whitequark | what do you think about having a convenient and directly searchable irc log archive for #qi-hw? | 00:12 |
---|---|---|
whitequark | it's just me being fed up with stupid html-only logs which get badly indexed by Google, and I wrote a proper solution, and maybe it will be useful for someone. | 00:13 |
whitequark | can be seen here: http://irclog.whitequark.org/ | 00:16 |
wolfspraul | we have a simple searchable one at en.qi-hardware.com/irclogs | 00:41 |
wolfspraul | uses logs2html by Marius Gedminas | 00:42 |
wolfspraul | but sure, if there is something better, let's try it out | 00:42 |
wpwrak | whitequark: i think most of us are just "always on" ;-) | 00:46 |
wolfspraul | whitequark: looks nice! :-) | 01:40 |
kristianpaul | not bad indeed | 02:06 |
kristianpaul | a single window to broswe is atractive too | 02:07 |
qi-bot | [commit] Xiangfu Liu: this commit break the ks7010 driver. (master) http://qi-hw.com/p/openwrt-xburst/ae1b777 | 07:13 |
kyak | xiangfu: have you tested if ks7010 works on your side? | 07:16 |
xiangfu | kyak, yes. just tested. | 07:16 |
kyak | did you check with WPA2 as well? | 07:16 |
xiangfu | kyak, no. I just run 'ifconfig -a' | 07:17 |
kyak | xiangfu: i think this is not sufficient.. i see the interface in ifconfig -a, but it doesn't actually work | 07:19 |
kyak | not with OPEN, WEP or WPA2 | 07:19 |
kyak | if you confirm that it really works on your side, then i would be sure that i'm having hardware issues with this card | 07:24 |
xiangfu | kyak, I will test that now. | 07:31 |
kyak | xiangfu: thanks! | 07:37 |
xiangfu | kyak, http://pastebin.com/0Rz566vM | 08:01 |
xiangfu | I can connect to an open network. | 08:01 |
xiangfu | I have to manually set ip address. | 08:01 |
xiangfu | but RX and TX are all 0. | 08:01 |
kyak | Access Point: Not-Associated | 08:02 |
kyak | this is because you are not actually connected | 08:03 |
kyak | and you are not connected because nothing can be sent or received, indeed | 08:03 |
kyak | when i try connecting via wpa_supplicant, i always get "no data received" | 08:04 |
kyak | or something like this | 08:04 |
kyak | "iwlist scan" says "no scan results" | 08:04 |
kyak | as far as i remember | 08:04 |
xiangfu | kyak, just reboot and try again. now it is working: | 08:06 |
xiangfu | http://pastebin.com/N6Qjnepd | 08:07 |
xiangfu | http://pastebin.com/gDfSpkTa | 08:07 |
xiangfu | oh. I have enable the DEBUG options of ks7010. now it's working. | 08:08 |
kyak | xiangfu: great! i'll try it at home with the latest image :) | 08:08 |
kyak | do you mean it only works with DEBUG option? | 08:09 |
xiangfu | kyak, maybe some 'timing' issue. I am not sure. more ourput mean there are some delay inside. :) | 08:12 |
xiangfu | http://pastebin.com/gTkaRc1w | 08:12 |
xiangfu | kyak, just recomple the ks7010 and re-install it. then reboot. | 08:12 |
xiangfu | kyak, iwlist scanning only give me one open network 'xiangfu' . there are a lot of WPA2 wireless around me. | 08:13 |
xiangfu | kyak, maybe I just enable this 'KS_WLAN_DEBUG=9' in the next release. at least it make the wifi card working. | 08:13 |
xiangfu | it needs some time to fix the ks7010 driver. the 'KS_WLAN_DEBUG=9' and another issue: ' SDIO IRQ handling for single irq under linux-3.0' | 08:15 |
kyak | xiangfu: i just hope this debug output is not going to console :) | 08:16 |
kyak | damn this card (or driver?) is weird.. | 08:16 |
kyak | so i guess it is currently working with OPEN networks and with DEBUG=9 only | 08:17 |
xiangfu | kyak, I disable the debug try again. now it is working without the KS_WLAN_DEBUG=9. | 08:21 |
xiangfu | now I try to reboot 5 times. see what happen. | 08:21 |
xiangfu | 1. ks79xx: firmware load failed !! retern code = 8 | 08:24 |
xiangfu | 2. works fine. | 08:24 |
xiangfu | 3. ks79xx: firmware load failed !! retern code = 8 | 08:28 |
xiangfu | seems there are some other SDIO wifi card. support by linux 3.0 | 08:28 |
xiangfu | 4. works. fine. | 08:28 |
xiangfu | 5 works fine. | 08:28 |
xiangfu | so the SDIO wifi card, hardware(or software) not stable. | 08:28 |
xiangfu | kyak, you makebe can modify the command line disable the 'console' , check 'fw_printenv' the 'bootargs' | 08:29 |
kyak | xiangfu: ok, thanks. Will try that at home | 08:55 |
whitequark | wolfspraul, kristianpaul: so I'll then let it flow | 11:20 |
whitequark | by the way, I think I'll convert all the old logs for the sake of convenience. that's quite trivial | 11:20 |
wpwrak | (irc log) hmm, contrast is extremely low. please have some mercy with us living in sunny places :) | 12:20 |
wpwrak | even with the blinds down, the calender is almost unreadable | 12:20 |
wpwrak | i like the overall look, though. tidy and modern. | 12:23 |
whitequark | wpwrak: sure | 12:27 |
whitequark | I'm almost done converting your archives | 12:28 |
whitequark | that python thingy chokes on some messages and emits garbage :/ | 12:28 |
whitequark | like here http://en.qi-hardware.com/irclogs/qi-hardware_2010-07-10.log.html at the bottom. | 12:28 |
wpwrak | the \e } of gbraad ? | 12:33 |
wpwrak | \'e even | 12:33 |
whitequark | wpwrak: yeah | 13:28 |
Artyom | hi KristianPaul! :) | 13:39 |
Action: C-Keen reflashes his nanonote | 14:22 | |
whitequark | ok, archive conversion is complete | 14:51 |
whitequark | http://irclog.whitequark.org/qi-hardware | 14:52 |
wolfspraul | whitequark: cool, great initiative, thanks a lot! | 14:52 |
C-Keen | who is logged there? | 14:53 |
wolfspraul | whitequark: how about the #milkymist and #homecmos logs? | 14:53 |
whitequark | wolfspraul: no problem | 14:53 |
whitequark | can you point me to urls? | 14:54 |
whitequark | ah, already found | 14:54 |
kristianpaul | Artyom, hey, gps clock working now wiht milkymist, still some freezes but for now it seems boot to bios ;) | 15:00 |
whitequark | C-Keen: what do you mean by "who"? The criteria for selecting channels? | 15:02 |
C-Keen | whitequark: I mean by "who" that stuff I type is not showing up in your logs | 15:02 |
whitequark | C-Keen: http://irclog.whitequark.org/qi-hardware/2011-11-17#1321542159; | 15:03 |
C-Keen | ah k, nvm :) | 15:03 |
whitequark | there was a small glitch due to conversion of old logs. a 40 Mb SQL file, mind you | 15:03 |
whitequark | wolfspraul: done | 15:18 |
C-Keen | hm how can I switch off the audible terminal bell again? | 15:18 |
whitequark | C-Keen: setterm -vbell off IRIC | 15:18 |
whitequark | *IIRC | 15:18 |
whitequark | ah no | 15:19 |
whitequark | setterm -blength 0 | 15:19 |
whitequark | oh no, my mysql is red hot | 15:20 |
whitequark | I guess the indexes were invented for a reason... | 15:20 |
C-Keen | thanks whitequark | 15:22 |
LunaVorax | Hello everyone! | 15:42 |
LunaVorax | How is everyone doing? | 15:42 |
whitequark | wpwrak: an attempt to make it more readable: http://kaunan.whitequark.org:4567/ | 15:54 |
wpwrak | nice works better, thanks !:) | 15:57 |
whitequark | ok, then I'll put that to production | 15:57 |
whitequark | done | 15:58 |
whitequark | wolfspraul, wpwrak: ok, it looks like someone finally needs that piece of software. so feel free to poke me with any features | 16:04 |
kristianpaul | a good feature will be cool to indentify thread talks in irc :) | 16:15 |
whitequark | kristianpaul: thread talks? | 16:17 |
whitequark | i.e. highlight a chain of messages replying to one another? | 16:18 |
whitequark | I doubt this could be done automatically unless all participants always put the name of the one they're talking with at the start of message | 16:20 |
whitequark | and not everyone has that habit | 16:20 |
wpwrak | semantical recognition with automatic summaries ? :) | 16:20 |
whitequark | I think that for 99% of conversation two summaries would suffice | 16:20 |
whitequark | like "bullshit" and "flamewar" | 16:21 |
whitequark | *of conversations on irc | 16:21 |
whitequark | so, that's not a very hard task | 16:21 |
wpwrak | hmm, i don't think it's quite so grim around here. but there are certainly channels where you'd be quite right :) | 16:22 |
whitequark | yeah, not here indeed :) | 16:22 |
whitequark | #qi-hw logs have circa 290000 messages | 16:25 |
whitequark | 203153, to be precise, including this one. | 16:26 |
kristianpaul | whitequark: yes | 16:33 |
kristianpaul | yes indded not here, but nice feature | 16:34 |
whitequark | kristianpaul: well that's trivial to implement | 16:34 |
kristianpaul | :D | 16:34 |
whitequark | might be a bit cpu-hungry, through, but not a lot I guess | 16:35 |
kristianpaul | hum | 16:36 |
whitequark | irclog currently costs me like $0.15 a day | 16:36 |
kristianpaul | i just now electricty bill at home is 10usd per month :) | 16:36 |
whitequark | it's hosed on a VDS | 16:36 |
whitequark | the whole VDS eats $0.5 | 16:36 |
whitequark | a day, that is | 16:36 |
whitequark | *hosted | 16:37 |
kristianpaul | i see | 16:37 |
whitequark | kristianpaul: http://irclog.whitequark.org/qi-hardware/2011-11-17 and click on some yellow arrows. | 18:07 |
whitequark | wpwrak: (semantic analysis) check out http://irclog.whitequark.org/ruby/2011-11-17#1321483534-chain; | 18:39 |
wpwrak | so where the part where it summarizes ? "<...200 lines of mindless drivel...>" or such :) | 18:40 |
whitequark | I only implemented semantic analysis :/ | 18:41 |
whitequark | note the absence of nicks on several lines through | 18:41 |
wpwrak | as in <nick> or as in nick: ? | 18:42 |
whitequark | the latter | 18:42 |
wpwrak | ah, i guess i would only notice when comparing with the original :) | 18:43 |
whitequark | well, there's a sophisticated algorithm which tracks the chains and tries to guess whether one had replied or not | 18:44 |
lars_ | not bad | 18:46 |
kristianpaul | nice | 18:46 |
wpwrak | does it reorder messages ? | 18:47 |
lars_ | and it is realtime | 18:47 |
wpwrak | let's see | 18:47 |
wpwrak | you still have to reload, i guess | 18:48 |
lars_ | yes | 18:48 |
Artyom | hi KristianPaul! :) | 19:01 |
kristianpaul | Artyom: hey | 19:02 |
Artyom | I've read about your success ^) | 19:03 |
Artyom | :) | 19:03 |
kristianpaul | not quite, i still not understand why system hangs when i ran a bios read command (mr 0xa0000000) | 19:03 |
Artyom | And I also have good news: I wrote e-mail to Carlos and he updated the code on github :) | 19:04 |
kristianpaul | but i guess or is a simple mistake when disabling the usb core or milkymist soc timings are not friendly with a 65535 clock | 19:04 |
kristianpaul | nice ! | 19:04 |
kristianpaul | so you have a milkymist soc with sdram suport for your board? | 19:04 |
Artyom | Now I could synthsis the SoC | 19:04 |
Artyom | I have again questions ;) | 19:05 |
kristianpaul | thats not a problem :) | 19:05 |
Artyom | Where is bios stored? In flash? Or in block ram? Or it depends on some settings? | 19:06 |
kristianpaul | it depends | 19:06 |
kristianpaul | it shoudl be on flash, | 19:06 |
kristianpaul | but as you saw the tdc-core port uses a block ram | 19:07 |
Artyom | And how is it copied to flash? | 19:07 |
kristianpaul | well, it depends on your board port :) | 19:07 |
kristianpaul | for examplo, for the Milkymist One board | 19:07 |
kristianpaul | there is something called fjmem | 19:07 |
kristianpaul | its a bitstream (open source of course) | 19:08 |
kristianpaul | that allow to write that NOR chip in the board | 19:08 |
kristianpaul | as there is notaditiona MCU on the pcb | 19:08 |
Artyom | And how this program works? What interface is used? | 19:08 |
kristianpaul | for M1 is urjtag | 19:09 |
Artyom | uart? | 19:09 |
kristianpaul | fjmem talks with urjtag | 19:09 |
Artyom | and what is urjtag? (didn't hear about it before) | 19:09 |
kristianpaul | is a tool basically to talk some jtag based programers (as i understand) | 19:11 |
Artyom | BTW bios is used only for testing? And if I need to run my program what should I do? | 19:11 |
kristianpaul | so you can talk flash chips, or even cpu for debugging | 19:11 |
kristianpaul | Artyom: bios couod be hacked to run your own programs.. :) | 19:12 |
kristianpaul | if you need run own program well, you have three options | 19:12 |
kristianpaul | 1. link it again mmilkymist libc and include proper board intialization code (bios have then in boot.c if remenber well) | 19:12 |
kristianpaul | 2. Link it agains rtems operating system (not evry recomended at early development stages?) | 19:13 |
kristianpaul | 3. Compile it for milkymist uclinux port | 19:13 |
kristianpaul | Artyom: is this program osgps? :-) | 19:14 |
kristianpaul | cause i think it can even be already "called" from the bios | 19:15 |
kristianpaul | it dint havent threads or such isnt? | 19:15 |
kristianpaul | have** | 19:15 |
Artyom | At first I want to run bios. Then something like hello_world. And finally | 19:15 |
Artyom | I want to test some OS | 19:15 |
Artyom | RTEMS at first | 19:16 |
Artyom | may be something else later | 19:16 |
kristianpaul | hello world, let see there is one template somewhere.. | 19:16 |
kristianpaul | http://www.lekernel.net/presentations/masteri2l/tp/tp_files.tar.bz2 | 19:18 |
Artyom | I also want to ask you about programs in "tools" folder. Where are they used? (what for?) | 19:18 |
kristianpaul | check the base directory then main.c | 19:18 |
kristianpaul | yes, well therare some tools | 19:19 |
kristianpaul | for example flterm is used as a uart terminal and also to load binaries to milkymist | 19:19 |
whitequark | wpwrak: what do you mean by reordering? | 19:21 |
whitequark | and yes, it does every operation on client except for changing the date and searching | 19:21 |
whitequark | 300 lines of ruby and roughly the same amount of js | 19:22 |
kristianpaul | makeraw is for migrating images so can be loaded by bios at startup like logo | 19:22 |
Artyom | uart terminal = receives and transmits data through uart? And what does it mean "loads binaries to milkymist"? It loads binaries to milkymist flash? | 19:22 |
kristianpaul | othes to generate crc i think.. | 19:23 |
kristianpaul | no load to ram | 19:24 |
kristianpaul | Artyom: | 19:25 |
kristianpaul | http://www.milkymist.org/socdoc/paper_overview.pdf | 19:25 |
kristianpaul | it uses its own handshake protocol | 19:25 |
kristianpaul | so the bios ask for boot and flterm talks to ti and transmit binary | 19:25 |
wpwrak | whitequark: (reorder) does it change the sequence of messages. e.g., put this one below your question instead of after the other dialog by Artyom and kristianpaul ? | 19:27 |
lars_ | as far as i can see not | 19:31 |
kristianpaul | DocScrutinizer: Hi, what you think about a 65535000 clock for a SoC, or what multiple of xx clocks are actually be? | 19:46 |
whitequark | wpwrak, lars_: no it doesn't. I considered implementing it, but as the detection is less than perfect (and will always be so), such a feature will be very inconvenient | 19:49 |
whitequark | and the way it works now, it does not prevent you from reading all possibly-incorrectly-not-detected messages | 19:50 |
DocScrutinizer | kristianpaul: sorry you lost me | 19:50 |
wpwrak | good :) machines subtly changing semantic connections are scary :) | 19:50 |
DocScrutinizer | and it seems my IRC bouncer or freenode is ill | 19:50 |
DocScrutinizer | :-/ | 19:50 |
kristianpaul | i was asking your opinion about what a a System On Chip clock is usually be, for example multiple of ie, 12 | 19:51 |
kristianpaul | arghh, usually is | 19:52 |
DocScrutinizer | dang feenode! chat.eu.freenode.net failed | 19:55 |
DocScrutinizer | kicking the eu. part fixed it | 19:56 |
kristianpaul | odd you still been pinged at DocScrutinizer | 19:56 |
kristianpaul | 14:56 CTCP PING reply from DocScrutinizer: 0.662 seconds | 19:56 |
DocScrutinizer | thids is direct login, not via ZNC bouncer | 19:56 |
kristianpaul | ah | 19:57 |
DocScrutinizer51 | this however was dead | 19:57 |
DocScrutinizer | -> #freenode, ttyl | 19:58 |
kristianpaul | he, what you do if can go back to the past http://lwn.net/Articles/467852/rss ? :) | 20:10 |
DocScrutinizer | k, freenode says one server here has hickup, nice | 20:11 |
DocScrutinizer | kristianpaul: sorry, I have no idea how to answer your question about clock | 20:12 |
DocScrutinizer | cpu clock, time clock? what was the question? | 20:13 |
kristianpaul | cpu clock | 20:13 |
DocScrutinizer | still no question detected | 20:13 |
DocScrutinizer | I think CPU runs usually on NNN MHz, which gets synthesized internally by a programmable PLL | 20:14 |
kristianpaul | what are common values for a cpu clock? do they need to be multiple of a specific number(s)? | 20:15 |
kristianpaul | yes | 20:15 |
DocScrutinizer | from a xtal master clock that can be anything from 1 .. 1000MHz | 20:15 |
DocScrutinizer | kristianpaul: what are valid values for cpu clock depends largely on what you can set the PLL dividers to | 20:16 |
DocScrutinizer | there's no general rule | 20:16 |
kristianpaul | ok,so thats a green light for my 65535000 Hz clock :-) | 20:17 |
kristianpaul | Thansk! | 20:17 |
DocScrutinizer | if the datasheet for your SoC agrees, then yes :-D | 20:17 |
kristianpaul | milkymist dont have datasheet yet but you pointed one more excuse for starting one :) | 20:17 |
DocScrutinizer | often secondary system clocks get generated from CPU clock by further dividers, and those can't handle arbitrary ratios to create a sane Example Given memory bus clock | 20:19 |
DocScrutinizer | in new designs however this concept is frequently deprecated as it complicated cpufreq | 20:20 |
DocScrutinizer | on S3C2442 it been a nightmare to adjust CPU clock without messing up for example UART by glitches in clock | 20:21 |
wpwrak | kristianpaul: hmm, he may underestimate leadership skills. the whole BSD world also suffered from fragmentation due to incompatible egos. not sure this would have been different if the lawsuit hadn't happened. | 20:22 |
kristianpaul | egos egos :-) | 20:22 |
wpwrak | kristianpaul: basically, BSD had all the old guys who had a turf to defend. Linux was something new, so few pre-existing hierarchies were imported. and linus never went egomaniac | 20:23 |
kristianpaul | yes acepting patches not from the club is important | 20:26 |
wpwrak | in those days, there was the expression "BSD du jour", suggesting the BSD forks would pop up on a daily basis. usually because some guys couldn't get along, so one forked and then could implement his ideas. in linux, this kind of forking never happened. if someone forks, it's usually a dead end and doesn't get much public attention | 20:29 |
DocScrutinizer | Whois DocScrutinizer51 | 20:29 |
DocScrutinizer | err | 20:29 |
wpwrak | DocScrutinizer: identity crisis ? :) | 20:32 |
DocScrutinizer | nah, checking server | 20:32 |
DocScrutinizer | ZNC went apeshit when chat.eu.freenode.net each time again pointed to same server with a hickup, and timeout on login resulted | 20:33 |
DocScrutinizer | now I'm on chat.freenode.net and got kornbluth.freenode.net (Frankfurt, Germany). So basically exactly what I wanted | 20:34 |
DocScrutinizer | had to restart znc "the normalk way" via initscript, as the prov instance was shell interactive to read debug prints, and that time I got some server in Budapest | 20:35 |
wpwrak | whitequark: maybe rename your logger so that "wh" doens't autocomplete to it instead of you ? a _ prefix (if that's allowed) may be good | 23:07 |
whitequark | wpwrak: oh well. I've missed that fact | 23:11 |
wpwrak | kewl | 23:13 |
wpwrak | whitequark: much easier to talk to you this way :-) | 23:13 |
whitequark | indeed | 23:16 |
wpwrak | wolfspraul: https://sharism.cc/shop/product_reviews.php?products_id=1 | 23:44 |
wpwrak | "The m.." and then the censor struck ;-) clicking on the name produces a cheerful greeting from PHP | 23:44 |
--- Fri Nov 18 2011 | 00:00 |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!