Perl 6 - the future is here, just unevenly distributed

IRC log for #openframeworks, 2015-01-15

| Channels | #openframeworks index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:03 w4ffles joined #openframeworks
00:07 vade joined #openframeworks
00:39 vade joined #openframeworks
00:56 mhashmi joined #openframeworks
01:37 ar_tama joined #openframeworks
01:38 ar_tama_ joined #openframeworks
01:45 billautomata joined #openframeworks
01:49 reqshark_ joined #openframeworks
02:27 w4ffles joined #openframeworks
02:32 burper joined #openframeworks
02:32 burper left #openframeworks
02:36 w4ffles joined #openframeworks
02:41 billautomata joined #openframeworks
02:46 mhashmi joined #openframeworks
03:03 payload joined #openframeworks
03:03 wjf joined #openframeworks
03:05 payload hi. can I compile of for linux amd64 with OpenGL ES 2.0 support? is there something special i need to do? i currently download it and try it
03:07 billautomata joined #openframeworks
03:10 payload I added PLATFORM_DEFINES = TARGET_OPENGLES2 in libs/openFrameworksCompiled/project/makefileCommon/config.linux.common.mk
03:12 payload i install and compile now and hope it works
03:17 admsyn any luck?
03:23 payload code is compiling
03:23 payload ;)
03:58 payload hm, examples run as OpenGL 3.0 still
03:59 admsyn I remember there being discussion about this for the NVIDIA Jetson I think, since it supports both
03:59 ar_tama joined #openframeworks
03:59 admsyn there's a pretty big forum thread about it, don't remember if it got resolved though
03:59 payload i compile it again now with TARGET_OPENGLES  ... not GLES2
04:05 payload i have to somehow get EGL in
04:06 lalou joined #openframeworks
04:07 payload i dont find the mentioned forum thread about this topic
04:07 admsyn http://forum.openframeworks.cc/t/trying-to-compile-an-example-on-tegra-jetson-tk1/16128
04:08 payload oh thank you :)
04:10 mhashmi joined #openframeworks
04:10 payload okay, I learned that I can look at ofConstants.h ... maybe I find clues. second option is looking for getting it compiled step by step: next would be getting egl included
04:12 admsyn just to check, you're on a graphics card that supports ES 2 right?
04:14 payload yes, my mesa supports es 2 and i run es2gears just fine for example
04:14 admsyn ok then
04:15 payload i changed ofConstants.h ifdefs accordingly. the headers were not included dispite of TARGET_OPENGLES. it dependet on IOS or ARM
04:16 admsyn I figured that sort of thing would have been discussed in the jetson forum thread?
04:16 payload yes it was
04:17 admsyn @jvcleave do you know any tricks for forcing ES 2?
04:17 payload but without details and he wanted normal opengl on an arm platform... but its easy to change
04:17 payload code is compiling, need to wait
04:17 payload errror
04:17 payload argh, glX still in the code
04:18 payload because its only fenced with #ifdef TARGET_LINUX
04:27 payload goes on, excluded ofAppGlut*
04:31 payload its always the same with this ifdefs. implicit assumptions everywhere
04:33 payload Done!
04:33 payload now checking examples
04:36 payload hm ./compileOF.sh compiles a different OF than examples/gl/*/Makefile needs. what is the difference between libs/openFrameworks and libs/openFrameworksCompiled?
04:37 admsyn libs/openFrameworks is the .cpp files, openFrameworksCompiled is usually just the IDE files etc
04:38 admsyn it'll be your makefiles for linux though
04:48 ar_tama joined #openframeworks
04:52 payload oh, glfw is shipped as a compiled dependency in lib but not with EGL support
04:52 payload i should had used git clone :/
04:53 payload or maybe not
04:54 payload also in git its a precompiled .a
05:06 payload oh well how do I link EGL into the libglfw3.a... build system is cmake
05:07 payload cmake is hard for me because i dont know it well enough
05:07 admsyn I don't know off the top of my head
05:08 admsyn there's some work being done right now on a build script system for all the individual OF libs (apothecary)
05:08 admsyn it's a little bleeding edge to use right now, but the scripts inside might help?
05:08 admsyn you're definitely heading off the beaten path here, though
05:10 admsyn that'd be in scripts/apothecary/formulas
05:10 payload meh I cant link a libEGL.so into an archive. should have known that
05:10 admsyn if you're using the git head, not the OF download
05:10 payload no, i dont use the git head cause it already includes precompiled stuff
05:11 payload i crafted my glfw3 together for linux64 with egl support
05:11 admsyn alright well https://github.com/openframeworks/openFrameworks/tree/master/scripts/apothecary/formulas
05:11 payload now i have to tell the OF makefiles to always link libEGL.so and such into itself
05:14 payload added PLATFORM_LIBRARIES = EGL Xinerama
05:14 payload oh yes, off the beaten path i am
05:14 payload but it looks promising
05:14 admsyn good!
05:16 payload i wonder why it tries to compile ofAppGLFW than there is ofAppEgl
05:16 payload didnt found the switch yet
05:16 payload GLFW makes problems
05:22 payload uhhh shaderExample compiles, links and runs
05:22 payload and is detecting the correct GLSL
05:24 admsyn cool
05:26 payload yes, definitly works
05:27 payload so basically hacked my way through. so I try to document the needed changes so someone can do it in a cleanly manner
05:31 payload I make a pullrequest on github? is this the way to go?
05:33 admsyn it might be better as an issue / feature suggestion plus a description of what you did
05:33 admsyn assuming what you did is more of a hack / experiment
05:33 admsyn if it's pretty clean though then yeah, go for it!
05:34 payload its more like an hack
05:34 admsyn IMO you should open an issue as a feature suggestion, explain what you did and why
05:34 admsyn though there's a lot  that's changed between 0.8.4 and the current git HEAD, so hard to say if what you did is out of date
05:34 payload okay, issues get read by someone?
05:35 admsyn yepp
05:35 payload fine :)
05:35 admsyn e.g. this is the activity this week https://github.com/openframeworks/openFrameworks/pulse
05:38 saravanans joined #openframeworks
05:53 payload https://github.com/openframeworks/openFrameworks/issues/3559
05:55 admsyn great, thanks!
06:04 payload joined #openframeworks
06:27 mhashmi_ joined #openframeworks
07:43 mhashmi joined #openframeworks
07:52 saravanans joined #openframeworks
07:58 payload joined #openframeworks
08:34 w4ffles joined #openframeworks
09:00 saravanans joined #openframeworks
11:55 mhashmi joined #openframeworks
11:55 DrCode joined #openframeworks
12:26 saravanans joined #openframeworks
13:07 mhashmi joined #openframeworks
14:07 wjf joined #openframeworks
14:20 w4ffles joined #openframeworks
14:24 saravanans joined #openframeworks
14:52 saravanans joined #openframeworks
15:13 saravanans joined #openframeworks
15:20 w4ffles joined #openframeworks
15:21 reqshark_ joined #openframeworks
15:48 admsyn joined #openframeworks
15:50 w4ffles joined #openframeworks
16:19 saravanans joined #openframeworks
16:40 reqshark_ joined #openframeworks
17:52 admsyn joined #openframeworks
17:53 admsyn joined #openframeworks
17:57 saravanans joined #openframeworks
18:02 saravana_ joined #openframeworks
18:06 saravanans joined #openframeworks
18:27 saravanans joined #openframeworks
18:54 pizthewiz joined #openframeworks
19:10 gabrielstuff joined #openframeworks
19:13 reqshark_ joined #openframeworks
19:40 saravanans joined #openframeworks
20:41 vade joined #openframeworks
20:43 pizthewiz joined #openframeworks
20:50 vade god I fucking hate this job
20:50 admsyn go on
20:50 vade id flip a table but co workers I like are sitting at it, ha.
20:55 pizthewiz vade: I meant to ask about your git rebase workflow, do y'all require clean histories for PRs or something?
20:56 vade thats it
20:56 pizthewiz day-um
20:56 vade and basically, that issue could be solved via a UX fix on github
20:56 pizthewiz They really hate the merge commit?
20:56 vade but instread, we get to force push and cherry pick
20:56 vade instead
20:56 vade I just lost some work, somehow
20:56 vade I have no idea how
20:56 pizthewiz I personally like the merge commit as it is pretty explicit about what came from where
20:56 vade its just fucking stupid
20:56 vade yes exactly
20:56 admsyn yeah same
20:56 vade and you dont overwrite and MAKE UP history
20:57 vade squashing commits and rebase - i lets you really fuck things up
20:57 admsyn except when you have a PR that's like 2 commits and 7 "merge master into x" merge commits but *shrug*
20:57 vade esp with an errant force push or something
20:57 pizthewiz Otherwise it looks like some changes got all time-machine'y and just splatted in rapid succession on HEAD
20:58 pizthewiz Yeah, I found a good rebase'ing tutorial from the eduX guys and the force push at the end is pretty… scary.
20:58 pizthewiz Is it at least like a pre-integration thing, you don't have to rebase multiple times right?
20:59 pizthewiz The more I think about it, the more weird it is… isn't this the whole point of DAGs?
21:02 jvcleave DAG?
21:03 pizthewiz Directed Acyclic Graph
21:04 vade well, I guess I could only rebase once pre integration
21:04 vade and merge during dev
21:04 vade I was trying to get used to it, and rebased a co workers branch that my branch depended on
21:04 vade lost a ton of work
21:05 vade it wouldnt be so bad if I enjoyed the work but this place is killing me
21:05 jvcleave sounds like svn
21:05 vade apparently stuff should be in the git reflog somewhere but
21:05 vade whatever.
21:05 admsyn maybe you should have some sort of version control in place :x
21:05 vade haha
21:05 vade we should git checkout our git repos
21:05 pizthewiz One of my favorite fix techniques back when I used monotone were "DaggyFixes" http://wiki.monotone.ca/DaggyFixes/ where you'd essentially branch from the bad rev and then merge back in instead of fixing it upstream, that way anything that descended from it would get a crack at pulling it in.
21:09 admsyn wow, OF doesn't compile right now and it took way too long for me to find the issue
21:09 admsyn derp
21:09 admsyn http://i.imgur.com/CvldZHp.png
21:09 admsyn I had, like, git-vision on
21:12 vade ha git-vision
21:12 vade I know that feeling
21:12 vade got C++ errors are the worst
21:13 jvcleave looks legit to me
21:13 jvcleave what is causing the error?
21:13 admsyn +looks legit. +nothing to see here. +move along
21:13 admsyn (someone copy-pasted a git diff or something, there's + chars at the start of those lines)
21:14 vade :D
21:14 vade yea the errors are just totally inane and unhelpful
21:14 admsyn but I was looking at it for like 5 mins like "I DON'T GET IT WHAT THE FOUABSF"
21:14 jvcleave ah - I thought that was from your editor
21:16 admsyn the fix diff kinda makes my brain hurt a little actually
21:16 admsyn https://github.com/openframeworks/openFrameworks/pull/3562/files
21:17 jvcleave -1 , needs spaces in the sqrt parens - please read the code formatting standards
21:17 jvcleave ;)
21:17 admsyn hey stop it
21:20 jvcleave you can merge right?
21:20 admsyn yeah I'm just gonna go for it
21:26 jvcleave so apparently it’s pretty much impossible to seek to a specific timestamp with h264
21:26 jvcleave http://www.mjbshaw.com/2012/04/seeking-in-ffmpeg-know-your-timestamp.html?showComment=1381872525667#c2065142213049942847
21:27 jvcleave this is throwing off all my “getCurrentFrame” - wondering if I should just throw it out
21:27 vade jvcleave: you can do it, but you have to decode the various frames around your frame, then search for that frames presentation timestamp, and then present that one
21:27 vade since h.264 frames in a GOP can be out of order
21:27 vade decode order and presentation order differ
21:28 jvcleave yeah -  it does somewhat work if you encode the movie with a GOP/keyframe as 1 (every frame)
21:28 vade no no i mean, you can do it, you just need to decode the entire set of frames
21:28 vade and then you get a a bunch of frames that you’d need to sort based on presentation time
21:29 vade then you select that one frame
21:29 vade that works for non iFrame h.264 video in AVFoundation
21:29 jvcleave so you basically have to read the whole file and build a dictionary?
21:29 sebllll_ joined #openframeworks
21:30 vade I think you can get the times for the keyframes
21:30 vade and then you know, im between this and that keyframe
21:30 vade and then you decode frames in whatever their decode order is
21:30 vade then you find the frame with the presentation time nearest your seek time,
21:31 vade I had to deal with this a bit with a custom transcoder in AVFoundation
21:31 vade we segmented and joined various videos and I had to split streams, and handle sample buffer stuff
21:33 jvcleave ffprobe kinda does this
21:33 jvcleave media_type=audio
21:33 jvcleave key_frame=1
21:33 jvcleave pkt_pts=1938432
21:33 jvcleave pkt_pts_time=0:00:40.384000
21:33 jvcleave pkt_dts=1938432
21:33 jvcleave pkt_dts_time=0:00:40.384000
21:33 jvcleave best_effort_timestamp=1938432
21:34 vade yea decode timestamp and presentation timestamp
21:34 vade yea fucking annoying
21:36 jvcleave I guess my thing is - if I can’t guarantee an accurate currentFrame #, should I give on at all?
21:38 jvcleave it’s more like decodedFrameCounter than currentFrame
21:40 vade hm
21:40 vade what Apple does is
21:40 vade you try to seek to a time
21:40 vade and it returns you the tme it actually seeked to
21:40 vade so its like
21:41 vade - (CMTime) tryToSeekToTime:(CMTime)someTime
21:41 vade and then you find a close frame thats easy, maybe a I frame
21:41 jvcleave right - and then you decode that and correct
21:41 vade yea
21:41 jvcleave i tried a little bit of that yesterday
21:42 jvcleave thanks tho - trying to weigh my battles
21:47 admsyn joined #openframeworks
22:03 vade yea that shit is so nuanced
22:03 vade :(
22:04 vade yea jvcleave basically AVFoundation gives an array of seekable time ranges
22:04 vade that might be one range, thats begin to end
22:04 vade or a few discontinuous ones
22:04 jvcleave so i just got a pts/dts after a seek
22:05 jvcleave but it looks like pts = 2.7e+07 dts=2.7e+07 (was seeking to 27s)
22:06 jvcleave well - that may be correct since this is an optimal move
22:06 jvcleave movie
22:06 jvcleave let me try with another
22:06 vade yea I think thats going to vary a lot with various encoders and profiles
22:07 vade because some profiles can optimize using out of order decode, and other weird shit I dont fully understand
22:07 vade like, the higher the profile the more errant wacko sometimes optimal for a few byte saving shit the encoder can do to make your life hell
22:08 jvcleave yeah - that is basically what the blog comment said - it really comes down to the container
22:18 jvcleave admsyn - you going to resonate?
22:18 admsyn yepp!
22:19 admsyn haven't locked in one of the long workshops yet but I want to
22:19 admsyn you?
22:20 jvcleave on the fence
22:20 jvcleave schedule conflict - MKX is released on 4.14
22:21 admsyn what's that?
22:22 jvcleave https://www.youtube.com/watch?v=1EROmqidZQc
22:22 admsyn ahhh
22:22 * admsyn is more of a street fighter type anyway <_< >_>
22:23 jvcleave that’s a while off - 2018 i heard
22:27 admsyn so I have time to rescue my arcade stick from storage then :)
22:27 admsyn resonate's skewing a bit more music festival-y this time, which *I'm* pretty ok with personally
22:27 admsyn last time was great
22:27 jvcleave interesting
22:28 admsyn well, 17 performances alongside 39 talks at the moment
22:45 sebllll joined #openframeworks
22:47 FergusL_ joined #openframeworks
22:49 ess_sing joined #openframeworks
22:52 qDot__ joined #openframeworks
22:53 barzi_ joined #openframeworks
22:53 jvcleave joined #openframeworks
22:53 pizthewiz joined #openframeworks
22:53 OliverUv joined #openframeworks
22:53 jacres joined #openframeworks
22:53 _rm joined #openframeworks
22:54 oneironaut joined #openframeworks
23:03 ilbot3 joined #openframeworks
23:03 Topic for #openframeworks is now Welcome :) Site: http://openframeworks.cc/ Logs: http://irclog.perlgeek.de/openframeworks/
23:05 barzi joined #openframeworks
23:05 connerb joined #openframeworks
23:05 tapout joined #openframeworks
23:13 _rm joined #openframeworks
23:13 7YUAAFKPK joined #openframeworks
23:13 OliverUv joined #openframeworks
23:13 esssing_ joined #openframeworks
23:13 jacres joined #openframeworks
23:13 The_Fly joined #openframeworks
23:13 davidnunez joined #openframeworks
23:30 DrCode joined #openframeworks

| Channels | #openframeworks index | Today | | Search | Google Search | Plain-Text | summary