**** BEGIN LOGGING AT Sat Jun 13 02:59:59 2015 Jun 13 05:06:22 gm Jun 13 15:00:29 apaar_ : pong Jun 13 15:19:32 karki_ :got in contact with ds2 he is travelling and available on email :) also i had some ideas for the control channel.Since i am passing the head and tail for each buffer in the structure itself there is no need to have them ,so control channel will only need to check- Jun 13 15:19:57 1)if buffer is full and possible data loss by overwriting Jun 13 15:20:28 2)some way to avoid the driver and pru trying to access the memory at the same time Jun 13 15:21:43 3)possibly the amount of data being tranfered if we want to move in a chunk at a time(not single char as i am doing right now) Jun 13 15:22:21 I'm more concerned about 2 Jun 13 15:22:43 discuss with Abhishek_ and alexanderhiam also Jun 13 15:22:51 alexanderhiam_ Jun 13 15:22:56 ^ Jun 13 15:23:16 apaar_ : are you having just one buffer? Jun 13 15:23:40 nope multiple right now i have implemented only one Jun 13 15:24:31 ds2 says that it would be better to keep the size variable as diferent applications can then optimise Jun 13 15:25:03 but number(i am still to decide) will probably be fixed Jun 13 15:30:19 each channel should have it's own buffer Jun 13 15:30:42 apaar_ : buffer size should be user defined Jun 13 15:30:45 as ds2 said Jun 13 15:30:54 don't fix it to an MB or something Jun 13 15:31:04 but yes, what about #2 Jun 13 15:32:46 #2 has one possible solution ie flags which are checked before every read and write on both pru and driver side but that may cause speed issues and maybe make it non realtime Jun 13 15:37:09 how can the PRU communicating with ARM have realtime - thats the main issue Jun 13 15:37:16 if we tackle that, rest is easy Jun 13 15:37:36 apaar_ : are you bale to stick to the week-by-week schedule I gave you? Jun 13 15:39:00 i am maybe 2-3 days behind as this week i am supposed to figure out and implement the algo but till last week i was on track Jun 13 15:48:28 karki : what if we mainain a offset between the read and write ie we know where the head and tail is in the circular buffer so we could tell the driver to always remain behind the tail(write) and head(read) to never be equal to tail and the same memory will never be accessed by driver or pru Jun 13 15:49:36 karki_:^ Jun 13 16:41:32 karki_ :Since the sizes of channel is varible can we implement one master sysfs file where one has to create/initialise the channels on driver side and pass the size to pru by adding one extra variable to the structure Jun 13 16:44:30 yello everyone Jun 13 16:48:05 hello Jun 13 16:50:58 what's everyone working on this fine Saturday day/evening/night/whathaveyou ? Jun 13 16:51:42 nerdboy: Didn't catch you after that meeting but I'll be hanging here for a while (insomnia as usual), so ping me if you're around Jun 13 16:55:18 good timing... Jun 13 16:55:31 * nerdboy has a proposal mtg Jun 13 16:56:21 hopefully i'll be back in a few hours... Jun 13 16:56:25 hah! Jun 13 16:56:33 on a saturday afternoon? Jun 13 16:57:00 well, my timing is impeccable, what can I say xD Jun 13 16:57:43 opensource never sleeps Jun 13 16:57:56 :) Jun 13 16:59:29 moto-timo: I like that, will use it from now on :) Jun 13 16:59:35 aaaaaanyway Jun 13 16:59:41 something very nice happened last night Jun 13 16:59:59 the good people at my dorm decided to go out everywhere in numbers Jun 13 17:00:30 which had a spectacular influence on my youtubes Jun 13 17:00:36 and make sure that everyone knew they were out in numbers, loudly ? Jun 13 17:00:42 and the video is *finally* up Jun 13 17:00:49 moto-timo: ofcourse! Jun 13 17:01:32 \o/ Jun 13 17:01:37 * moto-timo claps Jun 13 17:02:05 nerdboy: moto-timo: it lives here https://youtu.be/ojDqtxv9oGY Jun 13 17:02:14 bbl... Jun 13 17:02:31 pending approval that i didn;t mess something up spectacularly, I'll link it to elinux asap Jun 13 17:04:00 I doubt it needs "approval", just put it up on elinux until somebody complains :) Jun 13 17:04:12 will do Jun 13 17:04:32 moto-timo: feedback appreciated though, since I have a habit of overdoing audio/video stuff Jun 13 17:04:49 will do :) Jun 13 17:05:12 I'm usually the guy who will point to the right tools and do the original recordings. but I'm terrible at limiting myself on putting random stuff in the final cuts Jun 13 17:05:40 so i limited myself to only 1 thing for this one Jun 13 17:06:35 * moto-timo is the king of overcomplication Jun 13 17:06:39 :) Jun 13 17:09:20 hail to the King! Jun 13 17:09:42 nice video Jun 13 17:10:27 shubhangi: thanks! Jun 13 17:10:48 was a bit of a pain in the ass to do it on my laptop, but managed to do it with minimal lag Jun 13 17:11:07 http://nvisnjic.com/beaglesat has 404 error though Jun 13 17:11:22 coming up soon ? Jun 13 17:11:29 although chrome/impress.js presentation still lagged when I recorded it. Such is life when you're missing a proper desktop xD Jun 13 17:11:33 yeah, ofcourse Jun 13 17:11:45 didn't link it yet to a placeholder part of the blog Jun 13 17:11:54 to aggregate the updates and stuff Jun 13 17:12:09 will write it up in jekyll in a bit Jun 13 17:17:24 didn't regenerate my blog in a while, since I've been working on comments integration (and still am), but will update soon :) Jun 13 17:28:01 agree, nice video Jun 13 17:28:12 thanks Jun 13 17:28:33 still a bit torn if I should have excluded the easter egg Jun 13 17:28:35 it appeals to my ocd and adhd Jun 13 17:28:45 but I couldn't help myself :P Jun 13 17:29:01 always good to have some fun :) Jun 13 17:29:10 true dat Jun 13 17:29:28 I also learned that impress.js is a mighty fine alternative to PowerPoint Jun 13 17:29:51 or any other "classical" presentation program Jun 13 17:30:24 impressive Jun 13 17:30:28 * moto-timo slaps self Jun 13 17:30:32 glad to be rid of the MS creep for future presentations, although the browser compatibility could be better Jun 13 17:31:05 can't have everything I guess Jun 13 17:31:38 * neemo contemplates spending a lot of time building a gui application for impress.js to kill PP faster Jun 13 17:33:35 next summer Jun 13 17:41:19 yeah, would be nice Jun 13 17:41:25 we'll see Jun 13 19:17:21 karki_: ping Jun 13 21:53:40 inkscape -> jessyink -> svg slides Jun 13 21:53:55 can't get much geekier than that Jun 13 21:54:16 unless you'd rather code the whole thing in a .dot file Jun 13 22:13:40 nerdboy: I've started to like Inkscape Jun 13 22:14:05 it can do a lot and is quite easy to get a hang of it, once you invest the time and jessyink seems cool Jun 13 22:14:41 but I'm still super happy about finding Impress.js (and some other nice HTML5 pres tool) Jun 13 22:15:16 it was quite easy to remember the webprogramming stuff I did and heavily modify the demo presentaiton to turn it into the thing I want, but there's still room for improvement Jun 13 22:17:15 it doesn't scale well to all browser sizes (which is kinda my fault since I did a hack job of the html code and image sizing) and the css transitions are a bit laggy (probably because I hogged the resources to hell), but it works and is (almost) Web ready xD Jun 13 22:20:24 I'd say an advanced GUI editor (and internal presenter for standardization?) would help a lot to bringing this stuff to all those poor souls who didn't learn the joys of web programming yet Jun 13 22:20:45 if you can get those souls to try something better than PowerPoint that is Jun 13 22:21:07 s/better/"better"/ here **** ENDING LOGGING AT Sun Jun 14 02:59:58 2015