**** BEGIN LOGGING AT Tue Aug 08 03:00:01 2017 Aug 08 04:14:03 jkridner: async and '_device' problems? are you sure you're pulling from my repo and not from your fork which is 39 commits behind? Aug 08 04:14:24 good question! Aug 08 04:14:41 origin https://github.com/ravikp7/node-beagle-boot (fetch) Aug 08 04:15:06 hmmm... was out of date though Aug 08 04:15:16 not sure how. Aug 08 04:17:49 jkridner: what should I check on OSX before running the hoRNDIS kext unload command from code, so that it works on systems even where it isn't installed also ? Aug 08 04:21:16 jkridner: havn't updated uboot binary yet to mount sd card when emmc is not present, just forgot about it, will do that Aug 08 04:21:51 k, your master DOES work. Aug 08 04:22:40 I think you can simply execute 'kextunload /System/Library/Extensions/HoRNDIS.kext' since you'll need to be running as root anyway. Aug 08 04:22:52 should fail without issues if you don't have it. Aug 08 04:24:54 jkridner: I think I should add as an instruction to run this command prior to running server, instead of adding this to server code, as root permissions would be handled differently in flashing app Aug 08 04:26:44 jkridner: I could find only two modules which can provide elevation in electron based app, sudo-prompt and electron-sudo Aug 08 04:26:46 https://www.npmjs.com/package/sudo-prompt Aug 08 04:26:56 https://www.npmjs.com/package/electron-sudo Aug 08 04:27:15 only sudo-prompt worked for me Aug 08 04:27:17 noticed it takes about 12 seconds to boot. thought about any ways to speed it up? Aug 08 04:27:20 ravikp7: k. Aug 08 04:27:35 ravikp7: what about adding it in test.js? Aug 08 04:28:18 jkridner: Tried that, but this module doesn't work in spwan or fork fashion, it supports only exec Aug 08 04:29:15 jkridner: so using it, we can't run sub process with elevated access from same executable Aug 08 04:29:45 jkridner: I looked into etcher code, to find how they handle this Aug 08 04:31:09 jkridner: they spwan the writer process and redirect its stdout and stderror using ipc communication to the parent process Aug 08 04:33:36 jkridner: I still need to make more things clear, seems little complex. will take their help if needed Aug 08 04:38:31 jkridner: for file transfer, earlier I implemented code in such a way that it didn't wait for outend transfer to finish first before initiating inend transfer, it was like outend(aysnc) inend(async). That way file transfer was twice as fast. But it was failing in a Windows PC that I tried, so implemented it in same way as other requests were handled Aug 08 14:21:01 this connection still live? much of my Internet access seems to have died upstream from the router. Aug 08 14:21:07 ok, back online this morning Aug 08 14:21:26 ravikp7: I just discovered http://git.denx.de/cgi-bin/gitweb.cgi?p=u-boot/u-boot-arm.git;a=blob;f=doc/README.NetConsole. Could be interesting. Aug 08 14:21:36 * jkridner is trying to do more and more with uboot. Aug 08 14:33:36 jkridner: so, we can access uboot console logs over network. how we can make use of that? Aug 09 01:18:21 ravikp7: I need to parse the tftp request filename. Aug 09 01:20:02 ravikp7: seems you don't currently detect the TFTP opcode. Aug 09 01:20:17 * jkridner examines https://tools.ietf.org/html/rfc1350 Aug 09 01:26:11 yeah, identifyRequest makes little sense to me. Aug 09 01:26:23 just looking at the size of the packet?!? Aug 09 02:07:49 ravikp7: the UDP state saved during the TFTP transfer is managed in a global way rather awkwardly. **** ENDING LOGGING AT Wed Aug 09 02:59:59 2017