irclog2html for #picogui on 20030804

00:39.35*** join/#picogui CIA (CIA@cia.utility.freenode)
00:43.44CIA11trunk/tools/irc/util r084529 03micah: Updating the socket path in testClient
04:16.14*** join/#picogui thebolt (~thebolt@3ffe:4008:24:0:0:0:0:1)
05:10.04*** join/#picogui gonkulator2 (~brandon@12-252-175-3.client.attbi.com)
06:34.44*** join/#picogui fragglet (~fraggle@pc2-glfd1-4-cust207.glfd.cable.ntl.com) [NETSPLIT VICTIM]
07:15.08*** join/#picogui fragglet (~fraggle@pc2-glfd1-4-cust207.glfd.cable.ntl.com) [NETSPLIT VICTIM]
07:15.09*** join/#picogui thebolt (~thebolt@3ffe:4008:24:0:0:0:0:1) [NETSPLIT VICTIM]
07:15.09*** join/#picogui CIA (CIA@cia.utility.freenode)
07:15.09*** join/#picogui gonkulator (~brandon@h-64-105-20-186.DNVTCO56.covad.net) [NETSPLIT VICTIM]
07:15.09*** join/#picogui file (~file@mctn1-0191.nb.aliant.net)
07:15.09*** join/#picogui prot-work (~jupiter@161.97.199.99) [NETSPLIT VICTIM]
07:15.09*** join/#picogui Xentac (~jchu@amga45cyy30x9.bc.hsia.telus.net) [NETSPLIT VICTIM]
07:15.09*** join/#picogui mourad (foobar@122.192-200-80.adsl.skynet.be) [NETSPLIT VICTIM]
07:15.09*** join/#picogui scanline (~micah@12-252-175-3.client.attbi.com) [NETSPLIT VICTIM]
07:15.09*** join/#picogui captain_proton (~blardebla@12-203-245-57.client.attbi.com) [NETSPLIT VICTIM]
07:15.09*** join/#picogui jojo (~jojo@io.infidyne.com) [NETSPLIT VICTIM]
07:15.10*** join/#picogui nemo (nemo@61.95.53.30) [NETSPLIT VICTIM]
08:31.32*** join/#picogui thebolt (~thebolt@protego.ipv6.svanfeldt.com)
09:37.34*** join/#picogui TD (~mike@m142-mp1.cvx5-c.man.dial.ntli.net)
10:16.08*** join/#picogui Dentoid (Miranda@h242n1fls35o814.telia.com)
10:16.48DentoidHi
10:16.54Dentoidscanline, ping
10:19.04scanlinehi
10:20.22DentoidHi :)
10:21.20DentoidI remade our crystal video driver a bit yesterday, so now it just uses Pico's internal video handling (vbl_linear32) and echoes it to CS's framebuffer.
10:21.39DentoidSo for now the output works fine. Not the fastest, but it works. :)
10:21.56scanlineprobably good for starters, but it will be nice to eventually do some mad acceleration like pg's opengl drivers had
10:22.04DentoidYeah, exactly
10:22.14DentoidBut this works for now, for testing.
10:22.18scanlineyep
10:22.25DentoidBut I've got a problem with input...
10:22.57DentoidWhen I click something and drag it, if I drag outside the rectangle of the original position of what I drag, it stops.
10:23.01thebolt|working(hi scanline)
10:23.06DentoidLike for example the titlebar of a window...
10:23.15scanlinehi thebolt|working
10:23.29DentoidI can only drag it to the edge of the original position of the titlebar.
10:23.42DentoidJust wondering if you've got one of those "Oh, that might be..."? :)
10:23.47scanlineDentoid: I know I've seen that before.. IIRC it's what happens when the mouse move event doesn't have the right button bits set
10:24.00scanlineI'd suggest enabling the input filter debug and sanity checking what it prints
10:24.44DentoidWell, I'll first just look at button bits. :)
10:29.37Dentoidscanline, what is infilt_pointing_send expecting? I mean for the button int?
10:30.11scanlineIIRC that's a button number for the up/down events and a mask for the move event
10:32.05scanlineoh, looks like they all use button masks
10:32.24DentoidSo it's 1<<button?
10:32.26scanlinewhen in doubt, copy what sdlinput does :)
10:32.27scanlineyep
10:32.49scanlinebut it should indicate all the buttons that are down, not just the one most recently pressed
10:32.53DentoidYeah, I was looking at SDLinput, but copying didn't work, so I was uncertain.
10:33.01DentoidOkay
10:33.28DentoidEven for mousedown events?
10:33.36scanlineyes
10:34.32DentoidOk
10:38.03DentoidCool, now it works.
10:38.16scanlineyay
10:38.33thebolt|working:)
10:38.40thebolt|workingDentoid: have any recent screenshot?
10:38.45thebolt|workingwith some more color maby ;)
10:38.56TDscanline: i think CIA is still eating xfree log messages
10:39.15scanlineTD: I never changed it :)
10:39.24TDcunning. that'd be why :)
10:39.31Dentoidhttp://dentoid.mine.nu/cspics/pico2.jpg
10:39.38DentoidIsn't that color enough for you? :)
10:39.49thebolt|workingDentoid: bah
10:39.52scanlinehaha
10:43.41Dentoidhttp://dentoid.mine.nu/cspics/pico5.jpg
10:44.04DentoidThe buttons are still wrong, but otherwise it works pretty ok.
10:44.31thebolt|workingthats better
10:44.38thebolt|workingfast?
10:44.45DentoidNo :)
10:44.53thebolt|workinghow slow? :)
10:45.00thebolt|workingand what is it that takes timè?
10:45.28DentoidI think it's drawing.
10:45.35DentoidIt's all software
10:46.14DentoidBtw scanline, that "select" statement. I'm not sure what it does, but... When it had 5 second timeout I got 5 seconds between each screenupdate too. Is that normal? :)
10:47.16scanlinehah
10:47.39scanlineselect() is a way to block until either a timeout is reached or there's activity on a file descriptor
10:47.51scanlineso it uses that to wait until a client or input driver wants attention
10:48.00scanlineI wouldn't expect it to work right in win32 :)
10:48.05Dentoid:)
10:48.09DentoidWell it works with SDL...
10:48.24thebolt|workinghm..
10:48.28scanlineah.. then it could be a problem with your input driver
10:48.31DentoidBut then again I think SDL relies on select to time its polls, so it decreases the timeout
10:48.44scanlineyes, the SDL driver has to poll
10:48.56DentoidThere's a call in the input drivers to alter the timeout right?
10:49.06scanlinesomething like that
10:49.15DentoidCS is event driven, so it doesn't need to poll.
10:49.55DentoidAh, hmm... I think I understand (I hope)... So basically what we're having here is that Pico is waiting for an event, and thereby stalling CS?
10:49.56scanlinewell picogui is event driven too.. the issue though is how you wait for an event in a portable way
10:50.16scanlineyes
10:50.31scanlineyou'll want to set that timeout to zero if CS is going to be continuously rendering
10:50.36DentoidOk... I reduced the timeout to 0
10:50.39DentoidYeah
10:50.42scanlineif CS has a way of only rendering when things change though, it could get more complicated
10:50.56DentoidCS does continuous rendering.
10:51.02scanlineif CS was blocking on some other event, pg wouldn't get a chance to read info from clients
10:51.02scanlineok
10:51.12DentoidBut then I'll do it in the input driver, since I guess that's cleaner.
10:51.23scanlineyes
10:51.38Dentoidthebolt|working, 60FPS
10:51.58scanlineit looks like your button and cursor problems are probably due to alpha blending
10:51.58thebolt|workinghm, Dentoid do you feed pg from CS eventqueue?
10:52.06scanlinenot sure why that would be an issue if you're using linear32 though
10:52.17DentoidNo idea... :/
10:55.43thebolt|workingDentoid: how do you send input to pg?
10:55.49DentoidEvents
10:56.13thebolt|workingDentoid: ok, and you send them directly? then you shouldn't need to poll
10:56.50scanlinethebolt|working: it's not really input events that are the issue, it's pg's event loop that is mostly responsible for network events
10:56.51Dentoidthebolt|working, yeah I know, but the problem was just that pico waits for events too. I've set it to just wait 0 time, but I did it directly in the pico source.
10:57.29theboltscanline: k.. hm..
10:57.35*** join/#picogui vIkSiT (~vIkSiT@164.100.35.114)
10:57.41vIkSiThey all!
10:57.50scanlinehello
10:57.59vIkSiTscanline: hey!
10:58.36vIkSiTa very basic question.. can pico replace the Xserver, for a very low end laptop which i want to switch over to Linux, AND maintain productivity.
10:59.10scanlineno
10:59.37scanlinethere are very few applications for picogui, and there's no way to do proper accelerated video in linux without X yet
10:59.42vIkSiTscanline: i see. so how can i use it?
10:59.50vIkSiTokay..
11:00.11vIkSiToh, acceleration isnt my priority..presently i run MuLinux with a basic, stripped down version of X..
11:00.28vIkSiTand was wondering if it can be replace by a better looking alternative..
11:00.48scanlinemost people don't realize how disgustingly slow video gets at normal desktop resolutions without any acceleration at all
11:03.56TDyou don't really have to replace X to make it look nice
11:05.40vIkSiTTD: well, i cant get a decent window manager to run on it..all require too many libs..
11:05.51vIkSiTfluxbox isnt exactly "Eye Candy"? :)
11:06.36TDwell, just use text mode ;)
11:06.57vIkSiTscanline: hmm. do you have icons on your desktop and things?
11:07.06scanlineno.. I hate desktop icons
11:07.17scanlineall I need is a root menu and a half dozen terminal windows
11:07.23DentoidI never see the desktop anyway, so I dunno what I'd use them for. :)
11:08.08vIkSiThehe.. yeah thats true!
11:10.20TDi have them. but likewise, i never really use them
11:10.42TDi do use nautilus sometimes though, so i might as well have them there
11:11.07vIkSiTTD: true..
11:11.37TDwell, RMS only uses emacs on the console. so it's possible :)
11:11.40TDas long as you like emacs
11:12.56vIkSiTTD: but then, RMS is RMS! i met him last summer, and well, he's 'different!'
11:12.58vIkSiT:)
11:15.10vIkSiTthebolt: but it IS powerful..:) not having used too much of it!
11:15.20DentoidWho the heck is RMS? :)
11:16.25theboltDentoid: Richard Stallman.. founder of GNU
11:17.49vIkSiTiirc, someone (jorrit maybe) had a nice tiff with RMS last year..?
11:17.50DentoidAh, okay. And just because he uses something that would be the best for everyone? :)
11:18.25vIkSiTDentoid: hmm.. i guess it would be avoided for the simple reason he Does use it! :D
11:18.27theboltvIkSiT: it was a few years ago
11:18.33DentoidvIkSiT :)
11:19.24vIkSiTthebolt: ohkay..
11:19.37vIkSiTi guess the interview is still online!
11:19.53theboltnot interview, it is a mail conversation
11:21.56vIkSiTthebolt: oh yes..
11:23.49Dentoidthebolt, correction... The FPS isn't that bad. I'd forgotten I put a csSleep(5) for each frame (to make the client start faster, since it is in another process now :)
11:24.31DentoidWhen nothing happens in Pico it's pretty much just as fast as not drawing anything as usual.
11:24.56thebolt:)
11:38.14*** join/#picogui file (~file@mctn1-8101.nb.aliant.net)
11:40.59*** join/#picogui TD_ (~mike@m213-mp1.cvx5-c.man.dial.ntli.net)
11:55.53*** join/#picogui file (~file@mctn1-8101.nb.aliant.net)
11:57.21*** join/#picogui Dentoid (Miranda@h242n1fls35o814.telia.com)
12:04.48*** part/#picogui Dentoid (Miranda@h242n1fls35o814.telia.com)
12:08.29*** join/#picogui Dentoid (Miranda@h242n1fls35o814.telia.com)
13:20.08*** join/#picogui gonk2 (~brandon@h-69-3-79-10.DNVTCO56.covad.net)
13:30.29*** join/#picogui file (~file@mctn1-8101.nb.aliant.net)
14:15.16*** join/#picogui Oktal (~mat@pc2-rdng3-5-cust28.winn.cable.ntl.com)
14:23.59*** join/#picogui file (~file@mctn1-8101.nb.aliant.net)
14:46.06CIA11trunk/pg1/server/input r084530 03sambau: Added KEY_FORWARDING constants definitions.
17:20.26*** join/#picogui ibot (ibot@208.186.182.172)
18:05.52gonk{work}captain_proton: ban him in here too
18:05.59captain_protoneh
18:06.08captain_protoni don't think there's as big a danger of him coming here
18:06.21gonk{work}alright
19:00.09*** join/#picogui Oktal (~mat@pc2-rdng3-5-cust28.winn.cable.ntl.com)
22:32.15*** join/#picogui lurgyman (~lurgyman@161.97.199.99)

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with blootbot logs, split per channel, etc.