| 00:01.29 | Mirell | landley: Odd though about the Swiss Francs. |
| 00:06.28 | solar | Mirell: sorry. Anyway if you still want to see the example of mysqltcl in use. http://dev.gentoo.org/~solar/irc/bugzilla.tcl.txt |
| 00:16.28 | *** join/#edev RobotGuy (n=n7pkt@pool-173-50-247-236.ptldor.fios.verizon.net) |
| 00:32.33 | landley | is pondering writing an expect implementation in C. |
| 00:32.57 | landley | Mirell pointed me at one a few days ago, but it has a different syntax than most expects. |
| 00:33.39 | landley | 80% of it's netcat, and I've already got one of those in toybox... |
| 00:42.10 | solar | landley: there is already an expect-lite |
| 00:42.27 | solar | think the andriod phones use it for the ssh-shell |
| 00:45.49 | landley | solar: Um, I just googled for expect-lite, found the sourceforge page, downloaded it, and it's an expect script. |
| 00:45.52 | landley | Starts with #!/usr/bin/expect |
| 00:46.11 | landley | It's implemented in expect, so I'm not sure what exactly this thing is for. |
| 00:46.15 | landley | What were you referring to/ |
| 00:54.12 | Mirell | Hmm... |
| 00:55.51 | Mirell | landley: It's for doing expect, but with nicer syntax, really. |
| 00:57.07 | Mirell | Hmm... |
| 00:58.09 | Mirell | And phone interview with Fog Creek over. |
| 01:06.13 | landley | Mirell: how'd it went? |
| 01:07.04 | Mirell | landley: Went pretty much the full hour, but of course I have no idea if it was good or not. |
| 01:08.15 | Mirell | landley: It at least didn't go down in flames |
| 01:09.30 | Mirell | landley: He did seem to like my take on any design goal is figure out how to deal with pathological use cases first. |
| 01:18.57 | landley | That's pretty much finding the edges of the problem, yeah. |
| 01:19.06 | landley | So you actually talked to The Joel? |
| 01:19.24 | Mirell | landley: Nah, just one of the developers |
| 01:20.24 | Mirell | landley: http://www.reddit.com/r/AskReddit/comments/a7alu/how_much_actual_cooking_goes_on_in_a_corporate/ |
| 01:21.08 | solar | prpplague_afk: When you said you expected to see it hit 1200 today, I thought you were pretty crazy. Even tho spot went down to 1159 I see BD lowest bid price on an AB is 1240. Went up over $60 today alone. |
| 01:21.43 | solar | landley: naa I don't think whats on SF is the same expect lite. |
| 01:22.08 | solar | landley: I'll poke the driod users in about an hr if you want me to find out more |
| 01:29.02 | *** join/#edev lilalinux_away (n=lilalinu@ist.deswahnsinns.de) [NETSPLIT VICTIM] |
| 01:29.44 | *** join/#edev mpk (n=mpk@209.20.91.138) |
| 01:31.34 | landley | solar: so where's the expect lite you were referring to? |
| 01:33.24 | solar | it's not been an HOUR :P |
| 01:34.05 | solar | I'm waiting for Kingtaco to get off the caltrain to ask him for more details |
| 01:38.30 | landley | No rush, I'd missed that your earlier comment was directed towards me. :) |
| 01:38.43 | landley | (Twelve windows at once...) |
| 01:40.49 | solar | landley: if you wrote a min-expect what functionality would you give it? |
| 01:41.00 | solar | the use of $variables ? |
| 01:45.58 | landley | Well, to start with I just want "search for regex" "respond with $BLAH" pairs. |
| 01:47.21 | landley | I don't really care what the syntax is, I'm just used to expect-like things working off a config file instead of a half-dozen separate command invocations. |
| 01:48.40 | landley | I vaguely recall "set timeout", "expect", "send", and "interact" commands. |
| 01:49.14 | landley | Variables sound nice, though. |
| 01:50.20 | landley | Hmmm... possibly the separate commands approach is a good one and I was just confused by the -a -b -c many options syntax of the thing Mark showed me, I need to dig it up and look at it again... |
| 01:54.57 | landley | I very, very vaguely recall using a version a while ago where the TIMEOUT was an environment variable, and "expect" and "set" were command line commands... |
| 01:55.06 | landley | send, I mean. |
| 01:56.37 | landley | I might call 'em "xpect" and "xsend" and XPECT_TIMEOUT=30 to avoid the more obvious collisions? Dunno. |
| 01:57.04 | landley | "send" is part of mailutils-mh and nmh, but I don't care all that much. :) |
| 01:57.14 | landley | Dinner time cometh, back in a bit... |
| 01:57.18 | solar | good call. |
| 01:57.35 | landley | Good call on xpect/xsend, or on not caring about "send" in the mail utils? :) |
| 01:58.01 | solar | xpect/xsend to avoid name conflicts |
| 01:58.17 | landley | Sure, why not. |
| 01:58.30 | landley | Maybe I'll code it up tonight, the actual implementation's pretty trivial... |
| 01:59.14 | landley | Anyway, dinner now. |
| 01:59.41 | solar | yeah a simple one should be. I might be able to lend a hand in testing/replacing some of my existing expect scripts |
| 01:59.46 | solar | enjoy. |
| 02:43.28 | Mirell | Hmm... |
| 02:47.06 | Mirell | I seriously hate my new commute. |
| 03:10.08 | solar | how far is it and what method of travel do you use? |
| 04:01.14 | Mirell | 1.5 hour roundtrip, and car. |
| 04:14.31 | *** join/#edev bsdfox (n=h36sa@c-67-161-174-21.hsd1.ca.comcast.net) |
| 04:15.12 | solar | ouch |
| 04:21.17 | *** join/#edev florian_kc (n=fuchs@Maemo/community/contributor/florian) |
| 04:21.22 | *** join/#edev kergoth (n=kergoth@ip24-251-170-95.ph.ph.cox.net) |
| 04:28.48 | Mirell | Yep..Yep.. |
| 04:28.55 | Mirell | So Fog Creek would be interesting |
| 05:11.02 | *** join/#edev ohsix (i=ohsix@66.220.111.99) |
| 05:12.15 | Mirell | That's it, I need to go into work tomorrow and outline why they have to rewrite their entire build system... |
| 05:54.00 | *** join/#edev landley (n=landley@cpe-66-68-109-179.austin.res.rr.com) |
| 05:54.01 | *** mode/#edev [+o landley] by ChanServ |
| 05:55.43 | landley | The uClibc guys are doing musical rotating git branches. *whimper* |
| 05:56.05 | landley | On the bright side, I seem to have wielded cluebat to get the darn nptl merge at least somewhat unstuck. |
| 05:56.25 | landley | They're still doing the "X, Y, and Z must happen before it's merged" thing, but that doesn't matter anymore. |
| 05:56.36 | landley | Because there is now an NPTL branch that tracks master automatically. |
| 05:56.52 | landley | I.E. the merge has _been_ done, they just need to be taught to ignore the old "master" branch which is now flipping useless. |
| 05:57.22 | landley | Unfortunately, the new branch was origin/nptl_merge. And origin/nptl got renamed to origin/nptl_old, and then origin/merge_nptl got renamed origin/nptl. |
| 05:57.32 | landley | Meaning I just did a pull and I have ALL THREE of those in my repository and have no clue what is where. |
| 05:57.34 | landley | I hate git. |
| 05:57.50 | landley | Correction: I DESPISE every aspect of the git user interface. |
| 05:58.00 | landley | And don't care about the infrastructure. |
| 05:58.18 | landley | The high point of my git experience is the bits I'm apathetic about. |
| 06:00.03 | Ramiz_ | 0x011b will be 0x400 how ?? |
| 06:11.47 | solar | landley: nevermind on the expect-lite already existing. what the guy had described and what exists differ quite a bit. http://www.androlib.com/android.application.com-elegion-gui-qBBm.aspx |
| 06:12.14 | *** join/#edev mbuf (n=mbuf@118.102.130.6) |
| 06:40.46 | landley | Alas. |
| 06:41.01 | landley | I'm doing some design work to figure out what exactly I want the thing to do. |
| 06:41.16 | landley | If you're on tomorrow I might bounce ideas off you, right now I've got some plate-clearing to do first. |
| 07:23.39 | ZeZu | check the # of revisions might tell you :) |
| 08:26.29 | *** join/#edev linac (n=lin@114.114.116.140) |
| 08:50.36 | mnemoc | `expect` in C would be sweet |
| 08:59.02 | mnemoc | it would probably be pretty simple to implement doing a simple stdin eater with pattern matching and some sh helper functions |
| 10:01.34 | *** join/#edev SandGorgon (n=OmNomNom@122.160.41.129) |
| 10:49.15 | *** join/#edev eox425 (n=root@213.133.177.11) |
| 11:09.27 | *** join/#edev mbuf (n=mbuf@118.102.130.6) |
| 11:24.26 | *** join/#edev mapd (i=mapd@mapd2.portfrog.com) [NETSPLIT VICTIM] |
| 11:25.13 | *** join/#edev pirho (i=pirho@gateway/gpg-tor/key-0x2CEEC9CB) |
| 11:34.48 | *** join/#edev likewise (n=chatzill@82-171-51-231.ip.telfort.nl) |
| 12:21.19 | *** join/#edev gustavoz (n=gustavoz@host108.201-253-243.telecom.net.ar) |
| 14:30.49 | *** join/#edev mbuf (n=mbuf@122.169.132.210) |
| 15:05.44 | chouimat | morning |
| 15:09.08 | prpplague | chouimat: greetings |
| 15:09.36 | chouimat | prpplague: I hate being awaken by a phone call ... |
| 15:09.46 | prpplague | indeed |
| 15:10.14 | chouimat | but it was the fedex guy ... so it's ok I guess |
| 15:31.27 | *** join/#edev GPSFan (n=kenm@64.92.145.112) |
| 15:31.27 | *** mode/#edev [+o GPSFan] by ChanServ |
| 16:04.54 | Mirell | Hmm... |
| 16:05.22 | Mirell | Fog Creek wants me to move on to step two... |
| 16:10.50 | *** join/#edev mbuf (n=mbuf@122.169.132.210) |
| 16:19.55 | *** join/#edev SpeedEvil1 (i=1000@tor/regular/SpeedEvil) |
| 17:05.46 | *** join/#edev VK4MSL (n=vk4msl@gentoo/developer/redhatter) |
| 17:06.51 | *** join/#edev SandGorgon (n=OmNomNom@122.173.249.152) |
| 17:12.19 | *** join/#edev Madsy (n=Madsy@88.90.30.138) |
| 17:25.34 | *** join/#edev JoeLlama (n=snork@206.169.63.10) |
| 17:40.29 | Mirell | So sleepy... |
| 17:40.57 | chouimat | gives some coffee to Mirell |
| 17:47.21 | Mirell | Just took a Foosh mint... |
| 17:50.31 | likewise | Mirell: Fog Creek as in you the customer, or you the employer? (I was a customer) |
| 17:50.49 | likewise | employee (sp?) |
| 17:51.54 | likewise | employee is what I meant (was temporarily low on caffeine myself while accessing English vocabulary) |
| 18:19.54 | *** join/#edev CosmicPe1guin (n=nobody@66-162-99-237.static.twtelecom.net) |
| 18:26.57 | *** join/#edev Madsy (n=Madsy@fu/coder/madsy) |
| 18:27.14 | *** join/#edev ja2 (n=a@81.233.153.190) |
| 18:29.18 | ja2 | Hi, Is there any expertise on PLL in here? Is there any rules of thumb for the values of feedback dividers. Eg. I got two feedback dividers, and two "multipliers". Should values be choosen in any particular way? Im not good with this, maybe I use the wrong terminology. |
| 18:30.03 | SpeedEvil | Is this a multi-stage PLL? |
| 18:30.18 | ja2 | Think so. EP9307, cirrus. |
| 18:30.32 | SpeedEvil | I mean - does it generate an intermediate frequency - and then the second PLL uses this as a clock? |
| 18:30.42 | SpeedEvil | Carefully read the datasheet. |
| 18:30.58 | SpeedEvil | You may get less noise at some divisor ratios |
| 18:31.12 | SpeedEvil | Or faster lock time, or ... |
| 18:31.53 | ja2 | Yes, there's PLL1 and PLL2. My hope was to get less noise. My hypothesis is square values to reach my desired freq. But maybe a low first stage is better? |
| 18:32.10 | SpeedEvil | If the datasheet doesn't specify it - you can't really tell. |
| 18:32.36 | SpeedEvil | Carefully check for anything saying 'PLL1 under 24MHz is not recommended' or ... |
| 18:32.47 | SpeedEvil | But I'd first tend towards square too |
| 18:33.12 | ja2 | They name specific frequencys, but never how to calc the freq from the actual parameters/dividers. They only give the total formula for input freq->output freq. |
| 18:33.27 | ja2 | Yes, there is a lot of constricts like that. |
| 18:33.49 | ja2 | But running the chip outside seem better than some values inside the constrict!!! |
| 18:34.02 | SpeedEvil | Also - in practice - you can't guarantee it will be better than worst case of the spec sheet. |
| 18:34.26 | ja2 | When do you satisfy with an empirical value? |
| 18:34.26 | SpeedEvil | For example - over in ##electronics someone had real fun with attinys (IIRC) |
| 18:34.38 | SpeedEvil | The clock was specified as +-0.5% |
| 18:34.40 | SpeedEvil | (IIRC) |
| 18:35.08 | ja2 | oh, they might be a good source for info. I never talked to them. |
| 18:35.12 | SpeedEvil | Using the onboard clock to generate PWM to run a servo motor worked fine on one silicon revision. As the oscillator was stable. |
| 18:35.23 | SpeedEvil | Well - it was a constant frequency drifitin with temp |
| 18:35.40 | ja2 | So, do PLL drift to? |
| 18:35.43 | ja2 | too |
| 18:35.48 | SpeedEvil | The silicon revision then happened, and there was a clock-clock jitter of .5% |
| 18:35.52 | SpeedEvil | It was still in spec. |
| 18:36.07 | SpeedEvil | It's just that the second made the product fail in the field. |
| 18:36.16 | ja2 | Auch. |
| 18:36.26 | ja2 | Difficult bug to find. |
| 18:36.44 | ja2 | I got failing USB communication now. |
| 18:36.47 | SpeedEvil | No - this wasn't a PLL - it's just an example that you should always assume the worst possible interpretation of a datasheet - unless you buy only one lot. |
| 18:37.21 | SpeedEvil | And that even if it works for this batch, it may not work for hte next. |
| 18:37.26 | *** join/#edev RobotGuy (n=n7pkt@pool-173-50-247-236.ptldor.fios.verizon.net) |
| 18:37.37 | ja2 | Is there any use of running PLL at double freq, and divide it by two or more for every unit using it. Or should you go for as low PLL freq as possible, just dividing by 1 where appropriate? |
| 18:37.53 | SpeedEvil | ja2: Does it not work at the recommended figures? |
| 18:38.05 | SpeedEvil | ja2: you've checked obvoious - decoupling and ... ? |
| 18:38.31 | ja2 | No. The design of the board is poor, so there's a lot of sources for error in my case. |
| 18:39.12 | ja2 | I just want it to work somewhat. I dont mind sacrifising top speed to get better function. |
| 18:39.29 | ja2 | And lowering freq seems to work out. |
| 18:39.33 | ja2 | I just want to trim now. |
| 18:41.11 | ja2 | But I never had any deep knowledge about PLL, and this unit seems pretty complex. So i figure I better get some rules of thumb atleast. :) |
| 18:48.59 | Redhatter | I would have thought leaving a clock as-is rather than multiplying it would be better for stability. |
| 18:49.18 | *** join/#edev flo_lap (n=fuchs@Maemo/community/contributor/florian) |
| 18:49.20 | Redhatter | Also better for EMC, since a lower frequency clock will have fewer harmonics. |
| 18:51.47 | ja2 | Redhatter: Good points. The CPU is feed with ~14 Mhz, this is multiplied and divided via PLL to give the operating speed of 200 Mhz for this CPU. |
| 18:52.45 | Redhatter | Is that ~14MHz based on a crystal? |
| 18:52.56 | ja2 | Yes. |
| 18:53.14 | Redhatter | Ahh good... because an LC oscillator will be dancing all over the place. |
| 18:53.27 | ja2 | I still have thoughts about how much that xtal vary. |
| 18:53.46 | Redhatter | Should be a few parts per million... bugger all. |
| 18:54.01 | SpeedEvil | layout is important a bit though. |
| 18:54.07 | SpeedEvil | And the right values of caps. |
| 18:54.15 | Redhatter | Indeed |
| 18:54.17 | ja2 | so that little error shouldnt cause any harm even when multiplied 30 times in the PLL. |
| 18:54.18 | SpeedEvil | (well - not grossly wrong values) |
| 18:54.36 | SpeedEvil | ja2: you're getting problems due to clock errors? |
| 18:54.53 | ja2 | Infact, I do not know for sure. |
| 18:55.09 | *** join/#edev archae0pteryx (n=snewman@207.47.42.130.static.nextweb.net) |
| 18:55.38 | ja2 | One thing I just noted is that is the clock for the peripherials, PCLK, is divided by a "large" value (4-8), USB communication fail at points. |
| 18:55.54 | ja2 | There are some registers (DMA) which have restrictions |
| 18:56.13 | ja2 | for how many clk cycles they must wait, depending on the PCLK divider settings. |
| 18:56.25 | ja2 | So there might be pure SW errors. |
| 18:57.13 | ja2 | It is difficult to look into this, as errors occur on random occations. |
| 18:57.42 | ja2 | I run linux on the board and usually I can force error by dd a large file. |
| 19:02.57 | SpeedEvil | you've rechecked the wait-states and timing on your RAM? |
| 19:03.57 | ja2 | No, I didnt. Will do. |
| 19:09.53 | *** join/#edev roxfan2 (i=dunno@87.67.249.82) |
| 20:22.57 | *** join/#edev davidc__ (n=davidc__@user-387g3od.cable.mindspring.com) |
| 20:36.26 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 21:25.58 | *** join/#edev eox425 (n=root@213.133.177.11) |
| 21:35.22 | *** join/#edev archae0pteryx (n=snewman@207.47.42.130.static.nextweb.net) |
| 21:53.05 | *** join/#edev kergoth (n=kergoth@ip24-251-170-95.ph.ph.cox.net) |
| 22:05.08 | *** part/#edev kergoth (n=kergoth@ip24-251-170-95.ph.ph.cox.net) |
| 22:05.26 | *** join/#edev kergoth (n=kergoth@ip24-251-170-95.ph.ph.cox.net) |
| 22:06.13 | *** join/#edev GPSFan (n=kenm@64.92.145.112) |
| 22:06.14 | *** mode/#edev [+o GPSFan] by ChanServ |
| 22:28.39 | RobotGuy | prpplague: ping |
| 22:47.19 | prpplague | RobotGuy: pong |
| 22:51.07 | RobotGuy | Ah, hi there! Question about the trainer board for Beagle.. Would it be possible (and reasonable) to include the option to use I2C to communicate with the Atmega328 instead of using the second serial port? |
| 22:52.57 | prpplague | RobotGuy: you can make the change by cutting the trace connections for the uart and solder in the i2c |
| 22:54.16 | RobotGuy | I just think it would be a useful selectable option to have included right on the board ready to go. :) |
| 22:54.26 | RobotGuy | Less required soldering is better, at least for me. :) |
| 22:54.50 | RobotGuy | (not that I can't do it, which I can if needed) |
| 22:54.51 | prpplague | i'll look, but i think we've already sent the board out for production |
| 22:55.17 | RobotGuy | Ah, OK. Well if so, maybe it could be considered for a later revision. |
| 22:56.01 | RobotGuy | I really need to use the second serial port for my servo controller. |
| 22:56.05 | prpplague | if sales for the zippy don't pick up i doubt there will be another revision |
| 22:56.21 | prpplague | rusty and i have pretty much given up on the hobby market, just no money in it |
| 22:57.03 | RobotGuy | That would be too bad. I think the Zippy should be a required item for Beagle. Too bad on giving up on that. I don't get why people don't see the need for these boards. |
| 22:57.21 | RobotGuy | You are making Beagle features easy to get at and use. |
| 23:00.30 | RobotGuy | I guess people are more into making their own custom boards to use with Beagle. |
| 23:01.15 | RobotGuy | I must be part of a very small minority. |
| 23:03.45 | *** join/#edev chiques (n=working_@pool-71-189-111-152.lsanca.fios.verizon.net) |
| 23:06.24 | Mirell | prpplague: I'm really annoyed at this point with just a Keurig brewer in an office environment |
| 23:06.27 | Mirell | prpplague: It takes too long. |
| 23:08.05 | prpplague | Mirell: hehe |
| 23:08.41 | prpplague | RobotGuy: no, most people aren't doing any expansion at all for the beagle, only with usb devices |
| 23:09.50 | RobotGuy | Ah. I'll be using I2C, UART, and USB stuff. |
| 23:10.38 | RobotGuy | I may end up going with a Gumstix Overo for my final robot brain to get the built in wifi and bluetooth. |
| 23:11.10 | RobotGuy | Beagle is a great dev platform for me though, and I am getting along well with OE. :) |
| 23:12.25 | RobotGuy | I want to at least get a Zippy and Trainer expansion board, and ShowDog too if it gets to see daylight. |
| 23:12.53 | RobotGuy | I still have those displays I got from you sometime back. |
| 23:14.17 | davidc__ | prpplague: If you want a suggestion; I think you need to market it better - the only place I've ever seen your stuff mentioned is here :( |
| 23:14.30 | davidc__ | prpplague: build something cool with it and submit to hackaday |
| 23:15.55 | prpplague | davidc__: yea, i hear that alot, problem is i'm not a marketing guy, i can build it, but not market it |
| 23:17.13 | davidc__ | prpplague: my suggestion would be "build something cool" |
| 23:17.26 | davidc__ | and then submit said cool thing to make, hackaday, hell, if its cool enough, slashdot. |
| 23:17.56 | prpplague_afk | davidc__: i'm too mundane when it comes to that stuff |
| 23:17.56 | davidc__ | if you do a really good project writeup, and get it in make, or one of the hobbyist electronics magazines |
| 23:18.01 | davidc__ | hell, it doesn't even need to be too crazy |
| 23:18.04 | davidc__ | just build a robot! |
| 23:18.07 | prpplague_afk | hehe |
| 23:18.07 | davidc__ | a robot that runs linux! |
| 23:18.18 | prpplague_afk | has to go take his step daughter to get her hair done |
| 23:18.18 | prpplague_afk | bbiab |
| 23:18.22 | davidc__ | alright, talk to ya then |
| 23:23.06 | *** join/#edev JoeLlama (n=snork@206.169.63.10) |
| 23:26.15 | RobotGuy | davidc__: I am going to use a BeagleBoard (or similar) for the brain of my robot. |
| 23:27.05 | RobotGuy | I just need to figure out if I will have the resources I need as far as hardware ports and such. |
| 23:32.13 | RobotGuy | I've already figured out how to power Beagle from a battery supply, thanks to SparkFun. I still think TCT should get to be a SparkFun vendor - might be just the ticket to get TCT stuff into more hobbiest's hands. |
| 23:33.00 | RobotGuy | SparkFun carries the BeagleBoard now. |
| 23:36.57 | *** join/#edev archae0pteryx (n=snewman@207.47.42.130.static.nextweb.net) |
| 23:52.02 | Mirell | RobotGuy: I ended up getting a USB Magical Battery thing to power the BeagleBoard from USB |
| 23:52.20 | likewise | #dies_felices: you wrote |
| 23:52.33 | likewise | sorry |
| 23:53.05 | RobotGuy | Mirell: Hi! I'm going to run my battery into one of the SparkFun breadboard power supplies with a barrel plug coming off where the breadboard pins would be. |
| 23:53.51 | RobotGuy | More likely, I will use that to power the little USB hub with ethernet I got from Special Computing, and power Beagle from that. |
| 23:59.12 | Mirell | Ahhh, nice. |