you think should be the max mem usage wesnoth should have for embedded devices
20100929 00:04:52< alink> I am more considering ram usage, having them scaled on disk, allow to load it in memory without scaling. If not scaled on disk, we can scale them at loading (slower loading). Or scale them at render time (more ram used, possible speed cost for rendering)
20100929 00:05:21< GNUtoo|laptop> Ivanovic, it depend on what you want to target
20100929 00:05:32< GNUtoo|laptop> but I bet the ultimate answer would be testing right?
20100929 00:05:36< Ivanovic> alink: i think that we can already significantly reduce memory usage with simply "removing" images
20100929 00:06:04< Ivanovic> GNUtoo|laptop: i don't know what common devices that should be able to run wesnoth (and where people want to and do run it) have as hardware
20100929 00:06:19< Ivanovic> what the resolution there is as well as the available ram
20100929 00:06:33< Ivanovic> it is clear that devices with only 64MB won't work
20100929 00:06:51< Ivanovic> having 128MB devices working would be lovely, how much memory is available on those?
20100929 00:06:55< alink> Ivanovic: of course but we loose content. The goal is to keep a maximum of it
20100929 00:07:07< GNUtoo|laptop> Ivanovic, 128 is the minimum theses days
20100929 00:07:16< GNUtoo|laptop> I'll look
20100929 00:07:37< Ivanovic> alink: but we should only keep content in a way that keeps things usable
20100929 00:07:56< Ivanovic> alink: imagine a mobile phone with 480x320 as screen resolution on 3"
20100929 00:08:07< Ivanovic> now you have to use your fingers for input
20100929 00:08:24< Ivanovic> hitting a 36x36 hex really sucks there, so you likely want things bigger
20100929 00:08:55< GNUtoo|laptop> note that some phones say they have 192M but only have 110M usable
20100929 00:09:02< alink> Ivanovic: I was talking about removing image like terrain transitions 
20100929 00:09:12< GNUtoo|laptop> because of memory sharing between the main cpu and some other devices cpu
20100929 00:09:34< CIA-35> esr * r46808 /trunk/data/ (4 files in 4 dirs): Fix variuous small glitches found by wmlscope/wmllint.
20100929 00:09:45< Ivanovic> alink: looking at what we currently got we can save most space with transitions
20100929 00:10:05< Ivanovic> alink: that is: on those high dpi devices those transitions are often less noticable anyway
20100929 00:10:38< alink> Ivanovic: yeah I never used these devices, so I have no real idea of the dpi
20100929 00:10:54< Ivanovic> alink: they are rather high
20100929 00:11:01< GNUtoo|laptop> indeed
20100929 00:11:22< GNUtoo|laptop> I need to go to sleep soon btw
20100929 00:11:39< Ivanovic> eg the pandora screen has about 9cm horizontally
20100929 00:11:46< Ivanovic> on those 9cm you get 800px
20100929 00:11:55< CIA-35> zookeeper * r46809 /trunk/data/campaigns/Under_the_Burning_Suns/ (2 files in 2 dirs): Added the Small Mudcrawler back to UtBS, but using the core Mudcrawler as a [base_unit].
20100929 00:12:17< Ivanovic> should be easy to do the math regarding how many mm a 36x36 hex would get
20100929 00:12:34< GNUtoo|laptop> with 128M GUI,telephony I've 40M/123M
20100929 00:12:36< GNUtoo|laptop> on htop
20100929 00:12:40< GNUtoo|laptop> s/on/in
20100929 00:12:54< GNUtoo|laptop> with no app launched
20100929 00:12:57< Ivanovic> phone displays have a little less resolution but tend to be smaller than 4" diameter
20100929 00:13:16< Ivanovic> GNUtoo|laptop: so, uhm, only 83MB available?
20100929 00:13:21< GNUtoo|laptop> yes
20100929 00:13:46< Ivanovic> not much ram at all
20100929 00:13:53-!- evilshadowmaster is now known as Shadowmaster
20100929 00:14:36< GNUtoo|laptop> 10% used by enlightenment
20100929 00:14:37< CIA-35> alink * r46810 /branches/ogl/configure.ac: Add OpenGL header and library in configure.ac
20100929 00:14:50< GNUtoo|laptop> 9,3% by some python stuff
20100929 00:15:09< GNUtoo|laptop> that are necessary for telephony
20100929 00:15:18< GNUtoo|laptop> 8,3% for the GUI daemon for the phone
20100929 00:15:39< GNUtoo|laptop> basically GUI+phone stuff
20100929 00:15:45< GNUtoo|laptop> eat all this mem
20100929 00:15:59< GNUtoo|laptop> on non-phone you can remove the telephony stuff
20100929 00:16:17-!- stikonas [debian-tor@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20100929 00:16:43-!- stikonas [debian-tor@wesnoth/translator/stikonas] has joined #wesnoth-dev
20100929 00:16:51< GNUtoo|laptop> if I remove telephony killing dbus
20100929 00:16:54< GNUtoo|laptop> it's 21M
20100929 00:16:57< GNUtoo|laptop> that is used
20100929 00:17:12-!- Blueblaze [~Blueblaze@adsl-76-202-19-44.dsl.hstntx.sbcglobal.net] has quit [Quit: Blueblaze]
20100929 00:20:45< GNUtoo|laptop> Ivanovic, I'll try to sleep now
20100929 00:20:52< Ivanovic> n8 GNUtoo|laptop
20100929 00:21:11< alink> GNUtoo|laptop: yes gn, thanks for the info
20100929 00:21:31< GNUtoo|laptop> I'll come back tomorrow
20100929 00:21:41< GNUtoo|laptop> so I could give more infos
20100929 00:21:45< GNUtoo|laptop> bye
20100929 00:21:48-!- GNUtoo|laptop [~gnutoo@host245-55-dynamic.180-80-r.retail.telecomitalia.it] has quit [Quit: Program received signal SIGSEGV, Segmentation fault.]
20100929 00:23:03< Ivanovic> alink: personally i'd really love to get rid of tinygui since this is a, uhm, imperfect solution
20100929 00:23:25< Ivanovic> if there are other ways that basically work as well i'd prefer using those and the opengl switch seems to *maybe* open up some
20100929 00:24:00< alink> yes me too, but OtOH it makes sense too, so the other way need to be good enough
20100929 00:24:18< Ivanovic> yes, it needs
20100929 00:25:15< Ivanovic> but currently i am of the opinon that since we already do in fact remove all animations and everything from units beside the base frame for those devices we could/should go the step further and remove more stuff that barely makes a difference which means terrain transitions
20100929 00:26:25< alink> I don't know, it's more an artist and device users decision. I am just a coder :)
20100929 00:29:13< Ivanovic> check this video starting at about 5:50: http://www.youtube.com/watch?v=zs4Z0It7mA4
20100929 00:29:21< Ivanovic> afterwards tell me if you think if it is usable or not
20100929 00:30:11-!- Shakey [~jbailey@12.190.80.225] has quit [Quit: Ha ha, charade you are.]
20100929 00:30:34< Ivanovic> considering that many terrains appear to be missing (villages and forests at least) i'd say that if we want to keep tinygui in the current way it does need some serious fixes
20100929 00:32:00< alink> yeah but you terrain are smooth, not sure how it could look with sharp hexes
20100929 00:32:55< alink> btw quick UI idea: with fast zoom, the view could zoom when selecting an unit and dezooming when deselecting it, or similar behavior
20100929 00:33:08< alink> I mean for these devices
20100929 00:33:21< Ivanovic> yeah, some interesting things would be possible
20100929 00:33:23< alink> or small on fight and things like that
20100929 00:33:34< alink> s/small/zoom
20100929 00:33:45< Ivanovic> you select a unit and it zooms out to show everything that you can reach with the unit
20100929 00:33:53< Ivanovic> (or zooms in if it is not far)
20100929 00:34:15< Ivanovic> anyway, i should head off to bed, t00
20100929 00:34:20< Ivanovic> s/t00/too
20100929 00:34:22< Ivanovic> n8
20100929 00:34:29< alink> and I should go back to my branch
20100929 00:34:33< alink> Ivanovic: gn
20100929 00:35:02< Ivanovic> alink: and if you want to see wesnoth on a high dpi / small screen device: i'll most likely be at fosdem with my pandora
20100929 00:35:04< Ivanovic> ;)
20100929 00:35:52< alink> everybody will have a pandora at that time
20100929 00:36:37< alink> this IPhone thing is so old
20100929 00:36:57< esr_> zookeeper: Who's doing the work on the new water terrains?
20100929 00:38:15-!- Aethaeryn [~Michael@wesnoth/umc-dev/developer/aethaeryn] has joined #wesnoth-dev
20100929 00:41:56-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20100929 00:42:27-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100929 00:43:53-!- noy_ [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100929 00:47:26-!- noy [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 264 seconds]
20100929 00:47:26-!- noy_ is now known as noy
20100929 00:47:52-!- noy [~Noy@wesnoth/developer/noy] has quit [Client Quit]
20100929 00:55:31< zookeeper> esr_, eleazzaar
20100929 00:56:18< esr_> zookeeper: I need to bitch at him about the new reef and coast terrains - Dead Water looks like shit.
20100929 00:56:50< zookeeper> well, go ahead, he's right here :p
20100929 00:57:00< esr_> eleazzaar: ping
20100929 00:57:57< zookeeper> not sure if he's usually around at this time, but at least he still was an hour ago.
20100929 01:06:12< eleazzaar> esr_: if you want to motivate me to consider your opinion of terrain art you might try a bit more tact
20100929 01:06:55< eleazzaar> there's also the possibility of reading my thread and finding out what's preliminary and what direction i'm taking
20100929 01:07:56< esr_> eleazzaar: What thread is that?
20100929 01:08:14< eleazzaar> http://forums.wesnoth.org/viewtopic.php?f=18&t=29691&p=458617#p458617
20100929 01:09:08< esr_> I mean, I am perfectly willing to credit that you're plannibfg something better, but I'm upset about DW because the maps used to be si pretty.
20100929 01:09:41< eleazzaar> this is a development version
20100929 01:09:47< Shadowmaster> esr_: since you probably haven't been told either.
20100929 01:09:51< Shadowmaster> DW has maintainers.
20100929 01:10:07< Shadowmaster> two maintainers to be exac t -- one of them is beetlenaut
20100929 01:10:15< esr_> I know beetlenaut is on it.
20100929 01:11:54< esr_> eleazzaar: Can we have shades of blue back for the coastal-reef terrain, please?  The present brown looks like toxic sludge.
20100929 01:13:02< eleazzaar> i doubt you are even looking at trunk
20100929 01:13:37< esr_> I did svn up recently.  I'll do it again.
20100929 01:13:54-!- alink [~alink@wesnoth/developer/alink] has quit [Remote host closed the connection]
20100929 01:14:45< eleazzaar> not that i think the reef ever looked like toxic sludge.  hyperbole doesn't strenghtne your case here
20100929 01:14:51< esr_> I have trunk version 46806.
20100929 01:15:15 * zookeeper is glad to be going to bed at this point
20100929 01:15:36-!- zookeeper [~l@wesnoth/developer/zookeeper] has quit []
20100929 01:16:35< Shadowmaster> the DW maps appear to be in need of an update.
20100929 01:16:49< esr_> eleazzaar: I'm looking at it now and it's an unpleasant purplish-brown, the color shift has made the "reef" things look like something nasty.
20100929 01:17:23< eleazzaar> what map are you looking at?
20100929 01:17:50< esr_> Invasion.  First scenario.
20100929 01:18:11< eleazzaar> i mean what's the map called so i can open in the map editor
20100929 01:18:24< esr_> Looking...
20100929 01:18:36< eleazzaar> home1?
20100929 01:19:24< esr_> Home_1.map
20100929 01:19:44< eleazzaar> nothing remotely purplish brown on my screen
20100929 01:19:54< esr_> ?
20100929 01:20:47< eleazzaar> either we aren't looking at the same thing, or you are bad at naming colors
20100929 01:20:50< eleazzaar> screenshot?
20100929 01:21:31< esr_> Qhat do the shallow-water hexes look like on your display? On mine they're a sort of purplish-brown with the ocean blue leaking through.
20100929 01:22:20< eleazzaar> esr_: do you know the proper names for the terrain?
20100929 01:23:07< esr_> When I mouse over it says Ford (Flat,Shallow Water)
20100929 01:23:26< eleazzaar> then why are you calling it "reef" and "shallow"?!?!
20100929 01:24:17< esr_> Shallow Water isn't "shallow"?  And the other terrain that looks awful on my display is Coastal Reef.
20100929 01:24:35< esr_> So in what way am I mislabeling these?
20100929 01:25:17< eleazzaar> ford != shallow water.  it has the move/def of flat and shallow water
20100929 01:25:34< eleazzaar> shallow water is it's own terrain
20100929 01:25:41< esr_> Oh, wait, I see.  OK< unmodified shallow water is flat blue, that's OK.
20100929 01:27:06< esr_> The ones that have the weird purplish-brown effect are Ford and Coastal Reef.  What do they look like on your display?  What are they supposed to look like?
20100929 01:29:40< eleazzaar> the water under the reef is the same as the medium shallow water since it is not worth making 3 reef terrains for a slight color difference.
20100929 01:30:12< eleazzaar> the ford is stand-in graphics
20100929 01:30:28< eleazzaar> but it's not going back to the rocks-sticking out of water look
20100929 01:30:38< esr_> but...what do *you* think that color is?
20100929 01:31:15< eleazzaar> under the reef: less saturated blue
20100929 01:31:30< eleazzaar> the ford: brown and blue
20100929 01:31:48-!- Daravel [~h_i_s_s@host81-129-92-64.range81-129.btcentralplus.com] has joined #wesnoth-dev
20100929 01:33:12< esr_> The rocks-sticking-out-of-water look is what made the DW maps pretty and interesting.
20100929 01:33:40< eleazzaar> that's what the reef is
20100929 01:34:15< eleazzaar> it's dumb to have two different terrain types so visually/conceptually similar
20100929 01:34:39< eleazzaar> feel free to change the map
20100929 01:34:47< esr_> Right now, on my display, those "rocks" look like floating sewage.  I'm not exagerating.
20100929 01:35:18< esr_> It's really quite repulsive.
20100929 01:35:24< eleazzaar> there's nothing more useful to accomplish here
20100929 01:35:29< eleazzaar> bye
20100929 01:36:51< esr_> OK, if you won't discuss the matter I'll just have to get you overruled.  Is that what you want?
20100929 01:43:38< eleazzaar> There are hundred of thing about the terrain i would like to improve, and will improve given time.  But i'm under no obligation to spend my efforts according to your priorities, especially when you insist on being so obnoxious and demanding about it.  I don't work for you.
20100929 01:47:05< esr_> eleazzaar: I'm being "obnoxious" because what I see on my screen is a disaster.  You could respond to this bty, for example, persuading  me that my display is screwing up the color values.  Or you could tell me what you're planning to improve and ask me to having me to have pattience; I'd probably respond to that by having patience,  Displaying your artistic temperment at me is, however,...
20100929 01:47:07< esr_> ...*not* the correct response.
20100929 01:48:12-!- stikonas [debian-tor@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20100929 01:57:32< Aethaeryn> Be careful with colors.
20100929 01:58:04< Aethaeryn> I know my laptop monitor and my 23'' monitor have totally different color appearances. Subtly different but obvious if you move a window back and forth.
20100929 01:58:18< Aethaeryn> What looks good on one monitor might look blah on another.
20100929 01:58:27< esr_> Aethaeryn: I'm listening.  It's the floating-sewage effect that's really getting to me, though.
20100929 01:58:45< Aethaeryn> Hmm, going to have to update.
20100929 01:59:02< Aethaeryn> Might take a bit to compile.
20100929 01:59:09< esr_> The ford is ugly but tolerable.  The reefs are intolerable.
20100929 01:59:45< esr_> You're right, it could be my display.  But...
20100929 02:00:14< esr_> ...do we want the terrains to look that bad on *anything*?  They didn;t before.
20100929 02:06:29-!- phlaem- [~a@212.84.227.31] has quit [Quit: Leaving]
20100929 02:06:49-!- Mica [~Mica@d27-96-76-1.nap.wideopenwest.com] has quit [Quit: ChatZilla 0.9.86 [Firefox 3.6.10/20100914125854]]
20100929 02:07:01-!- Blueblaze [~Blueblaze@adsl-99-182-53-14.dsl.hstntx.sbcglobal.net] has joined #wesnoth-dev
20100929 02:09:25< Aethaeryn> Again, compiling, I have to make sure I have the image.
20100929 02:10:48< Shadowmaster> esr_: I think you got it the other way around.
20100929 02:11:02< Shadowmaster> the ford is that brownish soil thing with a water overlay
20100929 02:11:18< Shadowmaster> that's supposed to be used for river fords. Reefs aren't.
20100929 02:11:56< Shadowmaster> and AFAICT DW's maps could benefit from revisions to take advantage of the current terrain set.
20100929 02:18:11< esr_> Shadowmaster: If we revise the terrain can I get back the pretty textured blue-on-blue effect the reefs used to have?  That was an important part of DW's visual appeal.  Well, I thought so anyway.
20100929 02:18:40-!- Daravel [~h_i_s_s@host81-129-92-64.range81-129.btcentralplus.com] has quit []
20100929 02:20:30-!- ancestral [~ancestral@166.137.141.169] has joined #wesnoth-dev
20100929 02:22:26< Shadowmaster> esr_: yes, you can.
20100929 02:22:40 * esr_ is all ears
20100929 02:22:44< Shadowmaster> I mean, can't you just play around in the map editor?
20100929 02:23:10< Shadowmaster> I'm not the one who wrote a full howto for campaign designers including instructions for map design, geez. :)
20100929 02:23:36< esr_> If you think I'll find the right combination by poking at it, I'll try.
20100929 02:23:52< esr_> What, you didn't like my HOWTO? ;-)
20100929 02:23:56< Shadowmaster> just know that most stuff has been converted into overlays. The reef terrain included.
20100929 02:24:24< Shadowmaster> so you should be able to put the reef on top of pretty much any kind of water tile type (and probably non-water too)
20100929 02:24:55-!- alink [~alink@wesnoth/developer/alink] has joined #wesnoth-dev
20100929 02:25:04< CIA-35> alink * r46811 /branches/ogl/src/ (sdl_utils.cpp sdl_utils.hpp video.cpp): 
20100929 02:25:04< CIA-35> Use (slow and bad) OpenGL for rendering and translate all SDL into OpenGL calls.
20100929 02:25:04< CIA-35> This works flawlessly on my box, but is several times slower (esp. with complex terrain)
20100929 02:25:04< CIA-35> It's just meant to be a temporary transition before real OpenGL optimizations,
20100929 02:25:04< CIA-35> and allows to have unoptimized parts still working in an OpenGL context.
20100929 02:25:04< CIA-35> alink * r46812 /branches/ogl/src/minimap.cpp: Fix broken minimap
20100929 02:25:05< CIA-35> alink * r46813 /branches/ogl/data/hardwired/tips.cfg: Use tip of the day to explain the goal of this branch and its current flaws
20100929 02:25:26< alink> \o/
20100929 02:25:37< Shadowmaster> esr_: no, actually, I liked your howto, although I'd have benefited from it if it had appeared a couple of months before its actual publishing date.
20100929 02:25:52< esr_> Shadowmaster: Sorry :-)
20100929 02:25:59< Shadowmaster> although in regards to map design it didn't tell me any news.
20100929 02:27:35< alink> btw I still didn't updated scons and cmake for the GL stuff. Since I am noob there, it would be cool if someone knowing that could take a look
20100929 02:28:07< alink> (in the ogl branch, of course)
20100929 02:28:18< Shadowmaster> I can take a look
20100929 02:28:24< AI0867> do we have a *new* ogl branch?
20100929 02:28:33< alink> Shadowmaster: many thanks :-)
20100929 02:28:34< Shadowmaster> yes
20100929 02:28:35< AI0867> because we had an old one, dating to 1.2 or something
20100929 02:28:47< AI0867> I may have gotten rid of it
20100929 02:28:56< Shadowmaster> yes, that one was removed
20100929 02:28:56< alink> AI0867: oh I reused the same name ?
20100929 02:28:58< Shadowmaster> tand its proper name was wesnoth_gl IIRC
20100929 02:29:03< Shadowmaster> alink: no
20100929 02:29:08< AI0867> alink: no, I think it was wesnoth-opengl or something
20100929 02:29:14< alink> ok good
20100929 02:29:34< Shadowmaster> it was wesnoth-gl according to git
20100929 02:29:46< eleazzaar> the reef and the ford are not overlays -- simply multi-layered
20100929 02:30:00< Shadowmaster> ah.
20100929 02:30:06< alink> anyway it's not OpenGL enough to have a full name yet
20100929 02:30:47< esr_> eleazzaar: Do you in fact think I'm just having a display problem?
20100929 02:30:54< Shadowmaster> alink: i'll take a look, tho I make no promises if new build tests are required ;)
20100929 02:32:13< alink> Shadowmaster: well, a first step is to make it build, details like build tests are for later
20100929 02:32:56-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100929 02:33:09< alink> but it's true that linux box don't always have working openGL, so at some point it will be needed
20100929 02:36:31< Shadowmaster> they usually should.
20100929 02:37:02< Shadowmaster> even if it's done via Mesa's swrast driver, there's usually a working OpenGL implementation installed.
20100929 02:37:17< alink> good
20100929 02:39:43-!- ancestral [~ancestral@166.137.141.169] has quit [Quit: Smell ya later!]
20100929 02:39:57-!- ancestral [~ancestral@97-116-168-17.mpls.qwest.net] has joined #wesnoth-dev
20100929 02:40:56-!- Blueblaze [~Blueblaze@adsl-99-182-53-14.dsl.hstntx.sbcglobal.net] has quit [Ping timeout: 276 seconds]
20100929 02:41:52< CIA-35> alink * r46814 /branches/ogl/data/game_config.cfg: Add a big visual thing to be sure that user of this branch reads my tip warning
20100929 02:41:54-!- Blueblaze [~Blueblaze@adsl-76-202-19-44.dsl.hstntx.sbcglobal.net] has joined #wesnoth-dev
20100929 02:41:57< alink> <:o)
20100929 02:44:24< alink> btw I already started to code real OpenGL texture stuff. Not working yet, but even when done, I will first wait to get tests of the current pixels hack
20100929 02:44:27< Shadowmaster> I love how cheap it's to switch branches with git-svn.
20100929 02:44:44< alink> Shadowmaster: yes me too :-)
20100929 02:45:01< Shadowmaster> oh, you are using git-svn ?
20100929 02:45:08< alink> yes
20100929 02:46:02< eleazzaar> esr_: i doubt it, but you can always run monitor calibration software
20100929 02:46:09< alink> that's why I didn't committed my work on this sooner, I was happy with local git branch
20100929 02:47:35< esr_> eleazzaar: I will do that if I can find any that runs under Linux.
20100929 02:48:37< alink> but my git logs were full of "disable this to prevent crash, understand how support it with ogl, re-enable it"
20100929 02:49:18< Shadowmaster> okay, working on an actual scons test.
20100929 02:49:35< alink> and other tweaks in the black magic arithmetic for sdl clipping 
20100929 02:50:14< alink> Shadowmaster: cool, and it builds ?
20100929 02:50:25< eleazzaar> esr_:  if you wish.  don't do it on my account
20100929 02:50:28< loonycyborg> Shadowmaster: Better let me port my silvertree stuff so it'll work on windows too.
20100929 02:50:31< Shadowmaster> I'll know after I finish working on it ;)
20100929 02:50:35< Shadowmaster> loonycyborg: aw.
20100929 02:50:40-!- SpoOkyMagician [~chatzilla@74-138-202-211.dhcp.insightbb.com] has joined #wesnoth-dev
20100929 02:50:43< Shadowmaster> okay, loonycyborg will do it.
20100929 02:50:49< Shadowmaster> I'll build using custom CXXFLAGS
20100929 02:50:54-!- Upth [ogmar@adsl-75-26-209-245.dsl.scrm01.sbcglobal.net] has joined #wesnoth-dev
20100929 02:51:52< alink> thanks guys, I really didn't want to work on that stuff.
20100929 02:52:15< esr_> eleazzaar: I *am* doing it on your account.  If I'm having a color-distortion prpblem, I don't necessarily have to hound you to fix those graphics.  (On the other hand, it's a problem if the terrain looks thios bad on any display that's not outright broken.)
20100929 02:53:48-!- Upthorn [ogmar@adsl-75-26-176-187.dsl.scrm01.sbcglobal.net] has quit [Ping timeout: 240 seconds]
20100929 02:54:08< eleazzaar> ...either way attempts at hounding would be ineffectual-- if not counter productive.
20100929 02:55:36< loonycyborg> Weird. GL branch builds for me even without additional flags.
20100929 02:56:16< Shadowmaster> alink: do you use any GL methods directly, or you just use SDL's wrappers?
20100929 02:56:27< Shadowmaster> if the latter then maybe the -lGL is not necessary.
20100929 02:56:52< alink> I use true GL calls
20100929 02:57:07< Shadowmaster> hm.
20100929 02:57:28< Shadowmaster> then maybe the SDL so pulls GL by itself
20100929 02:57:41< alink> not sure what you call SDL's wrappers, maybe SDL functions like SDL_GL_SwapBuffers() ?
20100929 02:57:57< Shadowmaster> alink: yees
20100929 02:58:14< Shadowmaster> hm, SDL loads GL via dlopen() IIRC, not compile-time linking
20100929 02:59:31< alink> there is only few such SDL_GL_* calls mainly for initiating the framebuffer, all rendering is done by true GL calls
20100929 03:00:52< alink> in fact, SDL_GL_SetAttribute it's not really needed here, so there is just SDL_GL_SwapBuffers()
20100929 03:00:52< Shadowmaster> let's say that loony's machine is smarter than ours and leave it at that.
20100929 03:00:56< loonycyborg> Probably SDL just pulls it in as a dependency. That seems to happen with dynamically linked libs.
20100929 03:01:13< Shadowmaster> because ldd wesnoth-1.8/wesnoth here doesn't mention any libGL.so.*
20100929 03:01:40< Shadowmaster> that could be caused by the funky linker flags I pass though.
20100929 03:02:05-!- ancestral [~ancestral@97-116-168-17.mpls.qwest.net] has quit [Quit: Smell ya later!]
20100929 03:03:24< loonycyborg> Even if I don't pass -lGL, libGL.so.1 => //usr/lib/opengl/nvidia/lib/libGL.so.1 (0xb6ee8000) appears in ldd's output.
20100929 03:03:40< loonycyborg> Though somewhat lower in that list than with -lGL
20100929 03:05:34< Shadowmaster> http://wesnoth.pastebin.com/WnHKdNA6
20100929 03:05:41< Shadowmaster> my trunk flags are even worse.
20100929 03:06:16< Shadowmaster> alink: now running my ogl branch build
20100929 03:06:49< Shadowmaster> it's slow as hell.
20100929 03:06:52 * alink hopes it doesn't crash too soon
20100929 03:07:40< alink> Shadowmaster: yes, but still usable here
20100929 03:07:54< CIA-35> loonycyborg * r46815 /branches/ogl/ (SConstruct scons/gl.py): Added scons check for opengl.
20100929 03:07:57< Shadowmaster> it's unusable here.
20100929 03:08:00< alink> Shadowmaster: any fps number ?
20100929 03:08:23< Shadowmaster> and there are several parts of the GUI that render as black rectangles for whatever reason
20100929 03:08:45< Shadowmaster> alink: 2 to 6 fps
20100929 03:08:51< loonycyborg> alink: 'big visual thing' you added is scaled non-uniformly on my wide screen :P
20100929 03:09:06< alink> Shadowmaster: Ah I had this bug, fixed here,
20100929 03:09:17< alink> loonycyborg: ah yes, indeed
20100929 03:09:48< alink> Shadowmaster: blinking black rectangle ?
20100929 03:10:46< alink> Shadowmaster: you mean fps when scrolling or animating, right ?
20100929 03:11:03< Shadowmaster> no, not blinking, just static: http://shadowm.rewound.net/wesnoth4.png
20100929 03:11:13< loonycyborg> alink: It doesn't seem to be very laggy for me.
20100929 03:11:31< Shadowmaster> fps when looking at a simple animated map view (test scenario, idle)
20100929 03:12:19< Shadowmaster> the black rectangles go away after resizing the window
20100929 03:13:08< alink> Shadowmaster: looks like my doublebuffer bug, there is optional operations there, I should check about it
20100929 03:13:13< loonycyborg> I'm getting 15 fps with first scenario of Dead Water though.
20100929 03:13:22< Shadowmaster> fps went to 50 fps after attacking with a unit.
20100929 03:13:29< Shadowmaster> the attack animation itself was played at less than 10 fps
20100929 03:13:58< Shadowmaster> some animated objects in the test scenario are rendered with various funny glitches when they are partially off-screen
20100929 03:14:05-!- Upth is now known as Upthorn
20100929 03:14:08< Shadowmaster>  that includes the floating labels and the campfire
20100929 03:14:41< Shadowmaster> there are more funny glitches if I leave the haloed Dark Adept partially offscreen
20100929 03:14:43< alink> ah btw window repaint is not implemented yet, so any desktop thing can cause glitches
20100929 03:15:17< Shadowmaster> the attack animation itself was played at less than 10 fpsmean that I should try at fullscreen anyway?
20100929 03:15:21< Shadowmaster> er.
20100929 03:15:22< alink> Shadowmaster: ok that's one of the clipping bugs that I hunted
20100929 03:15:38< Shadowmaster> I disabled KDE's window compositing temporarily. So you mean I should try at fullscreen instead anyway?
20100929 03:15:54< alink> Shadowmaster: testing fullscreen could be useful, things works differently with it
20100929 03:15:56< Shadowmaster> the attack animation itself was played at less than 10 fpsmean that I should try at fullscreen anyway? 90% of CPU time btw)
20100929 03:16:00< Shadowmaster> argh.
20100929 03:16:01< Shadowmaster> stupid touchpad.
20100929 03:16:32< Shadowmaster> Wesnoth's CPU usage is over 90% btw.
20100929 03:16:42< loonycyborg> alink: With plain trunk I'm getting 50 fps in first scenario of Dead Water :P
20100929 03:17:24< alink> loonycyborg: yeah I expected that ratio
20100929 03:18:33< Shadowmaster> as for the hardware specs
20100929 03:19:20< loonycyborg> nvidia GForge 6800 GS
20100929 03:19:33< Shadowmaster> this is an ATI Radeon HD 3200 (RS780) on Debian Squeeze+Sid+Experimental using Mesa's DRI2 ATI R600 driver from git master, on Linux 2.6.35.6 running with the drm/radeon driver in KMS mode. libdrm from git master, DDX from git master.
20100929 03:20:47< loonycyborg> And my cpu is AMD Phenom X3 at about 2GHz
20100929 03:20:58< alink> ok thanks for the info
20100929 03:21:05< Shadowmaster> there are also various glitches related to Wesnoth redrawing the fullscreen while its window isn't visible/minimized, btw.
20100929 03:21:34< Shadowmaster> starting TRoW and minimizing wesnoth iin middle of the loadscreen display caused it to stay as the background of the first story screen for some reason.
20100929 03:22:08< alink> "<Shadowmaster> some animated objects in the test scenario are rendered with various funny glitches when they are partially off-screen" I see it too. I know the cause. I will fix that tricky math tomorrow 
20100929 03:22:17-!- ancestral [~ancestral@97-116-168-17.mpls.qwest.net] has joined #wesnoth-dev
20100929 03:22:18< loonycyborg> alink: btw have someone considered adding support for disributing idle animations etc between cpu cores?
20100929 03:23:10< alink> Shadowmaster: yeah minimizing wesnoth, is in the "window repaint is not implemented yet" category
20100929 03:23:32< alink> loonycyborg: seriously ?
20100929 03:24:16< Shadowmaster> okay
20100929 03:24:42< loonycyborg> alink: Yes. Why not? Everyone are doing this lately :P
20100929 03:24:43< Shadowmaster> also, the CPU is an AMD Athlon(tm) X2 Dual-Core QL-62
20100929 03:25:09< Shadowmaster> loonycyborg: you know well that multithreaded programming is a double-sided knife.
20100929 03:25:30< alink> Shadowmaster: I know it's  annoying when multi-tasking, but at least now I am sure that desktop doesn't trigger it every second or weird things like that
20100929 03:26:50< Shadowmaster> alink: well, if I went to KDE's settings, enabled the compositing WM again and set window thumbnail generation to take place "always (breaks minimization)", that'd be a different matter. :P
20100929 03:26:53< alink> loonycyborg: with opengl at some point we will probably redraw all hexes every frames, so dividing things like that will not be wanted 
20100929 03:27:36< loonycyborg> alink: Maybe, but there might be other tasks.
20100929 03:28:00< Shadowmaster> I'll go back to trunk for now, btw
20100929 03:28:12< alink> loonycyborg: ah yes separating AI and rendering could make sense
20100929 03:28:28< alink> Shadowmaster: thanks for the testing
20100929 03:28:54< loonycyborg> Yes. Even AI itself could end up parallelizable.
20100929 03:29:29< Shadowmaster> go, ccache, go!
20100929 03:31:35< alink> ok, so the ogl branch doesn't crash and is not garbage, that is all I hoped for :-)
20100929 03:32:11< CIA-35> eleazar * r46816 /trunk/data/core/terrain-graphics.cfg: a quick, simple animated heat shimmer for the lava.
20100929 03:32:36< alink> (the first version of this hack were a crashy hall of mirrors)
20100929 03:32:47< Shadowmaster> wut, Linux 2.6.35.7 in less than two days after .6?
20100929 03:34:18< alink> anyway, I go now, I will fix the reported stuff later, bye
20100929 03:34:32-!- alink [~alink@wesnoth/developer/alink] has quit [Remote host closed the connection]
20100929 03:35:28< Shadowmaster> oh, it's a Xen fix.
20100929 03:54:24-!- noy [~Noy@wesnoth/developer/noy] has quit [Quit: noy]
20100929 03:57:47-!- SpoOkyMagician [~chatzilla@74-138-202-211.dhcp.insightbb.com] has quit [Quit: meh]
20100929 04:13:50-!- Elvish_Pillager [~eli@dhip-164.foss.residences.colby.edu] has quit [Ping timeout: 264 seconds]
20100929 04:18:50-!- Aethaeryn [~Michael@wesnoth/umc-dev/developer/aethaeryn] has quit [Quit: ...]
20100929 04:36:27-!- loonycyborg [~sergey@wesnoth/developer/loonycyborg] has quit [Quit: Zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz]
20100929 04:37:12-!- loonybot [~loonybot@wesnoth/bot/loonybot] has quit [Remote host closed the connection]
20100929 04:50:44-!- Mica [~Mica@d27-96-76-1.nap.wideopenwest.com] has joined #wesnoth-dev
20100929 04:51:54-!- ancestral [~ancestral@97-116-168-17.mpls.qwest.net] has quit [Quit: And that's the end of THAT chapter]
20100929 04:52:12-!- Ivanovic_ [~ivanovic@dtmd-4db2b876.pool.mediaWays.net] has joined #wesnoth-dev
20100929 04:53:54-!- Ivanovic [~ivanovic@wesnoth/developer/ivanovic] has quit [Ping timeout: 272 seconds]
20100929 04:54:08-!- Ivanovic_ is now known as Ivanovic
20100929 05:02:58-!- Arc [~arc@pysoy/developer/ArcRiley] has joined #wesnoth-dev
20100929 05:26:57-!- Gambit [~Gambit@unaffiliated/gambit] has quit [Read error: Connection reset by peer]
20100929 05:31:50< Arc> has anyone considered compiling wesnoth for android, instead of porting to java?
20100929 05:32:11-!- Shadowmaster is now known as shadowmaster
20100929 05:33:29< Arc> android supports native apps after all, though it'd require cross compiling to a few chipsets
20100929 05:47:23-!- Blueblaze2 [~Blueblaze@adsl-99-148-244-132.dsl.hstntx.sbcglobal.net] has joined #wesnoth-dev
20100929 05:50:03-!- Blueblaze [~Blueblaze@adsl-76-202-19-44.dsl.hstntx.sbcglobal.net] has quit [Ping timeout: 240 seconds]
20100929 05:50:03-!- Blueblaze2 is now known as Blueblaze
20100929 06:10:18< CIA-35> shadowmaster * r46817 /trunk/data/tools/helpheader.html: Update macro reference's header to match the live Glamdrol navbar
20100929 06:17:33< CIA-35> shadowmaster * r46818 /branches/1.8/data/tools/helpheader.html: 
20100929 06:17:33< CIA-35> Update macro reference's header to match the live Glamdrol navbar
20100929 06:17:33< CIA-35> (Backported from trunk, r46817.)
20100929 06:18:18-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100929 06:18:27< shadowmaster> Soliton: it might be a good idea to regenerate macro-reference.html according to r46817.
20100929 06:19:04< shadowmaster> I don't have write permission to it for a manual edit, since I don't belong to the "src" group
20100929 06:21:42-!- FAAB1 [~huajie@219.142.251.14] has joined #wesnoth-dev
20100929 06:23:24< shadowmaster> I just updated the main glamdrol CSS so wiki., forums., gettext., units. aren't considered external domains for the purpose of wiki link formatting btw
20100929 06:25:35< shadowmaster> okay, just wiki. and forums., since I just got frustrated at the enormous amount of subdomains we have listed in UsefulLinks :(
20100929 06:39:54-!- Blueblaze2 [~Blueblaze@adsl-99-148-246-221.dsl.hstntx.sbcglobal.net] has joined #wesnoth-dev
20100929 06:42:02-!- Blueblaze [~Blueblaze@adsl-99-148-244-132.dsl.hstntx.sbcglobal.net] has quit [Ping timeout: 264 seconds]
20100929 06:42:02-!- Blueblaze2 is now known as Blueblaze
20100929 07:09:34-!- noy [~Noy@wesnoth/developer/noy] has quit [Quit: noy]
20100929 07:30:16-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has quit [Disconnected by services]
20100929 07:30:31-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100929 07:41:26-!- Mica [~Mica@d27-96-76-1.nap.wideopenwest.com] has quit [Quit: I've left so now you're sad. - Night all]
20100929 07:48:07-!- Blueblaze [~Blueblaze@adsl-99-148-246-221.dsl.hstntx.sbcglobal.net] has quit [Quit: Blueblaze]
20100929 08:11:56-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has quit [Ping timeout: 240 seconds]
20100929 08:41:22-!- joo [~joo@unaffiliated/joo] has quit [Quit: college]
20100929 08:54:30-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100929 09:17:52-!- fangism1 [~davidfang@gilbert.achronix.com] has joined #wesnoth-dev
20100929 09:18:53-!- fangism [~davidfang@gilbert.achronix.com] has quit [Ping timeout: 245 seconds]
20100929 09:32:32-!- VladimirSlavik [~vladimir@PC-E233-396.ubmi.feec.vutbr.cz] has joined #wesnoth-dev
20100929 09:33:19-!- VladimirSlavik [~vladimir@PC-E233-396.ubmi.feec.vutbr.cz] has quit [Changing host]
20100929 09:33:20-!- VladimirSlavik [~vladimir@wesnoth/translator/VladimirSlavik] has joined #wesnoth-dev
20100929 09:46:22-!- crimson_penguin [~ben@wesnoth/developer/crimsonpenguin] has quit [Quit: crimson_penguin]
20100929 09:56:15-!- Ivanovic [~ivanovic@dtmd-4db2b876.pool.mediaWays.net] has quit [Changing host]
20100929 09:56:16-!- Ivanovic [~ivanovic@wesnoth/developer/ivanovic] has joined #wesnoth-dev
20100929 09:56:25-!- zookeeper [~l@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20100929 09:59:33< Ivanovic> moin
20100929 10:03:24-!- stikonas [debian-tor@wesnoth/translator/stikonas] has joined #wesnoth-dev
20100929 10:04:25-!- MeccaGod [~majs@h14n4fls31o279.telia.com] has joined #wesnoth-dev
20100929 10:24:58-!- FAAB1 [~huajie@219.142.251.14] has quit [Quit: Leaving.]
20100929 10:30:43-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has quit []
20100929 10:50:17-!- mjs-de [~mjs-de@vpw.wh.Uni-Dortmund.DE] has joined #wesnoth-dev
20100929 11:01:27< zookeeper> eleazzaar, the lavas aren't working...forgot to commit some images?
20100929 11:12:36-!- mjs-de [~mjs-de@vpw.wh.Uni-Dortmund.DE] has quit [Remote host closed the connection]
20100929 11:40:15-!- GNUtoo|laptop [~gnutoo@host245-55-dynamic.180-80-r.retail.telecomitalia.it] has joined #wesnoth-dev
20100929 11:49:35-!- beetlenaut [~dan@72-173-46-202.cust.wildblue.net] has quit [Read error: Connection reset by peer]
20100929 11:51:56-!- fangism [~davidfang@gilbert.achronix.com] has joined #wesnoth-dev
20100929 11:53:48-!- fangism1 [~davidfang@gilbert.achronix.com] has quit [Ping timeout: 272 seconds]
20100929 11:59:03-!- beetlenaut [~dan@72-173-46-202.cust.wildblue.net] has joined #wesnoth-dev
20100929 12:15:04-!- MeccaGod [~majs@h14n4fls31o279.telia.com] has quit []
20100929 12:47:01-!- beetlenaut [~dan@72-173-46-202.cust.wildblue.net] has quit [Write error: Connection reset by peer]
20100929 12:56:26-!- beetlenaut [~dan@72-173-46-202.cust.wildblue.net] has joined #wesnoth-dev
20100929 13:24:44-!- loonybot [~loonybot@ppp95-165-187-197.pppoe.spdop.ru] has joined #wesnoth-dev
20100929 13:24:45-!- loonybot [~loonybot@ppp95-165-187-197.pppoe.spdop.ru] has quit [Changing host]
20100929 13:24:45-!- loonybot [~loonybot@wesnoth/bot/loonybot] has joined #wesnoth-dev
20100929 13:25:45-!- loonycyborg [~sergey@wesnoth/developer/loonycyborg] has joined #wesnoth-dev
20100929 13:28:02< Ivanovic> mordante, alink: comment about the ogl branch: the tip of the day is not shown completely (the warning)
20100929 13:28:05< Ivanovic> 20100929 13:27:33 warning gui/draw: Text: text is too high for the canvas and will be clipped.
20100929 13:29:32< Ivanovic> alink: a glitch: using the pretty software mouse cursor in the main menu (windowed mode) leads to a corruption when moving the cursor to the buttom so that it is particially hidden and then moving it left/right
20100929 13:32:55-!- mjs-de [~mjs-de@vpw.wh.uni-dortmund.de] has joined #wesnoth-dev
20100929 13:33:29< Ivanovic> and yeah, you are right, it is slow
20100929 13:35:12< Ivanovic> getting 4fps currently in isle of the damned in the top right corner
20100929 13:36:01< Ivanovic> in general rendering looks correct though
20100929 13:39:13< CIA-35> ivanovic * r46819 /branches/ogl/ (CMakeLists.txt src/CMakeLists.txt): add basic opengl support to the cmake recipe
20100929 13:52:19-!- VladimirSlavik [~vladimir@wesnoth/translator/VladimirSlavik] has quit [Remote host closed the connection]
20100929 13:58:28-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Ping timeout: 245 seconds]
20100929 13:59:07-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100929 14:01:26-!- Gambit [~Gambit@pa-67-234-122-243.dhcp.embarqhsd.net] has joined #wesnoth-dev
20100929 14:01:26-!- Gambit [~Gambit@pa-67-234-122-243.dhcp.embarqhsd.net] has quit [Changing host]
20100929 14:01:26-!- Gambit [~Gambit@unaffiliated/gambit] has joined #wesnoth-dev
20100929 14:26:54< CIA-35> eleazar * r46820 /trunk/data/core/images/terrain/unwalkable/ (lava-A01.png lava-A02.png lava-A03.png lava-A04.png): adding animated lava images needed for previous commit.
20100929 14:32:45< zookeeper> eleazzaar, looks much better now ;)
20100929 14:33:14< eleazzaar> heh
20100929 14:36:39-!- Elvish_Pillager [~eli@dhip-164.foss.residences.colby.edu] has joined #wesnoth-dev
20100929 14:41:10-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Excess Flood]
20100929 14:41:52-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100929 15:24:38-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Ping timeout: 264 seconds]
20100929 15:26:09-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100929 15:28:36-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Excess Flood]
20100929 15:29:20-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100929 15:36:31-!- joo [~joo@unaffiliated/joo] has joined #wesnoth-dev
20100929 15:45:45-!- joo [~joo@unaffiliated/joo] has quit []
20100929 15:51:38-!- Elvish_Pillager [~eli@dhip-164.foss.residences.colby.edu] has quit [Ping timeout: 264 seconds]
20100929 15:56:02-!- joo [~joo@unaffiliated/joo] has joined #wesnoth-dev
20100929 16:07:25-!- EdB [~edb@tss37-1-89-84-18-220.dsl.club-internet.fr] has joined #wesnoth-dev
20100929 16:22:57-!- FAAB1 [~huajie@219.142.236.25] has joined #wesnoth-dev
20100929 16:23:38-!- Johannes13 [~Johannes@unaffiliated/johannes13] has joined #wesnoth-dev
20100929 16:29:34-!- _jbx_ [~jbailey@12.190.80.225] has joined #wesnoth-dev
20100929 16:29:35< _jbx_> kme
20100929 16:29:38< _jbx_> oops
20100929 16:32:35-!- dtiger [~dtiger@dynamic-vpdn-93-125-17-41.telecom.by] has joined #wesnoth-dev
20100929 16:33:05-!- stikonas [debian-tor@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20100929 16:38:17< Ivanovic> alink: wow, just tried the opengl branch with the gallium backend for mesa and wow, it takes about 2min for the logo to blend in
20100929 16:46:00-!- alink [~alink@wesnoth/developer/alink] has joined #wesnoth-dev
20100929 16:47:01< alink> Ivanovic: thanks for the tip, probably when I added the <big> pango markup
20100929 16:48:50< Ivanovic> alink: in general it seems to be really damn slow with the open source driver
20100929 16:48:55< alink> I thought that I fixed that cursor bug, I don't see here, but I will simplify that code to check if really safe in all cases
20100929 16:48:58< Ivanovic> still have to check with fglrx
20100929 16:49:21< Ivanovic> it seems to only happen at the buttom, not at the right edge
20100929 16:50:21< alink> about performance, my goal is make is usable for dev/testing on a *normal* map. Not a full water map with terrain animation on ;)
20100929 16:51:00< alink> and being able to scroll, move few units and open dialogs is enough for the moment
20100929 16:51:46< alink> test-scenario is also a bad test case, since it has a lot of floating things which are more expensive with ogl
20100929 16:52:21-!- MeccaGod [~majs@h14n4fls31o279.telia.com] has joined #wesnoth-dev
20100929 16:53:09< alink> Ivanovic: so it works even with the open-source driver ? (which one btw ?)
20100929 16:53:24< Ivanovic> the r600 classic from mesa git
20100929 16:53:31< Ivanovic> no render errors but *really* slow
20100929 16:53:43< alink> I can't make the nv works with my new screen, not sure why
20100929 16:53:47< Ivanovic> using the first scenario of two brothers as testcase i get a low one digit fps
20100929 16:54:04< alink> Ivanovic: when doing nothing ?
20100929 16:54:08< Ivanovic> yeah
20100929 16:54:11< Ivanovic> in :benchmark mode
20100929 16:54:23< Ivanovic> in 1024x768
20100929 16:54:38< Ivanovic> got to try how it works with fglrx, will do so later (got to reboot for this)
20100929 16:54:43< alink> oh ":benchmark" ok
20100929 16:55:39< alink> be sure to mention that when speaking about fps. :benchmark is not normal use, it's just used to compare things
20100929 16:56:20< Ivanovic> with plan :fps i get between 36 and 50fps in 1024x768 on the start area with some units and villages, no water
20100929 16:56:42< alink> and scrolling is decent enough ?
20100929 16:56:46< Ivanovic> in the top left corner i got constant 9fps
20100929 16:57:45< Ivanovic> it is rather sluggish with 6 to 14fps
20100929 16:58:11< alink> anyway, that works, I will clean the few glitches, and then continue the real work: using texture, which should provide a nice boost
20100929 16:58:42< Ivanovic> looking forward to hearing about mordantes performance
20100929 16:59:50< alink> as I said, I just want that it works without crash with usable interface
20100929 17:00:26< alink> because optimizing terrain make no sense if you can't pass titlescreen ;)
20100929 17:00:38-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20100929 17:01:12< Ivanovic> yeah
20100929 17:03:40< Ivanovic> beside the colored cursor i see no glitches
20100929 17:03:51< CIA-35> alink * r46821 /branches/ogl/data/hardwired/tips.cfg: Remove <big> pango markup from my custom tip because it triggers a bug
20100929 17:05:11< Ivanovic> alink: you might want to add some dumping of the current opengl render string in the normal startup info
20100929 17:05:23< Ivanovic> since this is good to have with testers of the gl branch
20100929 17:06:07< Ivanovic> what glxinfo puts out as "OpenGL renderer string" and "OpenGL version string"
20100929 17:08:26< alink> Ivanovic: very good idea. I will do that
20100929 17:09:16< alink> and I also forgot to add log and assert before tests
20100929 17:09:34-!- EdB [~edb@tss37-1-89-84-18-220.dsl.club-internet.fr] has quit [Remote host closed the connection]
20100929 17:10:54< Ivanovic> i think i saw some strange update behaviour with the animated water, but that might be due to the slowness
20100929 17:10:56-!- fangism1 [~davidfang@gilbert.achronix.com] has joined #wesnoth-dev
20100929 17:11:06< Ivanovic> okay, rebooting to get fglrx as test
20100929 17:11:23-!- fangism [~davidfang@gilbert.achronix.com] has quit [Ping timeout: 245 seconds]
20100929 17:12:40-!- FAAB1 [~huajie@219.142.236.25] has quit [Ping timeout: 255 seconds]
20100929 17:12:44-!- stikonas [debian-tor@wesnoth/translator/stikonas] has joined #wesnoth-dev
20100929 17:13:45 * alink is currently focusing on the alt-tab issue since it's really annoying
20100929 17:15:07< Ivanovic> okay, got some *serious* issues with fglrx
20100929 17:17:03< Ivanovic> it does not redraw stuff correctly, flickers in some strange way, ...
20100929 17:17:17< alink> which areas ?
20100929 17:17:19< Ivanovic> when composite is active and i run windowed the main menu does vanish
20100929 17:17:24< Ivanovic> alink: everywhere!
20100929 17:17:37< alink> and works with fullscreen ?
20100929 17:17:46< Ivanovic> when moving around on the map the terrain hex are not updated correctly
20100929 17:17:54< Arc> why is anyone using fglrx these days anyway?
20100929 17:18:28< alink> smells like my double-buffer problem.
20100929 17:19:41< Ivanovic> on the main menu eg the mouse over suff flickrs
20100929 17:20:22< Ivanovic> Arc: because it is a whole lot faster
20100929 17:20:37< Ivanovic> and there is no final release supporting acceleration on the latest cards
20100929 17:20:43< Ivanovic> alink: http://imagebin.org/116281
20100929 17:20:48< Ivanovic> the "ghosting" ingame
20100929 17:20:56< Ivanovic> this happend when i scrolled from the top to the buttom
20100929 17:21:05< Ivanovic> those redraw probs affect terrain, units, ...
20100929 17:21:12< Arc> true on the latest cards, but i thought they stopped developing fglrx in favor of putting their energy into the free driver
20100929 17:21:23< Ivanovic> no, fglrx is still developed
20100929 17:21:25-!- Elvish_Pillager [~eli@dhip-164.foss.residences.colby.edu] has joined #wesnoth-dev
20100929 17:21:27< Ivanovic> they are doing two things:
20100929 17:21:31< Ivanovic> 1) develop fglrx
20100929 17:21:42< Ivanovic> 2) support the open source driver with some coders and providing specifications
20100929 17:21:58< zookeeper> alink, i guess that you haven't thought more about solutions to the terrain alignment issue yet?
20100929 17:22:02< alink> Ivanovic: hmm having alpha blending between frames make no sense :-/
20100929 17:22:18< Arc> and funding, iirc
20100929 17:22:44< Ivanovic> Arc: funding is paying two developers to directly work on the open source driver
20100929 17:22:46< Ivanovic> ;)
20100929 17:23:04< alink> zookeeper: ah no sorry, I was busy with that ogl stuff. I'll try to get back on it soon
20100929 17:23:20< Ivanovic> but the framerate with fglrx is really fine
20100929 17:23:42< Ivanovic> just really glitchy atm
20100929 17:24:05< alink> and it's weird glitches
20100929 17:25:08< Arc> ah, it was my understanding they were paying the Xorg dri team directly in addition to having their own employees working on it, but it gets fuzzy with redhat and novell doing the same
20100929 17:25:27< alink> but I will add options for the double buffer update. I also had different behavior between fullscreen and windowed, so it's not a surprise that others system do things weirdly
20100929 17:25:49< Ivanovic> behaviour for the map glitch is identical fullscreen and windowed
20100929 17:26:05< Ivanovic> this is the problem with composite active: http://imagebin.org/index.php?mode=image&id=116282
20100929 17:26:32< Ivanovic> with every redraw it needs eg because the mouse over button text is displayed does make the top left corner move a little down to the right
20100929 17:26:42< alink> Ivanovic: I mean it looks related to the same part of the code
20100929 17:27:23< alink> that trippy look is almost nice ;p
20100929 17:27:41< Ivanovic> but unusable since the buttons are not where they are rerendered...
20100929 17:36:41< CIA-35> alink * r46822 /branches/ogl/src/controller_base.cpp: 
20100929 17:36:41< CIA-35> Redraw the map view after window/desktop event like alt-tab
20100929 17:36:41< CIA-35> Also log warning, so any desktop abuse (killing perf) will be visible.
20100929 17:36:56< alink> ^not done yet for GUI1/2.
20100929 17:37:40< Ivanovic> rebuilding anywa
20100929 17:37:42< Ivanovic> y
20100929 17:38:26-!- rigved [~rigved@116.72.163.243] has joined #wesnoth-dev
20100929 17:40:30< zookeeper> urgh...
20100929 17:41:05< Ivanovic> hey, now alt+tab is a great trick to get the game window displayed correctly for a short time!
20100929 17:41:07< Ivanovic> ;)
20100929 17:41:13< alink> hehe
20100929 17:41:44 * zookeeper is trying to write a WML event which detects when a wounded enemy unit could run away out of the player's reach if the player doesn't ZoC it, and then displays where the player would need to move units to do so
20100929 17:42:24< zookeeper> current mood: :K
20100929 17:42:30< alink> zookeeper: for new tutorial ?
20100929 17:42:33< zookeeper> yeah
20100929 17:42:41< alink> seems ambitious :p
20100929 17:42:48< alink> but nice
20100929 17:42:59-!- crimson_penguin [~ben@wesnoth/developer/crimsonpenguin] has joined #wesnoth-dev
20100929 17:43:32< zookeeper> well i just don't want to try to hardcode that kind of stuff since then that just leads to more hardcoding and it gets messy...i _think_ that can be done entirely dynamically with a bit of wizardry
20100929 17:43:37< Gambit> Achievement Unlocked: WML Masochist
20100929 17:47:43< zookeeper> it's not difficult to find out the locations to which you'd have to move to ZoC something, but to figure out which exact unit would need to do which move is a bit tricky...
20100929 17:49:19< alink> I am always surprised by the power of WML (in zookeeper's hands)
20100929 17:51:23< zookeeper> i wouldn't even try half of the stuff i'm doing for the new tutorial if lua didn't have wesnoth.find_reach() :x that's what really allows doing this without hardcoding a lot of locations and unit movements and such
20100929 17:54:18-!- Elvish_Pillager [~eli@dhip-164.foss.residences.colby.edu] has quit [Ping timeout: 240 seconds]
20100929 17:55:24-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Excess Flood]
20100929 17:56:27-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100929 18:07:21-!- phlaem [~a@212.84.227.31] has joined #wesnoth-dev
20100929 18:14:10< alink> I think I will try fast rendering with texture first (instead of hunting these glitches one by one). If fast enough, we can redraw everything for each frame (as we are supposed to do) and it should fix most of these glitches
20100929 18:15:12< Ivanovic> yeah
20100929 18:16:46-!- Aethaeryn [~Michael@wesnoth/umc-dev/developer/aethaeryn] has joined #wesnoth-dev
20100929 18:20:56-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100929 18:22:38-!- rigved [~rigved@116.72.163.243] has quit [Quit: Bye]
20100929 18:24:05-!- euschn [~euschn@wesnoth/developer/euschn] has joined #wesnoth-dev
20100929 18:49:51-!- FAAB1 [~huajie@219.142.236.25] has joined #wesnoth-dev
20100929 18:50:23-!- Aethaeryn [~Michael@wesnoth/umc-dev/developer/aethaeryn] has quit [Quit: class]
20100929 18:55:12-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100929 18:57:36-!- joo [~joo@unaffiliated/joo] has quit []
20100929 18:57:59< shadowmaster> wesbot: seen fabi
20100929 18:57:59< wesbot> shadowmaster: Person, who 5d 6h ago used nick fabi, last spoke 4d 11h ago.  4d 5h ago as fendrin was here and on the channels #wesnoth, #wesnoth-de and #wesnoth-umc-dev with the message: Remote host closed the connection
20100929 18:59:29-!- joo [~joo@unaffiliated/joo] has joined #wesnoth-dev
20100929 19:02:48-!- Elvish_Pillager [~eli@dhip-164.foss.residences.colby.edu] has joined #wesnoth-dev
20100929 19:08:23< euschn> alink: around?
20100929 19:08:31< alink> yes
20100929 19:08:48< euschn> I have seen that you made an opengl branch 
20100929 19:09:09< alink> yes, but just a first hack for now
20100929 19:09:16< euschn> i would love to contribute to that, since i have been starting to teach myself opengl lately
20100929 19:09:37< euschn> ok, so is there any task list or todos yet?
20100929 19:09:51< euschn> or is it mostly trying things out/hacking stuff together?
20100929 19:11:57-!- Blueblaze [~Blueblaze@adsl-99-148-247-33.dsl.hstntx.sbcglobal.net] has joined #wesnoth-dev
20100929 19:12:13< alink> ok, but a bit early now. I am just experimenting to get a base working. Later each special effect will probably need dedicated ogl, so help will be welcome for that.
20100929 19:13:12< alink> help to fix glitch specific to some hardware/driver is also welcome since it's often hard to debug
20100929 19:13:36< euschn> alink: ok, great
20100929 19:15:55< euschn> alink: is the following correct: the ogl branch already draws everything with opengl? i.e., the functions you added (sdl_blit(), sdl_fill_rect,...) do all the drawing in wesnoth?
20100929 19:16:25< alink> yes it translate all sdl calls in (bad) equivalent ogl calls
20100929 19:16:52< euschn> i see
20100929 19:16:57< alink> I avoid texture for now, because I want fine control of the texture stuff
20100929 19:17:17-!- Blueblaze2 [~Blueblaze@99.148.247.46] has joined #wesnoth-dev
20100929 19:17:53< alink> euschn: my main goal is to have a manageable transition, because it's impossible to update all the code in one pass
20100929 19:18:05< euschn> yea
20100929 19:18:34-!- Blueblaze [~Blueblaze@adsl-99-148-247-33.dsl.hstntx.sbcglobal.net] has quit [Ping timeout: 265 seconds]
20100929 19:18:35-!- Blueblaze2 is now known as Blueblaze
20100929 19:19:29< alink> I have started some texture stuff using a similar translation trick (sdl_surface to ogl_texture) but not ready yet
20100929 19:21:02< euschn> alink: can you access all individual pixels with sdl_surface? then it should not be to hard to do a conversion to a texture right?
20100929 19:23:28< alink> yes the actual data loading is not hard, but we use a wrapper object (surface) around sdl_surface
20100929 19:23:55< alink> and I wish to do different thing for cached surface and temporary surface
20100929 19:24:32< alink> and need to keep in mind spritesheet and texture atlas, since I don't want to create one texture by terrain image
20100929 19:25:58< alink> all of that without touching too much all the rendering callers spread in all the code
20100929 19:29:20< euschn> alink: I see, theres indeed lots of work ahead. thanks for all the info
20100929 19:29:46< alink> euschn: you are welcome
20100929 19:32:13-!- stikonas [debian-tor@wesnoth/translator/stikonas] has quit [Ping timeout: 245 seconds]
20100929 19:35:45-!- noy [~Noy@wesnoth/developer/noy] has quit [Quit: noy]
20100929 19:44:28< alink> zookeeper: is it the wanted output ? http://www.pastethat.com/corners
20100929 19:44:53< alink> also testing pastethat.com, I hope it works 
20100929 19:46:18< alink> done just with trunk WML but using an image trick (and still not clear why needed)
20100929 19:48:58< Ivanovic> euschn: what system are you using?
20100929 19:50:32< Ivanovic> (that is graphics card and driver)
20100929 19:52:12-!- FAAB1 [~huajie@219.142.236.25] has quit [Quit: Leaving.]
20100929 19:54:17< euschn> Ivanovic: Im using a laptop with some nvidia quadro and a desktop with a geforce gtc 460, with recent drivers
20100929 19:54:25< euschn> gtx*
20100929 19:55:21< Ivanovic> ah, all nvidia, too (just like alink)
20100929 19:58:00< euschn> Ivanovic: for testing purposes, I can also run wesnoth on my uni pc, which has some onboard intel graphics
20100929 19:58:16< Ivanovic> ah, okay
20100929 19:58:34< Ivanovic> already did some short tests with the open and close source ati drivers today
20100929 19:58:42< euschn> how did it work?
20100929 19:58:56< Ivanovic> the open source driver (classic) was rather okay, though really slow
20100929 19:59:14< Ivanovic> the open source driver (gallium) was by far slower (logo blend in took some 2min)
20100929 19:59:26< Ivanovic> the closed source ati driver does not work nicely at all
20100929 19:59:50< Ivanovic> many render probs, stuff like this
20100929 19:59:58< Ivanovic> (rendering was really fine with r600c)
20100929 20:01:37< zookeeper> alink, yes, that's right
20100929 20:01:54< alink> ok just sent you a PM with the data (now)
20100929 20:02:00< zookeeper> alink, no, wait, i'll check something..
20100929 20:02:06< alink> perfect timing
20100929 20:02:13< zookeeper> :P
20100929 20:04:02< zookeeper> alink, ok, yes, that's right. the image was missing convex or concave corners, but it seems my testcase didn't have both in the first place.
20100929 20:04:44< alink> yeah I just adapted what you gave me
20100929 20:04:52< zookeeper> looking at what you sent...
20100929 20:05:31< Soliton> shadowmaster: updated and you can do so too now.
20100929 20:05:42< shadowmaster> okay, thanks.
20100929 20:05:55< euschn> just finished building the ogl branch on my laptop
20100929 20:05:56< euschn> runs a little slow, but playable
20100929 20:06:37< Ivanovic> really slow over here with open source drivers, fast but glitchy with closed source ones
20100929 20:06:47< alink> good :)
20100929 20:06:51< Ivanovic> so "a little slow but without glitches" sounds like a decent compromise
20100929 20:06:53< Ivanovic> ;)
20100929 20:07:10< Ivanovic> but basically as expected, since you are using the nvidia drivers as is alink
20100929 20:07:15< euschn> mhm
20100929 20:08:37< alink> bbl
20100929 20:09:10< zookeeper> alink, cool, it seems to work. i have no idea why though :P
20100929 20:11:42< zookeeper> i'm guessing it's because the "invisible" top left hex is implicit in the rule in all rotations, not just in some?
20100929 20:12:11< Ivanovic> okay, back on the open source drivers again
20100929 20:13:30< zookeeper> alink, i'm guessing that happens because the "1" hex gets picked as the center hex around which the map is rotated, which in turn makes every rotation be of identical shape and thus the center= coordinates always point to the same location? urgh. complicated, but makes sense.
20100929 20:14:32-!- Blueblaze [~Blueblaze@99.148.247.46] has left #wesnoth-dev []
20100929 20:15:01< zookeeper> i'm gonna try if i can convert the human castle tiles to that format and keep them working right...
20100929 20:15:08-!- stikonas [~and@wesnoth/translator/stikonas] has joined #wesnoth-dev
20100929 20:15:16< Ivanovic> alink: just tried with setting things to the mesa (classic) software renderer
20100929 20:15:24< Ivanovic> WOW, *this* is what i call render problems...
20100929 20:15:38-!- wesbot changed the topic of #wesnoth-dev to: 160 bugs, 297 feature requests, 14 patches | logs: http://irclogs.wesnoth.org | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20100929 20:16:06< shadowmaster> try with the galium3d software renderer so I can know how well it does
20100929 20:16:07< shadowmaster> :p
20100929 20:16:15< shadowmaster> *gallium3d
20100929 20:16:52-!- joo is now known as joo|sleep
20100929 20:17:05< Ivanovic> looks rather the same
20100929 20:17:40< Ivanovic> http://imagebin.org/116301
20100929 20:17:52< Ivanovic> and not just imagine that all the villages you see are constantly flashing
20100929 20:18:26< shadowmaster> looks nice for a "drunk mode"
20100929 20:18:36-!- stikonas [~and@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20100929 20:19:01< zookeeper> alink, also i don't see any horizontal flipping... the images look like they're exactly the right way considering how the rule is written.
20100929 20:19:10< shadowmaster> can we backport the glitches to trunk and include them as an easter egg?
20100929 20:19:37< shadowmaster> especially designed for those who complain too much about the titlescreen
20100929 20:19:46-!- stikonas [debian-tor@wesnoth/translator/stikonas] has joined #wesnoth-dev
20100929 20:23:53-!- stikonas [debian-tor@wesnoth/translator/stikonas] has quit [Ping timeout: 245 seconds]
20100929 20:24:00-!- mordante [~mordante@roadie.xs4all.nl] has joined #wesnoth-dev
20100929 20:24:13-!- mordante [~mordante@roadie.xs4all.nl] has quit [Changing host]
20100929 20:24:13-!- mordante [~mordante@wesnoth/developer/mordante] has joined #wesnoth-dev
20100929 20:24:14-!- euschn [~euschn@wesnoth/developer/euschn] has quit [Quit: Leaving.]
20100929 20:24:16< mordante> servus
20100929 20:25:15< Ivanovic> hi mordante
20100929 20:25:20< mordante> hi Ivanovic 
20100929 20:25:21< Ivanovic> mordante: a lot to read for you in the logs
20100929 20:25:50< mordante> already read most during my lunchbreak ;-)
20100929 20:25:51< Ivanovic> and software rendering might look interesting for you, it does over here in latest mesa git when a force software mode
20100929 20:26:12< Ivanovic> using this: LIBGL_ALWAYS_INDIRECT=1 wesnoth-gl
20100929 20:26:40-!- stikonas [debian-tor@wesnoth/translator/stikonas] has joined #wesnoth-dev
20100929 20:28:30< Arc> looking at the lava anims, how difficult would adding animated rain be?
20100929 20:29:42< zookeeper> not difficult WML-wise...i did that something like 5 years ago
20100929 20:30:05< zookeeper> performance has been the only major obstacle
20100929 20:31:20< zookeeper> well, if anyone would suggest adding rain to the game, then balance and gameplay effects in mainline would be a problem too. but technically no other problem besides performance.
20100929 20:34:01< mordante> isle of the damned with before ogl is at 1000 fps in optimized build
20100929 20:34:31< Ivanovic> 1000?!?
20100929 20:34:49< Ivanovic> how have you optimized your build?
20100929 20:35:36< mordante> yeah yesterday tested with -o0
20100929 20:35:41< mordante> -O0*
20100929 20:36:03< mordante> will test further with optimized builds, makes more sense to test it like that
20100929 20:37:31< Ivanovic> i get below 70fps in 1024x768 at the top right corner of isle of the damned with animations active
20100929 20:37:56< mordante> with normal or ogl build?
20100929 20:38:00< Ivanovic> normal
20100929 20:38:04< Ivanovic> in :benchmark mode
20100929 20:38:24< mordante> what does benchmark mode do?
20100929 20:39:00< Ivanovic> it turns off the fps limit that is set to 50 and tries to redraw the screen as often as possible (or something like this)
20100929 20:39:27< mordante> ok I just start with --max-fps 1000 does the same
20100929 20:40:39< stikonas> Ivanovic: In future it is better to use vsync mode for OpenGL instead of limiting at 50
20100929 20:41:32< Ivanovic> stikonas: the stuff i just said is not about OGL mode
20100929 20:41:38< Ivanovic> those are numbers we currently have
20100929 20:42:03< Ivanovic> mordante: :benchmark might do a little more, alink should know
20100929 20:42:18< stikonas> I know, but I just remembered vsync, so I mentioned it
20100929 20:44:04< mordante> Ivanovic, it looks it also calls an invalidate all
20100929 20:44:13< mordante> so not something I want to test with
20100929 20:45:19-!- phlaem [~a@212.84.227.31] has quit [Read error: Connection reset by peer]
20100929 20:45:28-!- euschn [~eugen@wesnoth/developer/euschn] has joined #wesnoth-dev
20100929 20:45:33-!- phlaem [~a@212.84.227.31] has joined #wesnoth-dev
20100929 20:50:19< Ivanovic> you will get by far less in the opengl branch
20100929 20:50:33< Ivanovic> if it displays correctly at all
20100929 20:54:24< mordante> the good news is with OGL Wesnoth uses much less CPU... the bad news that the Xorg puts it at 100%
20100929 20:54:33< mordante> and that for a whooping 0 fps
20100929 20:55:13< mordante> all at 1920x1200
20100929 20:55:21-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100929 20:55:30< Ivanovic> better try 1024x768 for a start
20100929 20:55:56< Ivanovic> on some normal map like the first scenario of two brothers
20100929 20:56:00< mordante> will do that next, but like to know at large sizes as well
20100929 21:03:11< mordante> about 2fps at 1024x768 TB scenario 1 droided
20100929 21:06:01< mordante> but I see no big visual glitches at the moment :-)
20100929 21:06:11< Ivanovic> i get somewhere close to 50fps in the buttom right corner
20100929 21:06:28< Ivanovic> and about 8fps in the top left corner while doing nothing at all
20100929 21:06:32< mordante> so like the old ogl-branch very slow but no visual glitches :-)
20100929 21:06:53< mordante> doing nothing I also get more fps it drops when a unit moves
20100929 21:09:31< Ivanovic> i get something like 4 to 10fps when droiding the game
20100929 21:09:36< alink> yes :benchmark is a stress test redrawing the map(not the screen) as fast as it can. Usefull to evaluate the cost of a feature (example: grid as a visible impact)
20100929 21:09:55< alink> should maybe rename it
20100929 21:10:08< mordante> so better to test with --max-fps 1000 --fps
20100929 21:10:25< mordante> I think benchmark sounds scary enough for me ;-)
20100929 21:10:26< alink> mordante: to get results about normal use, yes
20100929 21:10:43< alink> especially with wide screen
20100929 21:10:44< mordante> yup and that is what we want for OGL
20100929 21:12:39< alink> Ivanovic: the bug of your screenshot is another doublebuffer problem. This doublebuffer thing was an unexpected difficulty of my hack (I found about it late when trying fullscreen). I hope that I can found a portable fix
20100929 21:12:56< Ivanovic> okay
20100929 21:15:28< alink> zookeeper: the images are all flipped compared to yours (and filename)
20100929 21:20:01< alink> zookeeper:it rotates the 7 hexes hexagon around (0,0) and then shift it to get all x,y positive. Because of the shape, "1" stays at the center but not because it's called "1"
20100929 21:20:45< alink> but maybe it's already why you meant
20100929 21:21:42< CIA-35> mordante * r46823 /trunk/changelog: Update changelog.
20100929 21:22:18< CIA-35> mordante * r46824 /trunk/data/ (4 files in 3 dirs): Strip trailing whitespace.
20100929 21:23:01< CIA-35> mordante * r46825 /branches/ogl/src/CMakeLists.txt: Fix compilation of tools.
20100929 21:28:52-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20100929 21:29:27-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100929 21:34:34-!- dtiger [~dtiger@dynamic-vpdn-93-125-17-41.telecom.by] has quit [Remote host closed the connection]
20100929 21:39:29-!- loonybot_ [~loonybot@ppp95-165-187-197.pppoe.spdop.ru] has joined #wesnoth-dev
20100929 21:39:29-!- loonybot_ [~loonybot@ppp95-165-187-197.pppoe.spdop.ru] has quit [Changing host]
20100929 21:39:29-!- loonybot_ [~loonybot@wesnoth/bot/loonybot] has joined #wesnoth-dev
20100929 21:39:29-!- loonybot [~loonybot@wesnoth/bot/loonybot] has quit [Read error: Connection reset by peer]
20100929 21:39:34-!- loonycyborg [~sergey@wesnoth/developer/loonycyborg] has quit [Read error: Connection reset by peer]
20100929 21:39:37-!- loonybot_ is now known as loonybot
20100929 21:40:44-!- loonycyborg [~sergey@wesnoth/developer/loonycyborg] has joined #wesnoth-dev
20100929 21:43:21< zookeeper> alink, yes, that's what i meant. however, i still don't see the image flipping; AFAIK, the difference in our images is that my weren't centered on the map= (because i wasn't using center=), whereas yours are centered. if i paste a hexmap on my images then i see that the center of my images is a junction of three hexes, whereas in your images the center is the middle of a hex.
20100929 21:43:54< zookeeper> i mean, obviously the images have a different layout but i still think that in your case there's no horizontal flipping happening
20100929 21:45:55 * alink feels very stupid and blame Konqueror
20100929 21:47:23< mordante> alink, you mentioned a hack for the `translation' of the title screen image a while ago, do you have some code for it?
20100929 21:47:25< alink> zookeeper: I initially incorrectly draw them flipped, noticed my error, flip them in Gimp, but my file browser didn't update the thumbnails :-/
20100929 21:48:16< alink> I saw them flipped on my file browser but wesnoth used the real files
20100929 21:48:41< alink> mordante: a hack-fix ?
20100929 21:49:34< alink> I don't remember any fix, just the bug
20100929 21:49:51< mordante> I think you said something about an image locator
20100929 21:50:40< zookeeper> alink, ah, i see :P
20100929 21:51:26< alink> ah no, I said that the image cache seems to be cleared, and since the logo didn't changed, I assumed that it used a surface instead of an image::locator
20100929 21:51:58< alink> *stored a surface
20100929 21:52:08-!- Mica [~Mica@d27-96-76-1.nap.wideopenwest.com] has joined #wesnoth-dev
20100929 21:52:37< alink> zookeeper: but glad that it has a simple explanation :)
20100929 21:53:37< mordante> alink, I'll have a look at it
20100929 21:53:51< zookeeper> alink, yeah... anyway, i'm still working on converting the human castle to use your method, it's just taking a fair bit of time...
20100929 21:54:14< alink> mordante: I'll retest what I just said
20100929 21:56:07< GNUtoo|laptop> hi do you need more screenshots?
20100929 21:58:16-!- GNUtoo|laptop [~gnutoo@host245-55-dynamic.180-80-r.retail.telecomitalia.it] has quit [Quit: Program received signal SIGSEGV, Segmentation fault.]
20100929 21:58:27-!- GNUtoo|laptop [~gnutoo@host245-55-dynamic.180-80-r.retail.telecomitalia.it] has joined #wesnoth-dev
20100929 22:00:00-!- Crab_ [~Crab_@wesnoth/developer/crab] has joined #wesnoth-dev
20100929 22:00:14< alink> mordante: yes seems valid, image cache is really flushed
20100929 22:00:24-!- fangism1 is now known as fangism
20100929 22:00:26< Crab_> wesbot: seen boucman
20100929 22:00:26< wesbot> Crab_: The person with the nick boucman last spoke 22h 27m ago.  21h 57m ago they left with the message: Quit: Leaving.
20100929 22:02:26< mordante> alink, I think I see the problem, let's see whether fixing it is as easy as I think
20100929 22:02:55< mordante> I indeed use a locator and keep the found image in the cache
20100929 22:04:03< fangism> loonybot: { std::cout << argv[0] << std::endl; } // :)
20100929 22:04:11< loonybot> /t
20100929 22:04:16< fangism> nifty!
20100929 22:11:17< shadowm_laptop> mordante: isn't it possible to align the version label in the titlescreen to the bottom left corner again?
20100929 22:11:27< shadowm_laptop> or alternatively, center it on the bottom edge.
20100929 22:13:20< mordante> shadowm_laptop, I like it as is so not about to change it
20100929 22:13:30-!- Crab_ [~Crab_@wesnoth/developer/crab] has quit [Quit: Leaving.]
20100929 22:13:33< shadowm_laptop> yes, I know you like things as they are ;)
20100929 22:13:40< shadowm_laptop> otherwise I'd not be asking.
20100929 22:16:19-!- anonymissimus [~chatzilla@HSI-KBW-095-208-089-203.hsi5.kabel-badenwuerttemberg.de] has joined #wesnoth-dev
20100929 22:16:35< mordante> but to answer you question it is possible
20100929 22:17:06< shadowm_laptop> I'll keep it in mind if I ever find myself standing in middle of the GUI2 code again.
20100929 22:18:48< CIA-35> alink * r46826 /branches/ogl/src/video.cpp: Temporary stop enabling V-sync, in case it triggers doublebuffer bug
20100929 22:19:13< alink> stikonas: ^ 
20100929 22:20:58< stikonas> I did not say that V-sync has to be enabled now, after all you are still in the development process 
20100929 22:21:56< CIA-35> mordante * r46827 /trunk/src/gui/dialogs/title_screen.cpp: Update wiki comment.
20100929 22:22:03< CIA-35> mordante * r46828 /trunk/src/gui/ (dialogs/title_screen.cpp widgets/window.cpp): 
20100929 22:22:03< CIA-35> Use the common hack for changing languages.
20100929 22:22:03< CIA-35> This removes the specialized code, the hack still needs to be removed at
20100929 22:22:03< CIA-35> some point.
20100929 22:22:05< CIA-35> mordante * r46829 /trunk/src/gui/auxiliary/canvas.cpp: Remove the deprecated canvas stretch option.
20100929 22:22:11< CIA-35> mordante * r46830 /trunk/ (changelog src/gui/auxiliary/canvas.cpp): 
20100929 22:22:11< CIA-35> Refresh the image from the locator upon redraw.
20100929 22:22:11< CIA-35> This fixes the title screen logo and map not to change after changing
20100929 22:22:11< CIA-35> the language. (Fixes bug #16631.)
20100929 22:22:15< alink> stikonas: yeah, I also didn't want to enable it now, it was a mistake when copy-pasting from my other OpenGL code
20100929 22:22:52< alink> also not even sure that I added it in the proper place. I will check that and add log to detect it
20100929 22:23:44< CIA-35> mordante * r46831 /trunk/players_changelog: Try for once not to forget the player changlog.
20100929 22:30:01< zookeeper> alink, ok, i think i'm onto something here. i converted the human castle to use your image layout and wrote a new macro for it...however, the layering of the different pieces gets messed up: http://imagebin.org/116318
20100929 22:30:06-!- cz1stf [~cz1stf@78.24.12.198] has joined #wesnoth-dev
20100929 22:30:14< zookeeper> that _might_ be the reason why the images have been cut up the way they are currently, but i'm not sure yet
20100929 22:30:39-!- cz1stf [~cz1stf@78.24.12.198] has left #wesnoth-dev []
20100929 22:30:49< zookeeper> also i think i can workaround that, but it'll require doubling the amount of rules...
20100929 22:31:41< alink> zookeeper: the rotation code do weird changes to base which affect layer, maybe causing this
20100929 22:33:11< alink> this is also the only impact of base.x IIRC
20100929 22:36:04< alink> zookeeper: if can check what happens to your castle when disabling this base rotation
20100929 22:38:13< alink> switching between branch with git is even more helpful now that I work between trunk and a svn branch :)
20100929 22:39:04< zookeeper> alink, actually i figured that it works like this: my macro places the top right corner first, and the rotations proceed clockwise (as always) after that, leading to the left and top left corners getting drawn above the bottom left corner, since the bottom left was placed before them.
20100929 22:39:26< zookeeper> well, still just a guess, but matches with the result
20100929 22:39:32-!- Johannes13 [~Johannes@unaffiliated/johannes13] has quit [Ping timeout: 276 seconds]
20100929 22:40:54< Ivanovic> alink: just made a tiny test: power saving with r600 drivers makes a *huge* difference for fps in wesnoth-gl
20100929 22:41:04< Ivanovic> that is: two brothers, top left corner
20100929 22:41:12< Ivanovic> with power level set to default i get 7 to 9fps
20100929 22:41:17-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20100929 22:41:24< Ivanovic> with level set to "mid" (the actual low power mode) i get 4fps
20100929 22:41:48< Ivanovic> so the limit is clearly on the *graphics* card, not some cpu bound thing
20100929 22:42:00< alink> zookeeper: http://www.pastethat.com/castles  <-- trunk without base rotation. Remind you something ? ;)
20100929 22:42:12< boucman> hey all 
20100929 22:42:18< alink> hi boucman
20100929 22:42:49< Ivanovic> hi boucman
20100929 22:43:24< Ivanovic> boucman: what graphics cards/drivers do you have at hand for looking for graphic issues in the ogl branch?
20100929 22:43:33< zookeeper> alink, somewhat, but doesn't look identical to what i'm getting
20100929 22:43:41< alink> Ivanovic: yes maybe I was not clear, there is some operation done by the GPU, but the blit comes from normal memory (I assume that it's the current bottleneck)
20100929 22:44:19< boucman> Ivanovic: lemme check
20100929 22:44:30< alink> zookeeper: yes small difference indeed, but similar incorrect layer order 
20100929 22:44:58< Ivanovic> boucman: over here with mesa r600 classic everything appears to be rendered correctly, the gallium version seems to just be a lot slower
20100929 22:45:01< boucman> radeon hd 3600 (not sure the X driver) and a poulsbo (no 3D drivers afaik)
20100929 22:45:12< Ivanovic> fglrx has serious graphics issues
20100929 22:45:33< Ivanovic> as has forcing indirect (which should be some software mode), there it is somehow worse than in fglrx
20100929 22:45:41< stikonas> r600g managed to render only about 2 spf
20100929 22:45:47< stikonas> on my laptop
20100929 22:45:50< Ivanovic> hmm, the poulsbo could be interesting
20100929 22:46:09< Ivanovic> stikonas: it takes about 2min on my box just to load the start menu animation, have not tried ingame yet
20100929 22:46:13< boucman> it's on a netbook so performance will probably be abismal
20100929 22:46:28< mordante> Ivanovic, do you know that is there a maximum size for the translated logo?
20100929 22:46:45< Ivanovic> mordante: uhm, no, i don't
20100929 22:46:54< Ivanovic> i never looked at the image translation stuff
20100929 22:47:04< alink> Ivanovic: maybe wait that I optimize it at least a little before asking everbody to test it ;)
20100929 22:47:05< Ivanovic> and i tend to completely ignore it, as i said to chusslove when he added this work
20100929 22:47:17< stikonas> mordante: I think they all have the same size
20100929 22:47:20< mordante> I like to set a maximum to make sure we don't need to scale at 800x480
20100929 22:47:29< Ivanovic> alink: looking for some render issues and getting a general idea what test hardware we have ain't wrong
20100929 22:47:31< Ivanovic> ;)
20100929 22:47:39< mordante> stikonas, they don't switch to Spanish and see the logo cut off
20100929 22:48:03< mordante> wesbot, seen chusslove
20100929 22:48:04< wesbot> mordante: Sorry, I don't know of chusslove.
20100929 22:50:19< boucman> zookeeper: did you notice that eleazzaar used your animated_water macro to animated heat over volcanoes ? you might want to rename that macro...
20100929 22:50:32< stikonas> mordante: you can find Chusslove in #kde-i18n
20100929 22:50:58< mordante> thanks stikonas 
20100929 22:51:14< alink> Ivanovic: ok, it's also true that more ogl-branch testers, more quickly we will be aware of hardware/driver problems
20100929 22:51:49< zookeeper> boucman, yes, i noticed, i'll probably rename it at some point, but it's not very important atm.
20100929 22:51:57< Ivanovic> mordante: http://pastebin.com/nNB2tj46
20100929 22:52:01< boucman> k
20100929 22:52:10< Ivanovic> looks like hungarian is currently the larges (by far!)
20100929 22:52:44< alink> mordante: ignore these crazy artists, rescale it at runtime ;-p
20100929 22:55:04< stikonas> Ivanovic: hungarian contains lots of blank space
20100929 22:55:18< zookeeper> alink, ok, after fixing all other glitches and problems, the castle works fine except for the layering glitch, which has now turned into the exact same thing as in your image
20100929 22:56:04< zookeeper> i'm not sure what would be the best way to deal with that..
20100929 22:56:40< alink> zookeeper: I would gladly kill this base rotation code if WML could be updated to avoid regression
20100929 22:57:36< zookeeper> alink, frankly i didn't quite understand what you said the code is doing to the base and layers...
20100929 22:58:02< alink> but I think base is used for foreground stuff, so need to be cautious 
20100929 22:58:45< zookeeper> _something_ needs to be done to the base and layers, but i guess it could be WML-specified instead
20100929 22:59:25< alink> zookeeper: I think it assumes that base is used as anchor point on map, and thus it update it when rotating it. But not even sure if it make sense to do that
20100929 22:59:59-!- Lancer2013_ [9fe68d16@gateway/web/freenode/ip.159.230.141.22] has joined #wesnoth-dev
20100929 23:00:45< zookeeper> alink, well, it might make sense. how exactly does it "update" the base when rotating?
20100929 23:01:12< zookeeper> it needs to add or subtract from it when rotating, but how can it know which one to do?
20100929 23:01:25< alink> it turns it around one point I think, not sure which one
20100929 23:02:33< Ivanovic> mordante: espreon has done this stuff lately
20100929 23:02:39< zookeeper> the center= perhaps?
20100929 23:02:57< alink> to be complete: each hex of the rule as a "base" shifted depending of its hex location. So, in fact, it makes sense to update it when moving them
20100929 23:04:13< alink> zookeeper: no idea yet, it's heavy math/code and comment says things like:" (demonstration left as an exercise for the reader)"
20100929 23:04:32< zookeeper> nice :P
20100929 23:05:28< alink> see builder.cpp:401 to get an idea
20100929 23:05:31< zookeeper> ok, could you fix the grammar in "each hex of the rule as a "base" shifted"? i'd like to be really sure what you're saying
20100929 23:06:02< boucman> alink: I think mordante an I did the design a long time ago... I remember what base is, i'm not sure how it acts and how it should act in rotations, though
20100929 23:06:38-!- Elvish_Pillager [~eli@dhip-164.foss.residences.colby.edu] has quit [Ping timeout: 240 seconds]
20100929 23:07:22< mordante> Ivanovic, *grumble* I know you like to increase the size of the tarball, but shipping all logos twice seems overkill
20100929 23:07:34< mordante> and of course guess which version I modified to test
20100929 23:07:40< CIA-35> ivanovic * r46832 /trunk/images/misc/l10n/hu/logo.png: 
20100929 23:07:40< CIA-35> remove extremely huge transparent area, this breaks displaying the icon!
20100929 23:07:40< CIA-35> @Espreon: please run the update stuff to fix the image tracking file
20100929 23:07:52< Ivanovic> huh?
20100929 23:08:17< mordante> data/core/images/misc/l10n/es
20100929 23:08:21< Ivanovic> honestly, i don't know how the logo stuff works at all, i just searched for the logo now and remove the extremely large trsnaprent areas
20100929 23:08:36< alink> zookeeper: sorry s/as/has :-/
20100929 23:08:37< mordante> me too only found that Spanish version 
20100929 23:08:43< alink> zookeeper: use :layers around a multihex image, each part will have different base (going north add 72)
20100929 23:08:53< alink> tested on great tree
20100929 23:09:15< zookeeper> i would if it worked in the editor ;)
20100929 23:09:59< alink> yeah really need to fix that, but the command handler is deep in game context
20100929 23:10:20< alink> and if not ":create" would allow to spawn unit in editor :-)
20100929 23:10:21< CIA-35> ivanovic * r46833 /trunk/data/core/images/misc/l10n/hu/logo.png: 
20100929 23:10:21< CIA-35> apply this logo to the other place where it is shipped
20100929 23:10:21< CIA-35> btw: why do we still have the /images/ dir outside of data?!? (including some duplicated stuff...)
20100929 23:10:43< zookeeper> hey, wow, i partially fixed it by changing to base=90,143
20100929 23:12:26-!- MeccaGod [~majs@h14n4fls31o279.telia.com] has quit []
20100929 23:12:47< mordante> boucman, not me and git blame blames gruikya
20100929 23:12:56< mordante> alink, ^
20100929 23:12:59< boucman> k, bad memory, then
20100929 23:13:08< mordante> I did the layer stuff
20100929 23:13:25< shadowm_laptop> Ivanovic: wmlscope needs the two separate dirs
20100929 23:13:38< shadowm_laptop> data/ is for stuff that is used by WML, the rest is for the C++, or so esr says
20100929 23:14:20< alink> boucman: ok, seems old, I would not be surprised if outdated. But still having terrrain WML relying/avoiding of this weird behavior
20100929 23:14:23< zookeeper> yeah, but really the whole separation between /, /data/ and /data/core/ is a bit messy.
20100929 23:14:31< shadowm_laptop> indeed it is.
20100929 23:14:36-!- SpoOkyMagician [~chatzilla@74-138-202-211.dhcp.insightbb.com] has joined #wesnoth-dev
20100929 23:14:47< zookeeper> like, wtf do about.cfg, help.cfg and co do in core?
20100929 23:14:50< Ivanovic> zookeeper, esr_: yes, the separation is a *real* mess
20100929 23:15:33< zookeeper> i just never touch it since it doesn't bother me personally and it'd be a lot of work to move a lot of stuff around.
20100929 23:15:45< Ivanovic> mordante: even better, svn up now and select hungarian (Magyar)
20100929 23:15:57< Ivanovic> since the logo is "too small" it is partitially repeated
20100929 23:16:15< Ivanovic> that is you see the start of the logo again on the right
20100929 23:16:38< mordante> Ivanovic, I know about that bug already fixed locally
20100929 23:16:44< Ivanovic> ah, okay
20100929 23:17:19< mordante> only needed a maximum so I no longer cut off widgets
20100929 23:17:28< mordante> s/widgets/the logo
20100929 23:17:37-!- Lancer2013_ [9fe68d16@gateway/web/freenode/ip.159.230.141.22] has quit [Ping timeout: 252 seconds]
20100929 23:17:37< Arc> zookeeper: i'd think rain would have to be explicitally added to a scenario, and would treat dawn/dusk as night and day as dawn/dusk, perhaps with some custom day cycle images for "its day but raining"
20100929 23:18:03-!- stikonas [debian-tor@wesnoth/translator/stikonas] has quit [Ping timeout: 245 seconds]
20100929 23:18:12< Arc> the day cycle stuff ive read how to do (thanks to UtBS) but how would you overlay animated rain over the whole board?
20100929 23:18:22< Arc> or perhaps, only part of the map
20100929 23:18:58< zookeeper> Arc, well, currently you'd use [terrain_graphics] for that. it can be added in a [scenario] as well, which is how my old code did it.
20100929 23:19:26-!- stikonas [debian-tor@wesnoth/translator/stikonas] has joined #wesnoth-dev
20100929 23:20:03-!- Blueblaze [~Blueblaze@adsl-99-148-247-46.dsl.hstntx.sbcglobal.net] has joined #wesnoth-dev
20100929 23:21:07< alink> zookeeper: math seems to rotate the base of each hex around (36,36) = center of hex (0,0).
20100929 23:21:46-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Excess Flood]
20100929 23:22:34< zookeeper> alink, yeah, but what meaning does that have when it's a multi-hex rule (like ours are)? for example in your testcase, the image has a base, not hexes.
20100929 23:23:21< zookeeper> or do you mean that the game sort of like cuts up the given base and calculates a separate base for each hex?
20100929 23:23:28< Arc> would terrain_graphics let you overlay the streaks of falling rain, not just their spatter on the ground?
20100929 23:23:49< alink> zookeeper: yes
20100929 23:23:51< zookeeper> Arc, yes, you can do pretty much _anything_ with terrain graphics WML
20100929 23:23:54< Arc> (sorry ive been looking on the wesnoth wiki to find the wml docs)
20100929 23:24:01 * alink has a deja-vu about this talk
20100929 23:24:05< Arc> do you have a page where terrain graphics wml is documented?
20100929 23:24:36< zookeeper> Arc, sure, but it's complicated: http://wiki.wesnoth.org/TerrainGraphicsWML i can try to dig up my old WML somewhere instead
20100929 23:24:58< alink> it's not complicated
20100929 23:25:03< alink> it's very complicated
20100929 23:25:06< alink> <:o)
20100929 23:26:16< Arc> oh speaking of this, is it possible to add text and graphics to the screen, not the map but an overlay either on the map or the sidebar which change based on variables?
20100929 23:26:48< Arc> ie, a counter for enemies slain, or total number of times you've crossed a river
20100929 23:27:17< Arc> i should rephrase - i know its possible - where should i be looking? is this a theme, or is there another way to do it?
20100929 23:27:41< zookeeper> Arc, ah well, i can't really find it anymore. anyway, if you really want to try to do it then i can help.
20100929 23:28:03< Arc> i'm reading up on terrain graphics, so yes i plan on doing it
20100929 23:28:21< Arc> would it be something for an addon or would it be useful enough to be included in the core game?
20100929 23:28:23< zookeeper> and no, that's not possible atm. it's something that everyone's wanted for a long time though.
20100929 23:28:42< CIA-35> mordante * r46834 /trunk/ (2 files in 2 dirs): 
20100929 23:28:42< CIA-35> Allow different sizes of the title screen logo.
20100929 23:28:42< CIA-35> Small logos are no longer tiled and large ones no longer cut off (at
20100929 23:28:42< CIA-35> least not unless they become larger as the maximum 550x250).
20100929 23:28:42< CIA-35> (Fixes bug #16632.)
20100929 23:29:01< mordante> Ivanovic, ^ that should fix the tiling
20100929 23:29:06< zookeeper> Arc, definitely add-on
20100929 23:29:25< Arc> cool
20100929 23:29:39-!- alink [~alink@wesnoth/developer/alink] has quit [Remote host closed the connection]
20100929 23:29:58< mordante> Ivanovic, guess most title screen issues have now been resolved
20100929 23:30:09< Ivanovic> mordante: cool
20100929 23:30:20< mordante> only the speed issue for some users is still there, but I prefer to wait and see what happens to OGL
20100929 23:30:27< Ivanovic> what about the logo blend in blocking basically everything since it uses too much resources?
20100929 23:30:29< Ivanovic> ah, okay
20100929 23:30:41< Ivanovic> in general i'd prefer if the speed of the logo blend in was about doubled
20100929 23:31:05< mordante> the main bottle neck seems to be the rendering of the whole screen since the logo can be behind the buttons on small resolutions
20100929 23:31:05< Ivanovic> it takes too long atm (especially considering that some users got to wait for it to finish blending in before they can actually use the game)
20100929 23:31:32< mordante> I need more info to reproduce that issue, it shouldn't be required 
20100929 23:31:44< mordante> unless of course the game takes too much CPU time
20100929 23:32:31< mordante> but if we move to OGL before 1.10 looking at fixing it for SDL is a waste of time
20100929 23:32:58< mordante> that's why I prefer to wait
20100929 23:33:02< Ivanovic> you should be able to reproduce it now with opengl mode, just start wesnoth and try to directly click something
20100929 23:33:10< Ivanovic> should still be dead slow and barely working
20100929 23:33:58< zookeeper> Arc, copying ANIMATED_WATER_15 (from data/core/terrain-graphics/internal-generic.cfg in trunk) would be a good starting point. then you just make it work on all terrains and maybe change the layering a bit and you should have it drawn on the whole map.
20100929 23:34:13< Arc> oh wow, you guys have been busy.  a lot of bugs have been fixed on the shavian translation
20100929 23:34:34< Arc> preferences screen is still messed, but title screen is beautiful
20100929 23:34:44< mordante> Ivanovic, true but if OGL gets faster fixing it for SDL seems mood
20100929 23:34:53< Arc> i'll actually break for a bit to do the new translations
20100929 23:35:07< mordante> Ivanovic, the problem is the high CPU usage due to a lot of redrawing
20100929 23:35:08< Ivanovic> mordante: yeah, that is right
20100929 23:35:30< mordante> and on larger screens there's no overlap and a lot of useless redrawing
20100929 23:35:48< mordante> and I assume OGL handles that better (even in software)
20100929 23:36:17< mordante> else I need to do tricks with dirty recs, which I hope OGL provides for free
20100929 23:37:01< mordante> alink also had some other ideas but I prefer to wait and see what OGL brings us (if before 1.10)
20100929 23:37:26< anonymissimus> Gambit: around ?
20100929 23:38:20< mordante> thanks Arc, the title screen has been converted to gui2
20100929 23:38:24-!- GNUtoo|laptop [~gnutoo@host245-55-dynamic.180-80-r.retail.telecomitalia.it] has quit [Quit: Program received signal SIGSEGV, Segmentation fault.]
20100929 23:38:25< mordante> the preferences no
20100929 23:38:35< mordante> not*
20100929 23:42:03< mordante> I'm off night
20100929 23:42:40-!- mjs-de [~mjs-de@vpw.wh.uni-dortmund.de] has quit [Remote host closed the connection]
20100929 23:42:56-!- mordante [~mordante@wesnoth/developer/mordante] has quit [Quit: Leaving]
20100929 23:43:43-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100929 23:44:55-!- Elvish_Pillager [~eli@dhip-164.foss.residences.colby.edu] has joined #wesnoth-dev
20100929 23:54:11-!- GNUtoo|laptop [~gnutoo@host245-55-dynamic.180-80-r.retail.telecomitalia.it] has joined #wesnoth-dev
20100929 23:55:10-!- _jbx_ [~jbailey@12.190.80.225] has quit [Quit: Ha ha, charade you are.]
20100929 23:55:40< zookeeper> could we get rid of the scattering of village/mountain/river/lake/etc labels on random maps already?
20100929 23:56:03-!- SpoOkyMagician is now known as SpoOkyMagician_
20100929 23:56:06-!- SpoOkyMagician_ is now known as SpoOkyMagician
20100929 23:56:13< zookeeper> in trunk they can't even be cleared anymore, not in-game and not in WML (not without iterating over every hex and setting an empty label anyway)
20100929 23:56:48-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has quit [Ping timeout: 240 seconds]
--- Log closed Thu Sep 30 00:00:07 2010