From: Sam Moore Date: Wed, 14 Aug 2013 17:00:07 +0000 (+0800) Subject: Automatic commit of irc logs X-Git-Url: https://git.ucc.asn.au/?a=commitdiff_plain;h=39f3324245b287e72696a25807e6c3129e51e511;p=matches%2FMCTX3420.git Automatic commit of irc logs --- diff --git a/irc/log b/irc/log index 8c7c90e..103a54c 100644 --- a/irc/log +++ b/irc/log @@ -716,3 +716,153 @@ 20:49 -!- jtanx [~asfa@124-169-120-181.dyn.iinet.net.au] has quit ["ChatZilla 0.9.89 [Firefox 22.0/20130618035212]"] 20:56 -!- jtanx [~asfa@124-169-120-181.dyn.iinet.net.au] has joined #mctxuwa_softdev 22:40 -!- jtanx [~asfa@124-169-120-181.dyn.iinet.net.au] has quit ["ChatZilla 0.9.89 [Firefox 22.0/20130618035212]"] +--- Day changed Wed Aug 14 2013 +07:58 -!- jtanx [~asfa@124-169-120-181.dyn.iinet.net.au] has joined #mctxuwa_softdev +09:13 < jtanx> I've done a version using fastcgi, which was surprisingly quite easy to set up +09:13 < jtanx> see here: https://github.com/jtanx/MCTX3420/tree/master/testing/fastcgi-approach +10:41 < sam_moore> That looks interesting, I'd prefer that to the PHP/database solution +10:42 < sam_moore> My concern is: You have to query the web server to start the process, even though it keeps running +10:42 < sam_moore> Also, I know "write a custom HTTP server" sounds scary, but it does actually work +10:43 < jtanx> nah +10:43 < jtanx> fastcgi +10:43 < jtanx> the process is started once and always runs +10:43 < jtanx> so you have a request loop +10:44 < jtanx> wait +10:44 < jtanx> how do you envision it working? +10:45 < sam_moore> Oh, right, I see how the fastcgi approach would work, it's basically the same idea; you have something responding to HTTP queries in one thread, something dealing with sensors/actuators in seperate threads +10:45 < jtanx> yeah +10:46 < jtanx> except that you don't have to worry about maintaining your own webserver +10:46 < sam_moore> Cool +10:46 < sam_moore> I do think that the HTTP server isn't that hard, and it's mostly working +10:46 < sam_moore> But we can probably switch to fastcgi +10:48 < sam_moore> I mean, we still have to parse a query string right? +10:48 < jtanx> well yes +10:48 < jtanx> if your request was +10:48 < jtanx> http://blah/cgi/?key=value&anotherkey=value2 +10:48 < jtanx> you'd get back key=value&anotherkey=value2 +10:48 < jtanx> as query string +10:50 < jtanx> try it out here: http://124.169.120.181:8080/cgi +10:51 < jtanx> (hopefully that forwards ok) +10:51 < sam_moore> Yep, that's cool +10:51 < sam_moore> The HTTP server I wrote basically does that +10:51 < sam_moore> But if it really bothers you, I'm fine with switching to fastcgi :P +10:52 < jtanx> well whatever works +10:52 < jtanx> but I thought it might be more maintainable if you used something that already exists :P +10:52 < sam_moore> Yeah, but thinking about it, what maintainence does the HTTP server require? +10:53 < sam_moore> Oh well, it's a good idea +10:53 < jtanx> for this situation not much +10:57 < sam_moore> We can pretty easily start with a structure that would allow us to switch, but I'd lean towards keeping the custom HTTP server +10:58 < sam_moore> Have you done multithreading in C before? +10:58 < jtanx> in windows yes +10:58 < jtanx> on linux not really +10:58 < sam_moore> Ah, I've only done it on linux +10:58 < jtanx> i did a bit in operating systems +10:59 < jtanx> but kinda forgot +10:59 < sam_moore> It shouldn't be too hard for our purposes +11:00 < jtanx> that thing with running a custom webserver is that if anyone wants to reuse this in the future, i doubt they'd really like the idea of that +11:01 < jtanx> say they had some new fangled thing that requires php +11:01 < jtanx> which isn't going to work +11:01 < jtanx> but with a standard webserver like apache or nginx, it's really easy to just install an extra addon +11:02 < sam_moore> I suppose... what would they need the new fangled php thing for? +11:03 < jtanx> well that's the thing - for now I don't know, but it could be required +11:03 < sam_moore> Yeah, good point +11:03 < sam_moore> Ok, another reason to use an existing web server, is we might require some advanced features if we want to be serious about the safety stuff +11:04 < sam_moore> For example, it would be a fair amount of work to add SSL to make the custom thing do https +11:04 < jtanx> oh yeah +11:04 < jtanx> that too +11:04 < sam_moore> Also we might want some authentication on it +11:04 < jtanx> nginx/apache are highly customisable +11:05 < sam_moore> I've never used nginx, what are it' +11:05 < sam_moore> s advantages over apache2? +11:05 < jtanx> nginx +11:05 < jtanx> yup +11:05 < jtanx> its really good +11:05 < sam_moore> Haha +11:05 < jtanx> a lot of sites use it because it's fast +11:05 < sam_moore> Fast is good +11:05 < jtanx> faster than apache +11:05 < sam_moore> Sold +11:05 < sam_moore> Well, we only need one user as well +11:05 < sam_moore> In fact, we deliberately do not want to have multiple clients able to use the thing at the same time +11:05 < jtanx> haha true +11:07 < sam_moore> Ok, so I think you've convinced me to use fastcgi +11:07 < sam_moore> Most of the custom HTTP server was reused code, so I didn't waste too much time on it +11:07 < jtanx> haha ok +11:07 < jtanx> have you written one before? +11:09 < sam_moore> Not specifically a HTTP server, but I've written a project for posix systems with a fair bit of networking +11:09 < sam_moore> git.ucc.asn.au/?p=matches/swarm.git +11:10 < jtanx> nice +11:10 < sam_moore> It allows you to run a shell, eg: bash accross multiple machines at once +11:11 < sam_moore> Not as useful as I thought it might be, but it was fun +11:12 < jtanx> heh +11:12 < jtanx> what did you end up using it for? +11:13 < sam_moore> I was setting up a programming competition where you have AIs play against each other +11:13 < sam_moore> I wanted to make a round robin +11:13 < sam_moore> So I had a bash script setup to do that, but then I got impatient, because it would have to run the games in order +11:14 < jtanx> oO +11:14 < sam_moore> Well it turned out no one entered the competition, so I probably didn't need to go overboard +11:14 < sam_moore> But maybe I can use it for something else +11:15 < sam_moore> For the threading I'm thinking of using pthreads +11:16 < sam_moore> The alternative is OpenMP, but it's more suited to numerical computation, rather than having seperate things talking to each other +11:16 < jtanx> okay +11:17 < sam_moore> My goal for this week was to have the framework started, so we just have some really basic threads running and then we can start implementing them +11:17 < jtanx> yeah that would be good +11:18 < sam_moore> If you want to do some work with the fastcgi part so that we'll be able to parse the query strings, that would probably be a good start +11:18 < sam_moore> We don't have the exact details of pretty much anything yet +11:18 < sam_moore> Maybe talk to James/Rowan about how the GUI is going to query the server +11:19 < sam_moore> Also, did you see my suggested naming convention? +11:19 < jtanx> no +11:20 < sam_moore> I don't really mind, although I'd prefer not to have hungarian notation +11:20 < sam_moore> As long as it' +11:20 < sam_moore> s consistent +11:21 < sam_moore> But the suggestion was: FunctionName, variable_name (local or member), Structure, ENUMVALUE, Extern_FunctionName, g_global +11:21 < jtanx> okay I'll try to keep to that +11:22 < sam_moore> Cool, I should probably go do something else +12:33 < jtanx> i gotta go +12:33 -!- jtanx [~asfa@124-169-120-181.dyn.iinet.net.au] has quit ["ChatZilla 0.9.89 [Firefox 22.0/20130618035212]"] +15:39 -!- jtanx [~asfa@124-169-120-181.dyn.iinet.net.au] has joined #mctxuwa_softdev +16:22 -!- justin_kruger [~justinkru@125.253.101.228] has joined #mctxuwa_softdev +16:23 -!- justin_kruger [~justinkru@125.253.101.228] has quit [EOF From client] +16:43 -!- jtanx_ [~asfa@124-169-120-181.dyn.iinet.net.au] has joined #mctxuwa_softdev +16:56 < jtanx_> a beagleboard +16:56 < jtanx_> hmm +16:58 -!- jtanx [~asfa@124-169-120-181.dyn.iinet.net.au] has quit [Ping timeout] +16:58 -!- jtanx_ is now known as jtanx +17:03 < jtanx> is it the beaglebone black? +17:03 < jtanx> sounds interesting +17:04 < sam_moore> They didn't specify +17:04 < sam_moore> Oh right, I accidentally cut out his message +17:04 < sam_moore> tl;dr "Beagle Board" +17:04 < jtanx> the beagle board is around $125 though +17:04 < sam_moore> Probably the Beaglebone black +17:04 < sam_moore> Oh really +17:04 < sam_moore> Heh +17:04 < jtanx> yeah +17:05 < jtanx> the beaglebone black is $45 +17:05 < jtanx> but just looking the A/D converter on the black has a max voltage of 1.8v +17:05 < sam_moore> That's probably what they meant, since they said it is "not much more expensive" +17:05 < sam_moore> That should be fine; the sensors team says things are in mV +17:05 < jtanx> okay +17:05 < jtanx> that's good +17:05 < sam_moore> Really though... electronics should be asking the sensors team stuff like this +17:06 < jtanx> heh yeah +17:06 < sam_moore> They should probably have asked us if it was feasable for software +17:06 < sam_moore> Maybe they had someone who knew it would be, but still +17:06 < sam_moore> It looks like it is anyway, so crisis averted +17:07 < jtanx> just with the sensors, I think that the diferences need to be amplified so it covers the full range of the A/D converter +17:07 < sam_moore> Yes, the sensors guy knows that +17:07 < jtanx> yeah so if its only 1.8v there's less range than 0-5v +17:08 < jtanx> is that compensated enough by having a 12 bit a/d converter vs a 10 bit (i think arduino is 10 bit) converter +17:10 < sam_moore> I think so; you get ~4x more resolution from the 12 bit ADC and lose ~3x from the lower range +17:10 < sam_moore> Also there's no set requirement yet on what the resolution should be +17:11 < jtanx> true +17:11 < jtanx> well in any case the bb black sounds quite nice +17:11 < sam_moore> Yep, it'll probably be good to use +17:14 < sam_moore> I'm much happier now that we actually have regular verbal communication with the other teams +18:23 < sam_moore> I'm looking into getting doxygen working for this +18:59 < jtanx> for the documentation? +18:59 < jtanx> seems good, but never used it before myself +19:01 < jtanx> looks a lot like javadoc +21:26 -!- jtanx [~asfa@124-169-120-181.dyn.iinet.net.au] has quit ["ChatZilla 0.9.89 [Firefox 22.0/20130618035212]"] +21:34 -!- jtanx [~asfa@124-169-120-181.dyn.iinet.net.au] has joined #mctxuwa_softdev +22:54 -!- jtanx [~asfa@124-169-120-181.dyn.iinet.net.au] has quit ["ChatZilla 0.9.89 [Firefox 22.0/20130618035212]"]