#jogamp @ irc.freenode.net - 20151029 05:05:19 (UTC)
20151029 05:05:19 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20151028050519.html
20151029 05:05:19 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20151029050519.html
20151029 07:01:50 * elect (~GBarbieri@anon) has joined #jogamp
20151029 07:24:12 <elect> hi
20151029 07:31:37 * jvanek (jvanek@anon) has joined #jogamp
20151029 07:50:59 * monsieur_max (~maxime@anon) has joined #jogamp
20151029 08:25:54 * bigpet (uid25664@anon) Quit (Quit: Connection closed for inactivity)
20151029 10:14:04 * doev (~doev@anon) has joined #jogamp
20151029 11:41:23 * Eclesia (~husky@anon) has joined #jogamp
20151029 11:41:28 <Eclesia> hi
20151029 12:48:36 <xranby> gouesseJ: i migrated my svn to git (one way conversion) using this guide http://john.albin.net/git/convert-subversion-to-git
20151029 12:49:13 <xranby> this guide had the nice feature that it allowed fixing up the commiters emails during the process
20151029 12:56:20 <xranby> http://www.theregister.co.uk/2015/10/29/robovm_open_source_not_working_for_us/
20151029 12:56:29 <xranby> Community respond by forking #RoboVM into #BugVM
20151029 12:56:36 <xranby> https://github.com/bugvm/bugvm/
20151029 13:51:52 <rmk0> do we have a documented list of caveats with shared contexts anywhere?
20151029 13:52:00 <rmk0> i have no doubt there are some!
20151029 13:52:18 <bruce-> I know of one
20151029 13:52:26 <bruce-> a caveat that is
20151029 13:52:28 <bruce-> not a list
20151029 13:53:18 <rmk0> yes?
20151029 13:54:17 <bruce-> I use a shared context for loading and unloading textures. On Intel HD 4000 (and probably newer models too) (only tested on windows) texture memory is not released if you don't call glClear once in a while
20151029 13:54:34 <rmk0> hehe, impressive
20151029 13:54:49 <bruce-> that took me a while to debug :)
20151029 13:54:52 <rmk0> glClear on which context?
20151029 13:54:56 <bruce-> on the shared context
20151029 13:55:03 <rmk0> the one that loads textured?
20151029 13:55:06 <bruce-> yes
20151029 13:55:28 <bruce-> on the context that loads and unloads textures, no actual drawing is performed on that context
20151029 13:56:02 <rmk0> wonder how they managed that one...
20151029 13:56:47 <bruce-> older Intel drivers did not have this issue, ever since mid 2013 it is like this
20151029 13:57:22 <rmk0> do they know about it?
20151029 13:57:48 <bruce-> I never reported it
20151029 13:59:21 <rmk0> is it trivial to reproduce?
20151029 13:59:35 <bruce-> with jogl yes
20151029 13:59:44 <rmk0> not in C?
20151029 14:01:17 <bruce-> I would say yes. I just don't work with C on a regular basis and I am afraid it would cost me a lot of time.
20151029 14:01:30 <rmk0> i'll have a go with it now
20151029 14:03:47 <bruce-> another fun thing is that on Intel HD graphics you can call glClear on a separate thread on the context as many times as you want. on nVidia Quadro K2000 cards on windows I have seen terrible performance and I had to change the interval
20151029 14:05:18 <rmk0> i assume i just need to keep allocating and freeing textures
20151029 14:10:21 <elect> I am trying Intellij, but I am having huge problems for simple things
20151029 14:11:14 <elect> like, new project from github, I know I have to fix dependencies, but there is no way to see missing dependencies error
20151029 14:11:38 <elect> you need to open all the packages up to a class to see it
20151029 14:11:59 <elect> moreover, I don't have the dependency viewer.. not included in the community edition
20151029 14:12:34 <elect> and each project requires its own window..
20151029 14:20:07 <zubzub> elect: that's because you are trying to use it like it's eclipse
20151029 14:20:19 <zubzub> you can use multiple projects inside the same window
20151029 14:20:25 <zubzub> but then they're not called projects
20151029 14:20:27 <zubzub> but modules
20151029 14:21:09 <zubzub> I don't use the build in github import thingywhatever in intellij so I can't help with that
20151029 14:21:20 <zubzub> they way I use it is by git clonig it manually first
20151029 14:21:31 <elect> I tried that
20151029 14:21:36 <zubzub> and then import project from existing source
20151029 14:21:45 <elect> ah, manually
20151029 14:21:46 <zubzub> select maven/gradle/whatever
20151029 14:21:48 <zubzub> yes
20151029 14:22:04 <elect> why doing that?
20151029 14:22:17 <zubzub> intellij will automatically detect that there is a git repo
20151029 14:22:19 <zubzub> and use that
20151029 14:22:26 <zubzub> no need to do anything githubby special
20151029 14:22:42 <elect> ok
20151029 14:23:48 <zubzub> at the bottom of your project overview you see a node called 'external libraries'
20151029 14:23:55 <zubzub> these are your project dependencies
20151029 14:24:04 <zubzub> if any of those are red, it means intellij can not find that dependency
20151029 14:28:43 <rmk0> bruce-: https://github.com/io7m/intel-texture-bug-20151029 <- doesn't reproduce the problem here, on sandybridge, Mesa 11.0.4
20151029 14:28:54 <rmk0> if you could try it at your leisure and see if there's something i've missed
20151029 14:28:55 <rmk0> no rush
20151029 14:29:10 <rmk0> it needs glfw 3 to build, but nothing else
20151029 14:30:06 <bruce-> maybe it is different on Windows?
20151029 14:30:14 <rmk0> it could well be
20151029 14:30:18 <xranby> http://www.europarl.europa.eu/news/en/news-room/content/20151022IPR98818/html/Mass-surveillance-EU-citizens'-rights-still-in-danger-says-Parliament
20151029 14:30:43 <xranby> this is good
20151029 14:33:37 <xranby> By 285 votes to 281, MEPs decided to call on EU member states to "drop any criminalcharges against Edward Snowden, grant him protection and consequently preventextradition or rendition by third parties, in recognition of his status as whistle-blower andinternational human rights defender".
20151029 14:33:59 <elect> only 4 votes..
20151029 14:34:09 <xranby> 4 votes was enough!
20151029 14:34:32 <elect> I find too much 281
20151029 14:34:42 <xranby> well if your MEP is one of those 281
20151029 14:34:48 <xranby> do call your MEP
20151029 14:35:01 <xranby> and tell them what you think
20151029 14:37:05 <elect> I cant find it anywhere
20151029 14:38:31 <xranby> you do not know which MEP you voted for?
20151029 14:38:47 <elect> yes
20151029 14:39:01 <elect> but I was looking for the votes list
20151029 14:47:43 * bigpet (uid25664@anon) has joined #jogamp
20151029 14:48:35 <xranby> elect: this document http://www.europarl.europa.eu/sides/getDoc.do?pubRef=-%2f%2fEP%2f%2fNONSGML%2bPV%2b20151029%2bRES-RCV%2bDOC%2bPDF%2bV0%2f%2fEN&language=EN search for 3. page 9
20151029 14:48:57 <xranby> there you have the names of the voters who woted -
20151029 14:49:01 <xranby> voted -
20151029 14:49:06 <rmk0> hrm... i'm looking at GLSharedContextSetter. is there a way to get the same functionality if all you have access to is a GLContext and not a (subclass of) GLAutoDrawable?
20151029 14:49:15 <elect> thanks
20151029 14:52:58 <elect> all +
20151029 14:54:33 <elect> I wonder why they dont link it on the side
20151029 14:54:38 <elect> of the news
20151029 14:56:13 <xranby> they link it from http://www.europarl.europa.eu/plenary/en/votes.html#banner_session_live
20151029 14:56:51 <xranby> click on results of votes
20151029 14:57:11 <xranby> clock on todays date and then click on Results of roll call votes available
20151029 15:30:26 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20151029 15:35:05 <rmk0> here's one for the list: on mesa, intel, if you try to use GL functions on a context that isn't current, you don't get a useful error. instead, you (may) get a message to stderr such as "function is a no-op" and will get malformed results from functions that return data
20151029 15:40:22 * rmk0 smacks straight into https://bugs.freedesktop.org/show_bug.cgi?id=41736
20151029 15:47:16 <rmk0> i... guess that's the end of attempting shared contexts
20151029 15:57:36 <elect> perfect, thanks
20151029 16:09:00 * elect (~GBarbieri@anon) Quit (Ping timeout: 265 seconds)
20151029 16:13:31 <rmk0> currently reading the sources, trying to determine... does GLMediaPlayer actually use shared contexts?
20151029 16:13:54 <rmk0> it's referenced from the mesa intel GLSharedContextBuggy quirk
20151029 16:14:12 <rmk0> "However, not all multithreaded use-cases are broken, e.g. our GLMediaPlayer does work."
20151029 16:14:33 <rmk0> it's not clear to me how it can possibly be working if it does use shared contexts, because it doesn't seem possible to even create them on the broken platform
20151029 16:26:36 <rmk0> i stand corrected
20151029 16:56:00 * Eclesia (~husky@anon) has left #jogamp
20151029 16:56:35 * doev (~doev@anon) Quit (Ping timeout: 264 seconds)
20151029 17:03:45 <rmk0> ok, so i can create shared contexts on intel, but with the following program:
20151029 17:04:03 <rmk0> https://github.com/io7m/jogl-scontext-20151029/blob/master/src/main/java/com/io7m/scontext/UseSharedContext1.java
20151029 17:04:23 <rmk0> i get:
20151029 17:04:25 <rmk0> Exception in thread "main" com.jogamp.opengl.GLException: This GL object is being incorrectly used with a different GLContext than that which created it
20151029 17:04:33 <rmk0> at com.io7m.scontext.UseSharedContext1.main(UseSharedContext1.java:113)
20151029 17:04:57 <rmk0> buffer objects are shareable across contexts, so it's not clear why i would be hitting this
20151029 17:05:43 <rmk0> ... oops
20151029 17:06:19 <rmk0> heh, used the wrong context
20151029 17:06:35 <rmk0> made the slave context current and then used the master context GL
20151029 17:06:48 * rmk0 heads back into the shadows
20151029 17:08:08 <rmk0> i guess i'm not clear exactly what the symptoms of 41736 are
20151029 17:08:51 <rmk0> there's a whole renderer quirk for it, so presumably they severe enough to warrant it. GLMediaPlayer claims to be able to work around it, but i don't understand exactly what it's working around or how
20151029 17:08:57 <rmk0> *they're
20151029 19:17:28 <bruce-> is there an easy way to actually flip the TextureData instead of messing around with uv coordinates?
20151029 22:25:49 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20151030 05:05:19 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20151030050519.html