**** BEGIN LOGGING AT Wed Jul 11 03:00:03 2018 Jul 11 16:20:00 Meeting in 10 Jul 11 16:20:11 Hi everyone Jul 11 16:27:06 hi all Jul 11 16:29:23 Hi all Jul 11 16:29:35 'm Jul 11 16:30:52 Hello everyone! Jul 11 16:31:47 Hello all Jul 11 16:32:02 howdy ds2, ravikp7, hendersa, vaishnav98_! Jul 11 16:32:20 and abhishek_[m]... saw you in the back log. Jul 11 16:32:36 muneeb17: you there? Jul 11 16:33:33 Appears that muneeb17 is offline Jul 11 16:34:01 Hello everyone ! Jul 11 16:34:21 anirban1998[m]: you around? Jul 11 16:35:03 topic #1: evaluations.... due in just 1 day! Jul 11 16:35:20 None of them are completed! (including mine... I've got 2 to do) Jul 11 16:36:25 The Speak-n-Spell eval will be submitted within the next hour or two. Jul 11 16:36:33 jkridner: hi, I have made the changes suggested by you to the typescript definitions , working on bringing in mobile browser compatibility to the UI , by resizing and redrawing all the canvas on window resize event , will this work well? Jul 11 16:36:35 zeekhuge: or zeekhuge_ : there? Jul 11 16:36:50 If ZeekHuge doesn't fill them up today, I will Jul 11 16:36:59 summer's on peak.. Jul 11 16:37:47 abhishek_: yeah ! Jul 11 16:38:23 vaishnav98_: great on typescript. Jul 11 16:38:49 vaishnav98_: for the UI, I like having a single UI for BB and PB, but can you not make a fixed resolution and have it work on mobile devices? Jul 11 16:39:05 I thought we'd picked a resolution that several mobile devices worked with reasonably. Jul 11 16:39:08 dunno. Jul 11 16:39:35 no anirban1998[m] or muneeb17[m] around? Jul 11 16:40:48 jkridner : Have all 4 students completed their mentor evaluations? Jul 11 16:41:22 e have Jul 11 16:41:24 3 have Jul 11 16:41:41 muneeb17[m]: is only 25% complete with his evaluation Jul 11 16:41:47 Sorry jkridner... a bit late in the meeting Jul 11 16:41:58 er... I read that wrong. Jul 11 16:42:31 muneeb17[m] is the only student who has not completed his 2nd evaluation yet. Jul 11 16:42:35 Ouch. Will send him a reminder email. Jul 11 16:44:06 ravikp7: any updates or blockers to discuss? Jul 11 16:44:43 Hi all - sorry to be late - I'm a little "under the weather" Jul 11 16:44:48 Looks like muneeb17 is here. Jul 11 16:45:25 muneeb17: hey ! Jul 11 16:45:47 muneeb17: This is the second time I have to remind you to post your weekly report before the meeting - ideally Monday. Jul 11 16:45:53 abhishek_: zeekhuge Wormo Extremely sorry for not being available for the past few days. Had a family emergency and had to go to Bangalore. Jul 11 16:46:32 Will describe more in the mail Jul 11 16:46:46 I'm leaving for Hyderabad now Jul 11 16:47:45 jkridner: :) .. currently it looks okay on many devices , so will a proper handling of touch events be sufficient regarding mobile browser compatibility ? I remember you saying that the UI didn't look good on mobile devices a few days ago , that's why I was concentrating on resizing and redrawing the canvas, were you talking about a previous version? Jul 11 16:48:00 muneeb17: more travelling after Hyderabad ? Jul 11 16:48:15 muneeb17: Hope things are well, and there shouldn't be a 3rd reminder in the weeks after Jul 11 16:48:32 jkridner: working on tcp headers, it's taking some time to understand the tcp states working, like how to assign the sequence number, acknowledgement number etc. Jul 11 16:48:34 No, my home s in Hyderabad Jul 11 16:49:27 jkridner: fyi, I got only the cdc interface up for networking on linux and windows, so working with that interface Jul 11 16:49:27 I will reach in the he morning Jul 11 16:49:44 ravikp7: Isn't there something already available for you in NodeJs libraries. Did you do a search and check? Jul 11 16:49:50 muneeb17: okay. A short report is >> no report . Jul 11 16:50:53 Will be there at the meetup time tomorrow. Ok I will send it now. Jul 11 16:51:17 abhishek_: I went through all APIs, couldn't find something to get such tcp header level info, only ports, and payload are available to work with Jul 11 16:51:30 welsh There? Jul 11 16:51:50 ravikp7: In npm packages too? Jul 11 16:53:46 zeekhuge: abhishek_ got to go now now. Again, very sorry! Jul 11 16:53:57 we'll be awaiting the email then Jul 11 16:54:15 abhishek_: couldn't find enough on that too, if you can please do a search, incase I've missed something Jul 11 16:55:48 ravikp7: I found this - https://github.com/mafintosh/turbo-net Jul 11 16:55:57 Check if it is helpful to you. Jul 11 16:55:59 welsh I needed some feedback regarding the GUI part...so that can make some suitable improvements...the GUI as of now is simple but serves the purpose...if you have seen my short demo video which I mentioned in the report Jul 11 16:56:07 Wormo: did you discuss something with muneeb17 last time ? I don't see anything in the logs .. Jul 11 16:56:07 Just curious where we are on the last task .. Jul 11 16:56:54 (sorry, was afk) Jul 11 16:56:58 I sent the email and mentioned the pastebin here, didn't hear back probably he left soon after that Jul 11 16:57:27 thought that should get him unstuck, we'll see Jul 11 16:57:31 ravikp7: looks like it isn't very helpful? Jul 11 16:58:25 abhishek_: yes, doesn't seem to do much different than node built-in api Jul 11 16:58:25 yeah that was 3 days ago that I sent it, and his emergency came up "a few days" ago Jul 11 16:58:26 vaishnav98_: how are things going with BaconBits? Did you wire up something for testing? Jul 11 16:59:22 ravikp7: https://github.com/mafintosh/turbo-net/blob/master/src/turbo_net.c Jul 11 17:00:04 anirban1998: welsh had to step into a meeting, but I'll make sure he takes a look Jul 11 17:00:26 muneeb17: If you're still there, don't forget to fill up your evaluation, or you won't continue further in the program. Jul 11 17:00:36 cwicks Okay sure! Jul 11 17:05:53 jkridner: I could get the thumbwheel and RGB led working from the UI which was easy , will add that too as a part of the UI , was trying to get the 7 segment leds working , my set-up uses MCP23S17, was trying using shiftout could not debug the issue (did this about two weeks ago: didn't spend time on it after that) Jul 11 17:07:01 k. Jul 11 17:11:54 abhishek_, jkridner: Is low level tcp ever exposed to applictions that we require to make tcp header to be sent to the board? I think os just provides api to use the sockets. Are we left with options other than doing it manually in code? Jul 11 17:12:49 ravikp7: I don't understand the question. Jul 11 17:13:24 the OS gives us a socket interface and the board sends us IP packets.... Jul 11 17:13:45 if it is an IP packet not headed to a server of ours, then we should route it using the OS sockets interface... Jul 11 17:15:29 also, if we want to open a socket to the board, such as a TCP/IP socket to port 80 or port 3000 for HTTP, then we should probably make a socket listener service served locally and then bundle up the packets and send them to the board.... Jul 11 17:16:16 jkridner: when we've to send tcp packets to the board, we get only socket info from the os, not all the tcp header fields info, how we keep track of other tcp header info? Jul 11 17:16:25 and if IP packets are routed back to that socket (and our fixed IP), provide them back to our application instead of routing them. Jul 11 17:16:31 routed? Jul 11 17:16:45 ds2: is that not the right term? Jul 11 17:16:48 what about exposing things on the board via a SOCKS5 interface? Jul 11 17:17:08 ds2: no software changes on the board... Jul 11 17:17:09 right term, just seems like a lot of work... plenty of tools with SOCKS5 support Jul 11 17:17:13 jkridner: the trouble here is that instead of letting the OS handle the networking through it's stack, we are sending low level packets over USB. Jul 11 17:17:19 'k Jul 11 17:17:50 I don't have anything against this approach, it's more convenient than configuring the network interface every time it's brought up. Jul 11 17:17:54 ds2: we are just removing the OS adapters in order to provide 100% predictable behavior. Jul 11 17:18:00 there is the libsockisfy stuff that you can shim in with LD_PRELOAD/LD_LIBRARYPATH Jul 11 17:18:11 ds2: so, the host OS never sees the RNDIS adapter. Jul 11 17:18:45 abhishek_[m]: we are already doing that part with the other parts of the stack. Jul 11 17:18:47 jkridner: yes but exposing to the OS seems like a lot of work as you do need to fiddle with source addresses Jul 11 17:19:04 jkridner: whereas SOCKS is available Jul 11 17:19:19 anyways, just a thought Jul 11 17:19:55 jkridner: Would it be better for ravikp7 to use a readymade TCP/IP stack like lwIP and integrate a layer of nodeJS on top of it? Jul 11 17:21:23 I know there are many other C- TCP/IP stacks but lwIP was the first one to come up Jul 11 17:21:36 ds2: We are trying to eliminate the virtual Ethernet adapter on the host to avoid it being configured wrong and also avoid needing to configure the OS "connection sharing" (or whatever it is called on the particular OS) by having the apps initiate the traffic onto the network. Jul 11 17:22:14 jkridner: are you familiar with SOCKS ? Jul 11 17:23:35 I don't see why a whole TCP/IP stack needs to be implemented.... this is just IP. Source/dest address and port (can't remember where socket/port is specified). Jul 11 17:24:11 forcing a source addr is very OS specific... Jul 11 17:24:28 need a SOCK_RAW in most cases Jul 11 17:27:56 jkridner: the point is how can we get the whole tcp header from the os? since the tcp header contains info other than the sockets also which are necessary and not fixed Jul 11 17:28:54 ravikp7: for Linux, look at SOCK_RAW Jul 11 17:29:09 (and tap/tun but that is also Linux specific) Jul 11 17:31:13 jkridner: what work should I be concentrating on after getting the remaining features of the UI working? , how should I deal with the issue that adresses making sync/async functions explicit? shall I proceed with the promises + submodules approach as suggested here ?: https://github.com/beagleboard/bonescript/issues/39 Jul 11 17:32:20 ugh... have to run to another meeting. Jul 11 18:22:02 vaishnav98_: have you looked at the way I tried to rewrite bonescript with every "blocking" action becoming a new function and then creating a series of those functions? Does the promised approached allow us to focus on a single implementation and still optimize for async performance as well as handle synchronous return values for the old API. **** ENDING LOGGING AT Thu Jul 12 03:00:01 2018