**** BEGIN LOGGING AT Wed Jul 30 03:00:01 2014 Jul 30 05:28:35 praveendath92: ping Jul 30 05:28:41 praveendath92: what is your status? Jul 30 05:50:05 av500: I'm testing the compression now. Jul 30 05:50:16 on BBB? Jul 30 05:50:34 With the RLE I'm adopting I got 1.6 MB -> 14.6 KB. Jul 30 05:50:44 No. Just standalone. Jul 30 05:51:05 But the data I'm testing on is the raw frame data. Jul 30 05:51:48 please try to move to BBB Jul 30 05:51:56 after all, this is what the project is about Jul 30 05:52:36 Yes. I ran a couple of tests on with debian image. Jul 30 05:53:05 The damage handle function is being called continuously. Jul 30 05:53:19 It is called only once on PC. Jul 30 05:54:11 I'm working in a messed up timings. I wanted to ask you about the damage handle calls. Jul 30 05:55:29 av500: I have logs collected from PC and BB. Jul 30 05:55:30 https://github.com/praveendath92/udlfb/tree/master/logs Jul 30 06:06:35 what about handle damage? Jul 30 06:06:41 you should not care about it Jul 30 06:06:51 its the fb telling you what part of the frame to update Jul 30 06:06:53 so you do that Jul 30 06:07:52 Did you check the log for bone? Jul 30 06:08:15 In that link? The fb makes too many calls when using on bone. Jul 30 06:08:36 And the data with calls is just blank image data. Jul 30 06:08:51 yes, but we cannot change that Jul 30 06:09:19 but do you get an image on the android? Jul 30 06:09:22 On PC, I only get one damage_handle call. Future calls are all made to deferred_io Jul 30 06:09:22 from BBB? Jul 30 06:09:27 Yes. Jul 30 06:09:30 ah Jul 30 06:09:39 and you can make a blurry phone cam vid of it? Jul 30 06:09:44 I got the setup image. Full green. Jul 30 06:09:50 well Jul 30 06:09:58 something other than green Jul 30 06:10:13 And then, it is being immediately overridden by the blank calls data :( Jul 30 06:10:38 on BBB, can you dump the fb? Jul 30 06:10:44 just to verify what you should see? Jul 30 06:10:57 I'm not able to run xinit because the driver is just sending data from damage calls. Jul 30 06:11:13 ? Jul 30 06:11:24 I tried. It says invalid ioctrl. Jul 30 06:11:41 I dont understand Jul 30 06:12:19 I think the invalid ioctrl is because the driver is busy. Jul 30 06:12:30 well, then comment all the USB foo Jul 30 06:12:34 just make it all do nothing Jul 30 06:12:37 does xinit start then? Jul 30 06:13:58 You mean, comment data sending? Jul 30 06:14:44 yes Jul 30 06:14:48 do nothing Jul 30 06:14:53 Let me see. Jul 30 06:15:46 you could also try xinit with std devfb Jul 30 06:15:54 just to make sure that works in principle Jul 30 06:31:55 Morning Jul 30 06:41:44 morning Abhishek_ Jul 30 06:53:42 Abhishek_: beaglelogic uses both the pru’s right ?? one for communication and one for actual data gathering Jul 30 06:55:45 anujdeshpande: yes Jul 30 06:55:58 Abhishek_: why not use just one PRU ? Jul 30 06:56:46 anujdeshpande: 100Msps is only possible with both PRUs operating in tandem Jul 30 06:57:25 Abhishek_: makes sense. also, the code for comm PRU is in asm ? Jul 30 06:57:54 It's hybrid actually Jul 30 06:58:19 One function that is timing-tight is in ASM Jul 30 06:58:24 the rest is in C Jul 30 06:59:24 I tried to get it use inline ASM in C code but thought it wasn't portable enough Jul 30 07:03:01 Abhishek_: right. also, any documentation on PRU input that you refered to ? Jul 30 07:14:34 anujdeshpande: I used the datasheet and generated the pinmuxes with tking Jul 30 07:14:56 Abhishek_: datasheet ? as in the trm ? Jul 30 07:16:07 Yeah the pru reference guide, especially the EGPIO and the scratch pad sections Jul 30 07:17:52 I experimented with cycle counts to ascertain the cycle count of key assembly instructions Jul 30 07:19:54 Abhishek_: thanks ! Jul 30 07:20:16 anujdeshpande: had time to flash the bbb? Jul 30 07:24:36 vvu: unfortunately no :( will try this weekend i guess Jul 30 07:27:57 oki Jul 30 12:51:56 hi jkridner Jul 30 12:53:08 jkridner: http://beaglelogic.github.io/webapp/ Jul 30 12:57:14 alexanderhiam: accelerometer is done =) Jul 30 12:58:37 rseethamraju: how far are you through with your proposal? Jul 30 12:58:55 the function the user calls when an interrupt occurs is right now yaking back-front, potrait-landscape, tap and shake Jul 30 12:59:36 so that has to be a standard then? Jul 30 12:59:53 Abhishek_: I only have IoT left. Jul 30 13:00:20 I was supposed to do BLE but no h/w Jul 30 13:00:32 but we added extra stuff all along so Jul 30 13:00:37 are you looking at CC3000 now? Jul 30 13:01:14 no. why? Jul 30 13:02:03 no, it's an embedded SPI-based WiFi module, IoT related Jul 30 13:02:26 yes Jul 30 13:02:42 I was supposed to do a BLE modue Jul 30 13:02:46 module Jul 30 13:03:02 I see Jul 30 13:03:31 how far have you come? Jul 30 13:04:00 I am doing the last part - the web interface, you did see a part of that in action yesterday Jul 30 13:04:29 it would need a custom HTML5 Canvas based render component that will have to be written from scratch mostly Jul 30 13:04:57 cool Jul 30 13:05:18 I really like your project. It’ll be very useful. Jul 30 13:05:22 I hope I am able to finish it within the timeline, might overrun a week, though, let's see Jul 30 13:06:24 I’m writing GRE 10 days after the end. so I hope I’ll finish off on time. Jul 30 13:06:33 your project would also be very useful to develop embedded applications and do I/O rapidly. Jul 30 13:07:15 oh ya! I’ll also be doing an example using a bunch of sensors Jul 30 13:07:22 All the best for GRE :) Jul 30 13:07:24 so any example senarios? Jul 30 13:07:29 thanks =) Jul 30 13:08:07 alexanderhiam was saying take a picture if motion is detected. Jul 30 13:08:18 thats one, any other ideas? Jul 30 13:08:46 well, you could develop something like a line following bot Jul 30 13:09:01 that's a "classic" application Jul 30 13:09:15 why is that the goto thing Jul 30 13:09:26 its super boring! Jul 30 13:10:11 have you seen the no of fests that have that competition? its crazy! Jul 30 13:10:43 I'm done with them. Jul 30 13:11:09 me too =P Jul 30 13:11:34 but if you can think of anything else let me know? Jul 30 13:11:42 sure Jul 30 13:12:25 btw, you giving GRE for which subject? Jul 30 13:17:26 just the general one Jul 30 13:17:52 they cancelled the comp science one last yr Jul 30 13:18:13 I don’t want to give math or physics so… Jul 30 13:20:00 I see, you have any univs in mind? Jul 30 13:20:23 depends on my marks Jul 30 13:20:32 I have a very bad % Jul 30 13:20:36 like 65 Jul 30 13:21:04 so unless GRE score’s amazing I won’t get into anything good Jul 30 13:21:40 in that case I’ll work for a year or maybe do a long internship and try again Jul 30 13:22:15 are you going to write this year? Jul 30 13:24:24 I have two years to go Jul 30 13:24:40 (mine is a 5 year dual degree course) Jul 30 13:25:17 oh? so will you be doing another master? Jul 30 13:25:21 master? Jul 30 13:25:23 s Jul 30 13:25:52 nope, I'd likely be going for a PhD Jul 30 13:27:54 nice! Jul 30 13:35:10 praveendath92: ping Jul 30 13:35:46 vvu: pong Jul 30 13:36:36 bugs now with the BBB ? Jul 30 13:37:36 Yep. I took some time off that and worked on the compression. Jul 30 13:38:06 I think I found the issue. Jul 30 13:38:35 A couple of runs and I hopefully I will get to a fix. Jul 30 13:38:38 praveendath92: are you able to measure FPS? Jul 30 13:39:26 praveendath92: fix for the deffered io thingie ? Jul 30 13:39:31 Abhishek_: Nope. Not running any FPS tests yet. Need to shift everything onto BB. Besides I just completed my compression functions. Jul 30 13:39:47 Oh. It's not related to deferred io. Jul 30 13:40:02 what is the root of evil ? Jul 30 13:40:22 The resolution used by driver is being changes somehow. Jul 30 13:40:45 it changes ? Jul 30 13:40:51 panto jkridner alexanderhiam : I'm done with soft PWMs. PWM[0] to PWM[100] => 0% to 100% of 3.3V Jul 30 13:40:51 I will just ignore any changes to res. besides the 1st one. Jul 30 13:41:08 I'm expecting that that should fix it. Jul 30 13:41:11 yet to push code though Jul 30 13:41:26 Abhishek_: That's a neat webpage! Jul 30 13:42:05 * vvu statically compiling QT is a pain... Jul 30 13:42:26 vvu: I may have a pull request for BBBIfs sometime tomorrow :) Jul 30 13:42:33 the rendering is going to change as wavedrom is likely to be replaced with a custom Canvas implementation Jul 30 13:42:45 more about it in my status report Jul 30 13:43:00 Abhishek_: That means no library usage for waveforms? Jul 30 13:43:03 praveendath92: regarding what? Jul 30 13:43:06 what is wrong there? Jul 30 13:43:21 Nothing. Just some README formatting. Jul 30 13:43:25 praveendath92: There are performance issues with the lib Jul 30 13:43:48 ah ok Jul 30 13:43:53 Let me see if I can get it to your liking :) Jul 30 13:43:56 it does not scale well for large datasets and ends up freezing the browser windows Jul 30 13:44:06 Abhishek_: When I hit plot on the site now, is it actually doing some work? Jul 30 13:44:27 Because I experience a brief freeze. Jul 30 13:44:28 yeah, it's getting some randomly generated data from the JS Jul 30 13:44:49 exactly, I generate ~20K samples Jul 30 13:45:06 Oh. So, you are going write one from scratch? Jul 30 13:45:19 Or switch to a different lib? Jul 30 13:45:19 yeah, I'll have to Jul 30 13:45:54 jQuery flot is one of the candidate libs, but I couldn't find anything else Jul 30 13:46:16 BTW, when it's deployed, those many samples are updated per sec right? Jul 30 13:46:38 no, it's not really real-time Jul 30 13:47:01 the BBB will capture and send over the samples to the browser for render Jul 30 13:47:12 (or raw download) Jul 30 13:47:13 So, I won't be seeing a running waveform? Jul 30 13:47:30 umm, no. This isn't really like an oscilloscope Jul 30 13:48:01 A soft real time is good enough I guess :) Jul 30 13:48:12 though triggers will be there for starting the capture. Jul 30 13:48:40 Nice :) Jul 30 13:48:59 they are soft triggers implemented into the sigrok drivers Jul 30 13:59:55 If BeagleLogic has to run real-time, it needs at least GbE LAN or a USB 3.0 device Jul 30 15:54:06 * alexanderhiam has to run right now and will miss the meeting, will be back on in an hour or so Jul 30 15:59:14 gm all Jul 30 15:59:25 sorry I've been missing over the last several days Jul 30 16:00:01 hello Jul 30 16:00:09 hi Jul 30 16:00:25 Thanks karki_, Abhishek_, DiegoTc for status updates... Jul 30 16:00:35 did I miss others? Jul 30 16:00:48 jkridner : hello! Jul 30 16:02:42 jkridner: I sent mine just now. Jul 30 16:02:48 thanks praveendath92 Jul 30 16:03:58 any agenda items? Jul 30 16:04:54 We have only 1 full week of coding left!!!!! Jul 30 16:05:07 jkridner: we need to define the live code Jul 30 16:05:19 Aug 11 is the pencils down date. Jul 30 16:06:17 the rest is only need to push to the branch and it's ready! Jul 30 16:06:20 DiegoTc: you mean where your code lives? I'd want a clean pull request to github.com/beagleboard/bone101 of what we want to take live... I'll duplicate the same to http://beagleboard.org/support/bone101 Jul 30 16:06:50 alexanderhiam : hi Jul 30 16:06:59 jkridner: no I mean to click run and run the javascript code Jul 30 16:08:13 DiegoTc: ah. Jul 30 16:08:50 jkridner: I sent mine just now Jul 30 16:08:54 sorry Jul 30 16:10:29 DiegoTc: I've thought about using web workers because they have a 'terminate()': https://developer.mozilla.org/en-US/docs/Web/Guide/Performance/Using_web_workers Jul 30 16:10:35 thanks rseethamraju Jul 30 16:10:48 DiegoTc: let's discuss when we get to your update... Jul 30 16:11:10 Let's start with karki_, since his report was in first. Jul 30 16:11:20 yay Jul 30 16:11:40 so soft pwm is working now :) Jul 30 16:11:45 for one pin....... Jul 30 16:11:47 :( Jul 30 16:11:51 karki_: does panto have an idea? Jul 30 16:12:02 panto is nowhere to be seen :'( Jul 30 16:12:35 alexanderhiam : yoo hoo Jul 30 16:12:55 karki_: any chance you are driving a conflict on that pin? Jul 30 16:13:06 as in? Jul 30 16:13:06 karki_: you don't have a scope, do you? Jul 30 16:13:11 no :'( Jul 30 16:13:13 as in, there is something else driving the pin. Jul 30 16:13:22 jkridner no Jul 30 16:13:42 if I set it explicitly to __R30 = 0b10 Jul 30 16:13:43 what happens when you use digital out of 1 on that pin? Jul 30 16:13:49 then 3.3V Jul 30 16:13:51 k Jul 30 16:14:07 weird thing is that it keeps varying. Jul 30 16:14:19 once even in pwm it reached 3.3v Jul 30 16:14:39 what other pins have PWM support? Jul 30 16:14:43 now if I set pin 1, pin 2,3 also randomly float about Jul 30 16:14:49 pins 0-7 Jul 30 16:15:31 * jkridner only knows pins by P8/9_x Jul 30 16:15:40 (yes, I do know how to look them up) Jul 30 16:15:43 ah Jul 30 16:15:57 (but documentation shouldn't require external references) Jul 30 16:15:59 P8-45/46 Jul 30 16:16:13 P8-43/44 Jul 30 16:16:23 jkridner : source https://github.com/deepakkarki/pruspeak/blob/master/src/pru-firmware/pru1_firmware.c Jul 30 16:17:15 * jkridner finds the use of only PRU-mode pins to be excessively limiting. GPIO banks should also be usable at the lower speeds, IMHO. Jul 30 16:18:04 Thats cool for DIO (I'll do it sometime), but what about PWM ( the current problem at hand?) Jul 30 16:18:58 I'd think the same in terms of what pins to support, but as to your problem, a scope plot would be extremely helpful. Jul 30 16:19:17 P8-39 to P8-46 are the pwm pins Jul 30 16:19:33 alexanderhiam told he could help me with the scope Jul 30 16:19:41 we should get someone else to run it and generate a scope plot. I'd want an exact script and steps to run such that it doesn't take excessive time communicating. Jul 30 16:20:11 please publish code/script to run along with instructions on where to hook the scope and follow-up with us here on IRC. Jul 30 16:20:33 * jkridner is sad his 8GB uSD card with all his working code has disappeared. Jul 30 16:21:00 I'm quite certain that the timings are (near) perfect :) and the readme has been updated for "how to use" Jul 30 16:21:00 anyway, should be easy enough to reproduce. Jul 30 16:21:52 pins P8-45/46 to scope for PWM Jul 30 16:22:12 and SET PWM[0] to SET PWM[100] will work Jul 30 16:22:14 :) Jul 30 16:22:35 just try "SET PWM[0], 50" Jul 30 16:22:47 followed by SET PWM[1], 50 Jul 30 16:23:27 please script it. Jul 30 16:23:27 pin 45 should be @0.16V initially, and pin 46 at 0.0V Jul 30 16:23:54 with the interface, it should be easy to pipe script commands in. Jul 30 16:24:26 script what the inst? or steps to use pwm? Jul 30 16:24:55 all of it. Jul 30 16:25:33 should be as easy as 'git clone git://github.com/XXX/pruspeak && cd pruspeak && scripts/run_test_script.sh' Jul 30 16:25:48 ok Jul 30 16:26:01 I would not expect that to be too difficult. Jul 30 16:26:12 quite easy Jul 30 16:26:16 you can require whatever base image. Jul 30 16:26:51 umm but the "SET PWM" should be done in the interpreter got by running bs_shell.py Jul 30 16:27:04 karki_: with this issue behind you, would you know how to wrap up coding on your project over the next 10 days? Jul 30 16:27:31 karki_: shouldn't that interpreter be running as part of your setup? Jul 30 16:28:49 karki_: shouldn't it be as easy as 'nc localhost port < script.txt' to run your script if it is listening on a tcp port? Jul 30 16:29:39 karki_: http://en.wikipedia.org/wiki/Netcat Jul 30 16:30:13 Abhishek_: I see your web interface is reaching performance limits. Jul 30 16:30:34 it still feels to me like small samples should be good for initial learning of SPI/I2C/1-wire/etc. Jul 30 16:31:21 I think jumping to Canvas at this stage is too much to expect with only 10 days of coding left. Jul 30 16:31:56 jkridner: yes, it could if restricted to 1K Jul 30 16:31:56 if 'sigrok' is working well, and there is a web-based demo/intro for small samples, I think that meets the base expectations. what does your primary mentor say? Jul 30 16:32:16 mranostay: ? Jul 30 16:34:23 mranostay: any objections to wrapping up with this functionality or is there something more you need to see? Jul 30 16:35:14 Abhishek_: can we integrate this into bone101? Jul 30 16:35:37 Abhishek_: where is the data for http://beaglelogic.github.io/webapp/ coming from? Jul 30 16:35:44 jkridner: yes, that's where it is to be Jul 30 16:36:00 Abhishek_: will you send me a pull request? Jul 30 16:36:18 jkridner: that's programmatically generated Jul 30 16:36:59 Pull request, as in for bone101? Jul 30 16:37:03 Abhishek_: any zoom functionality? Jul 30 16:37:11 Nope Jul 30 16:37:12 Abhishek_: yes, a pull request for bone101 Jul 30 16:37:20 k. Jul 30 16:38:11 Abhishek_: so, data from the actual beaglelogic isn't yet done or just not in this mockup? Jul 30 16:38:40 It's actually both Jul 30 16:38:44 we are moving very slow today. Jul 30 16:38:56 jkridner : my connection got lost..... Jul 30 16:39:29 jkridner : 10 days depending on what all features you'd want Jul 30 16:40:34 karki_: well, that was the question, do you know what to implement to wrap-up? Jul 30 16:40:45 jkridner : oh! Jul 30 16:41:39 I could go for supporting more commands or clean the current stuff (error handling etc.) Jul 30 16:41:43 Abhishek_: what do you mean 'both'? you need to integrate into the mockup as well as create the data transport? Jul 30 16:41:52 make it non user breakable basically Jul 30 16:42:03 jkridner: actually BeagleLogic has its own server at port 4000. So the bundle would be like bone101 but not into it Jul 30 16:42:26 Abhishek_: I see the mockup as good enough. Jul 30 16:42:47 karki_: I'd focus on cleanliness and stability, rather than adding more features, in my opinion. Jul 30 16:43:05 okay jkridner : I'll also add a TCP server Jul 30 16:43:05 karki_: especially documentation and modularity. Jul 30 16:43:16 so you can use netcat Jul 30 16:43:27 * mranostay pokes in Jul 30 16:43:29 karki_: what is the python thing today? I thought it was listening on a TCP socket. Jul 30 16:43:29 it was there before I removed it though Jul 30 16:43:35 oh. Jul 30 16:43:37 trains ran late again Jul 30 16:44:05 mranostay: wanting to make sure Abhishek_ knows what needs to be done to wrap-up over the next 10 days. Jul 30 16:44:06 jkridner : The python thing is just a shell! Jul 30 16:44:29 jkridner : I didn't port the TCP thing after I built the new compiler suit Jul 30 16:44:32 karki_: yes, but I thought you had a listener on a socket to accept commands. Jul 30 16:44:44 jkridner: the mockup needs some more code before it can render actual data Jul 30 16:44:45 no standard in Jul 30 16:44:45 karki_: so, you are using python scripts to pass all the commands? Jul 30 16:44:52 Abhishek_: what is you github again? Jul 30 16:44:54 yes Jul 30 16:44:57 ah, stdin. that's fine... Jul 30 16:45:01 i don't have it bookmarked on this laptop Jul 30 16:45:28 jkridner : but I will hae a tcp thingy ready, shouldn't take much time :) Jul 30 16:45:34 Github.com/abhishek-kakkar Jul 30 16:46:09 mranostay: also do look at this week stat report Jul 30 16:46:20 jkridner : TCP is the native interface that Labview uses Jul 30 16:46:21 karki_: I think that can be find for the testing. I still feel we want to make some final changes before publishing and it seems they'll happen after you wrap-up the GSoC, namely moving to Abhishek_/panto's split remote_proc and moving the interpreter into a kernel module. Jul 30 16:46:58 karki_: I was expecting the TCP because it is what is used by the web interface, labview and you had it working in the past. Jul 30 16:47:13 jkridner : It will be done! Jul 30 16:47:14 Abhishek_: did you send that the gsoc list? Jul 30 16:47:43 karki_: I'm less worried about you doing that TCP interface than making sure your code is clean, modular, documented and free of bugs. Jul 30 16:47:58 mranostay: yeah Jul 30 16:48:12 jkridner : I can do both, and I will do it. (bugs worry me too :( ) Jul 30 16:48:14 karki_: adding a TCP interface would only be an intermediary step anyway as alexanderhaim and I both want to move the interpreter into the kernel. Jul 30 16:48:38 karki_: k, but don't make extra work. 10 days is almost no time at all. Jul 30 16:48:47 jkridner : sure :) Jul 30 16:48:57 alexanderhiam is a bit late :p Jul 30 16:49:18 * jkridner sees this meeting is going rather long. Jul 30 16:49:23 hey, sorry guys, had to catch a train Jul 30 16:49:33 connection on train may cut out, not sure Jul 30 16:49:46 Abhishek_: I'll let mranostay chat with you about the finalization items while we move on to DiegoTc. Jul 30 16:49:49 jkridner : so I'm done? have to leave for a few min..... Jul 30 16:50:06 jkridner: Personally I do not feel happy about wrapping the web interface with serious performance limitations. I hope to revisit the components at some point Jul 30 16:50:10 alexanderhiam: anything for karki_ while you have a connection? Jul 30 16:50:37 ok jkridner Jul 30 16:50:38 Abhishek_: sure, but I think you have to keep thinking about it as a learning tool and not something to replace 'sigrok' for a full featured analyzer. Jul 30 16:50:42 Abhishek_: don't see it.. can you forward to me? have to head to a meeting Jul 30 16:50:44 not that I can think of, karki_: I'll check out that pwm code soon Jul 30 16:51:03 jkrider, karki_: did you guys figure out the plan for thew web interface? Jul 30 16:51:11 DiegoTc: sorry about Monday. Jul 30 16:51:17 alexanderhiam: I Jul 30 16:51:40 jkridner: don't worry. I was late also. Got a flat tire Jul 30 16:51:43 alexanderhiam: I am saying he should stick to WaveDROM, despite it being slow, and wrap up the project by adding the integration. Jul 30 16:52:16 jkridner: wait, that's Abhishek_? I mean the botspeak interface Jul 30 16:52:35 DiegoTc: is https://github.com/replit/jsrepl meant to run code in the browser? Jul 30 16:52:46 Abhishek_: i'll look over your web code (not my area of skills but i can give input) Jul 30 16:52:46 jkridner: yes Jul 30 16:53:18 alexanderhiam: oh, I'm not seeing how with existing bugs the web interface will be done, but maybe I'm wrong. Jul 30 16:53:24 quite similar to jsbin or jsfiddle.net Jul 30 16:54:05 alexanderhiam: I'm thinking that karki_ needs to clean it up and we'll need to follow-up with the kernel module and using the code from the Tufts Chrome extension to make the web UI ourselves. Jul 30 16:54:35 alexanderhiam: although, I think without a solid code review, clean-up might not be all that productive. Jul 30 16:54:52 jkridner: sounds good, so karki_ should focus on the botspeak terminal and getting that ready for Chris then Jul 30 16:54:54 DiegoTc: do you have confidence you can implement it? does it have a terminate() feature? Jul 30 16:55:09 a code review is a good idea Jul 30 16:55:39 jkridner: I'm confident I can do it on these 10 days. About the terminate want to talk about it Jul 30 16:55:41 alexanderhiam: having it in any shape Chris can review ASAP for a functional review would be good and we can perform a code review of the same. Jul 30 16:56:02 you told me at the beginning there was an issue how it's implement it right now Jul 30 16:56:10 that you need to refresh the page Jul 30 16:56:15 for running it again Jul 30 16:56:16 agreed Jul 30 16:57:08 rseethamraju: I saw your message about the still image not working at the moment, did you convert to building the entire pipeline upon initialization and just pausing the sinks? Jul 30 16:58:03 yes but I think there’s some typo I made because its not working at all Jul 30 16:58:05 DiegoTc: /me reads http://stackoverflow.com/questions/195149/is-it-possible-to-sandbox-javascript-running-in-the-browser Jul 30 16:58:15 alexanderhiam: I’m debugging it Jul 30 16:58:33 DiegoTc: yes, that is why I'm looking for a sandboxing function that includes the ability to terminate the running script. Jul 30 16:59:00 I tried the whole pipeline in tee.py - figured what typo I made. It was on the linking part. Jul 30 16:59:06 reading it Jul 30 16:59:36 rseethamraju: does it seem like that strategy will work though? Jul 30 16:59:55 alexanderhiam: can’t say yet. 80% yes. Jul 30 17:00:03 cool Jul 30 17:00:37 but I made the whole pipeline in tee.py and the picture still doesn’t work even though the others do Jul 30 17:00:50 and it messes up the recording Jul 30 17:01:27 DiegoTc: I don't see a way around using a web worker. :( Jul 30 17:01:41 rseethamraju: I wouldn't be against limiting it so you can't take still images while recording and/or streaming Jul 30 17:01:49 alexanderhiam meaning the recording isn’t smooth anymore and keeps freezing Jul 30 17:02:11 jkridner: I will work on this today and tomorrow. (Pure code, not visual not anything just functionality) Jul 30 17:02:19 rseethamraju: which might mean revising the pipeline construction again Jul 30 17:02:24 when it's done let you know and start implementing it Jul 30 17:02:26 yes Jul 30 17:02:41 jkridner: how to do you want it showing? Jul 30 17:02:48 DiegoTc: k, but I don't want this to keep you from finishing up the rest. Jul 30 17:02:50 alexanderhiam: thats not a big deal though. won’t take more than an hour or two Jul 30 17:02:53 you want a button call run Jul 30 17:03:02 DiegoTc: ideally, there'd be run and stop buttons. Jul 30 17:03:08 rseethamraju: actually, let's make that the new goal, it should be quicker to get that working Jul 30 17:03:13 could be the same button (modal). Jul 30 17:03:26 ok Jul 30 17:03:27 alexanderhiam: I finished the accelerometer so I didn’t exactly spend all my time on camera today Jul 30 17:03:36 actually, I think a modal button would be best and that navigating away should always end it. Jul 30 17:03:48 jkridner: the current bugs are small. Not big deal, On miday I found then with Steve hangout Jul 30 17:03:49 rseethamraju: awesome, so that's all set and working? Jul 30 17:03:56 so that's not taking much time Jul 30 17:04:04 DiegoTc: I was thinking today that one thing we didn't handle was the case where you have both HTML and JavaScript in the same page providing a dynamic visualization. Jul 30 17:04:07 alexanderhiam: actually if it can take pics while streaming that should be enough Jul 30 17:04:37 page = card Jul 30 17:04:53 jkridner: ufff Jul 30 17:04:54 alexanderhiam: I think the issue is that the pic part of the pipeline ends but the other part is still running Jul 30 17:05:27 DiegoTc: like http://jsfiddle.net/jkridner/rC4AD/ Jul 30 17:05:38 alexanderhiam: oh yes! accelerometer all done. I’m finishing the documentation. Should take another 10 min or so. Jul 30 17:06:05 DiegoTc: if it is too much, I can look at adding it later. Jul 30 17:06:09 alexanderhiam: but the interrupt callback function that the user calls has to be standard right? Jul 30 17:06:10 rseethamraju: great! And all the interrupt sources are working? Jul 30 17:06:16 jkridner: that will take some time because there's an issue with the editor right now and the code view, Jul 30 17:06:21 Steve found it on Monday Jul 30 17:06:25 rseethamraju: standard? Jul 30 17:06:30 alexanderhiam: yes all interrupts working. I’ll show you tomorrow Jul 30 17:06:37 DiegoTc: k, better to fix that issue. Jul 30 17:06:49 I can work after the program on that if you don't mind Jul 30 17:07:01 DiegoTc: that'd be fine. Jul 30 17:07:10 alexanderhiam: I mean it should take (back_front, portrait_landscape, tap, shake) in that order, coz thats what I give back Jul 30 17:07:13 DiegoTc: best to get to a good bug-free stopping point. Jul 30 17:07:25