2013-10-27 00:04 pcercuei has joined #qi-hardware 2013-10-27 00:41 viric has quit [Ping timeout: 240 seconds] 2013-10-27 00:43 viric has joined #qi-hardware 2013-10-27 01:00 viric has quit [Remote host closed the connection] 2013-10-27 01:00 viric has joined #qi-hardware 2013-10-27 01:04 kyak_ has joined #qi-hardware 2013-10-27 01:04 kyak_ has quit [Changing host] 2013-10-27 01:04 kyak_ has joined #qi-hardware 2013-10-27 01:08 kyak has quit [*.net *.split] 2013-10-27 01:48 dos1 has quit [Ping timeout: 272 seconds] 2013-10-27 02:00 pcercuei has quit [Read error: Connection reset by peer] 2013-10-27 02:00 pcercuei has joined #qi-hardware 2013-10-27 02:18 pcercuei has quit [Read error: Connection reset by peer] 2013-10-27 02:19 pcercuei has joined #qi-hardware 2013-10-27 03:40 pcercuei has quit [Ping timeout: 240 seconds] 2013-10-27 03:59 qi-bot has quit [Ping timeout: 256 seconds] 2013-10-27 05:01 panda|x201 has joined #qi-hardware 2013-10-27 05:18 viric has quit [Ping timeout: 240 seconds] 2013-10-27 05:19 Jay7 has quit [Ping timeout: 240 seconds] 2013-10-27 05:19 Jay7 has joined #qi-hardware 2013-10-27 05:20 viric has joined #qi-hardware 2013-10-27 05:28 panda|x201 has quit [Ping timeout: 245 seconds] 2013-10-27 07:17 qi-bot has joined #qi-hardware 2013-10-27 07:42 porchaso0 has joined #qi-hardware 2013-10-27 07:43 porchao has quit [Read error: Connection reset by peer] 2013-10-27 08:37 lekernel has joined #qi-hardware 2013-10-27 08:43 wolfspraul has joined #qi-hardware 2013-10-27 08:46 kilae has joined #qi-hardware 2013-10-27 09:31 rz2k has joined #qi-hardware 2013-10-27 09:37 wolfspraul has quit [Ping timeout: 272 seconds] 2013-10-27 09:39 wolfspraul has joined #qi-hardware 2013-10-27 09:42 jekhor has joined #qi-hardware 2013-10-27 10:55 rz2k has quit [] 2013-10-27 11:30 dos1 has joined #qi-hardware 2013-10-27 11:31 porchao has joined #qi-hardware 2013-10-27 11:33 porchaso0 has quit [Ping timeout: 252 seconds] 2013-10-27 12:26 pcercuei has joined #qi-hardware 2013-10-27 12:52 nice list of Free alternatives to closed systems: https://www.prism-break.org/ 2013-10-27 13:01 what's up with the capitalization of Free lately? 2013-10-27 13:02 is it free beer, or Free beer? 2013-10-27 13:02 porchaso0 has joined #qi-hardware 2013-10-27 13:02 it just sounds like some religious thing to me 2013-10-27 13:02 :/ 2013-10-27 13:02 free beer is when no money changes hands 2013-10-27 13:03 Free beer is when you also get to know how it's made 2013-10-27 13:03 wasn't that "libre" ? 2013-10-27 13:04 some use "libre" but "Free" is more common 2013-10-27 13:04 porchao has quit [Read error: Connection reset by peer] 2013-10-27 13:05 not sure what the origin of the use of "libre" is. maybe it's just because it sounds nice. maybe it's because people got tired of capitalizing "Free", and explaining it. 2013-10-27 13:05 maybe it's because the US have been busy making "freedom" a bad word 2013-10-27 13:06 "Fʀᴇᴇ Software" ;) 2013-10-27 13:06 well, in french or spanish, 'libre' clearly means Free, and not free. 2013-10-27 13:06 viric: I think that was the point 2013-10-27 13:06 viric: yup, it's clearer 2013-10-27 13:07 rz2k has joined #qi-hardware 2013-10-27 13:07 that's a good "origin of the use of libre", isn't it? 2013-10-27 13:07 whitequark: "Free(R) Software" (R) Registered Trademark. All Rights Reserved. ;) 2013-10-27 13:08 viric: yes, that qualifies as another one. so we have four reasons in total :) 2013-10-27 13:08 (for now) 2013-10-27 13:08 wpwrak: no no, I was thinking along the lines of "Our Lᴏʀᴅ and Savior RMS" 2013-10-27 13:08 ah :) 2013-10-27 13:09 capitalization makes me think like that each time 2013-10-27 13:11 i try to avoid that sort of literature :) 2013-10-27 13:12 Who came up with capitalization at all? 2013-10-27 13:12 jekhor has quit [Ping timeout: 245 seconds] 2013-10-27 13:13 i guess it's a natural consequence of using "free" in a world where there's lots of "gratis" yet not so much freedom 2013-10-27 13:13 e.g., Fakebook is "free", right ? :) free admission to the panopticum 2013-10-27 13:14 did you see the new Intro software by linkedin? 2013-10-27 13:15 btw, http://www.heise.de/ct/schlagseite/2013/17/gross.jpg "... we ask this year's employee of the year to come on the stage ..." 2013-10-27 13:15 wpwrak: I've once got sent http://s-ak.buzzfed.com/static/enhanced/terminal01/2010/9/19/22/enhanced-buzz-23247-1284949703-3.jpg, after which I became curious and read it a bit 2013-10-27 13:15 wpwrak: http://engineering.linkedin.com/mobile/linkedin-intro-doing-impossible-ios 2013-10-27 13:15 but that's probably far too offtopic 2013-10-27 13:15 :) 2013-10-27 13:16 linkedin, worst spammer ever 2013-10-27 13:16 wpwrak: now you can make it your mitm mail box checker 2013-10-27 13:18 i guess linkedin is great help for people in the Human Resources department. linkedin = applicant_has_linkedin_account(); lemming = job_requires_dumb_obeisance(); if ((linkedin && lemming) || (!linkedin && !lemming)) hire(); else ignore(); 2013-10-27 13:19 being a linked user is a bad sign 2013-10-27 13:20 whitequark: ah, now i see where you got that long dong reference from :) 2013-10-27 13:20 whitequark: the Skeptic's annotated bible always helps 2013-10-27 13:20 wpwrak: nah, that's just what I use idle capacity of my brain for 2013-10-27 13:20 what in the world is better than lame puns! 2013-10-27 13:21 viric: honestly I'm horrified enough of the plain one 2013-10-27 13:21 the long dong verse was the better part of its surrounding context... 2013-10-27 13:21 sc 2013-10-27 13:21 sc ? 2013-10-27 13:21 whitequark: this is classified in categories, to bring you to the best passages 2013-10-27 13:21 sc, the mad spreadsheet software 2013-10-27 13:22 whitequark: lemme guess, no white knight to save that damsel in distress ? 2013-10-27 13:22 wpwrak: iirc she was stoned or something. with rather extensive details 2013-10-27 13:23 whitequark: what weird font you#re using? 2013-10-27 13:23 small uppercase, sometimes even blurred :-o 2013-10-27 13:23 whitequark: ah, good old violence porn 2013-10-27 13:23 DocScrutinizer05: Aʟʟ Hᴀɪʟ Uɴɪᴄᴏᴅᴇ 2013-10-27 13:24 http://www.panix.com/~eli/unicode/convert.cgi 2013-10-27 13:24 o.O 2013-10-27 13:24 it often works really strange on monospaced terminals 2013-10-27 13:24 whitequark: well that's like every second szene in the old testament 2013-10-27 13:24 scene 2013-10-27 13:24 * DocScrutinizer05 just wonders why the ᴅᴇ is all blurred 2013-10-27 13:24 DocScrutinizer05: I would guess the small-caps have different metrics than the monospaced letters on terminal and your rendering engine gets confused 2013-10-27 13:24 I agree, about the old testament 2013-10-27 13:25 or you don't use monospace for irc? 2013-10-27 13:25 errr 2013-10-27 13:25 I sometimes forget not everyone uses irssi :) 2013-10-27 13:26 DocScrutinizer05: also it may be related that my display has 170 dpi and I've hinting turned off. 2013-10-27 13:27 probably that, in fact 2013-10-27 13:27 viric: larsc: then the same people make their children read the bible, then censor the hell out of today's media 2013-10-27 13:27 "oh god there is a NIPPLE FINE THEM $100K IMMEDIATELY" 2013-10-27 13:27 (I think that's an american thing but nevertheless) 2013-10-27 13:28 whitequark: they go for the new testament I think 2013-10-27 13:28 http://wstaw.org/m/2013/10/27/plasma-desktopSL3743.png 2013-10-27 13:29 DocScrutinizer05: I feel sorry for your eyes 2013-10-27 13:29 cause we all know, we shouldn't take everything in the bibel literally, only the things we like 2013-10-27 13:29 also that totally looks like hinting gone wrong 2013-10-27 13:29 or just lack of it 2013-10-27 13:30 "C" and "O" are slightly blurred too 2013-10-27 13:30 but only for the D and the E 2013-10-27 13:30 ooh yeah 2013-10-27 13:30 ᴀʙᴄᴅᴇꜰɢʜɪᴊᴋʟᴍɴᴏᴩqʀꜱᴛᴜᴠᴡxyᴢ 2013-10-27 13:31 do you use vector fonts for terminals? 2013-10-27 13:31 http://wstaw.org/m/2013/10/27/plasma-desktopmC3743.png 2013-10-27 13:32 viric: on 170 dpi, totally 2013-10-27 13:32 http://wstaw.org/m/2013/10/27/plasma-desktopdO3743.png 2013-10-27 13:32 and it looks amazing 2013-10-27 13:33 for 170 dpi maybe it makes sense 2013-10-27 13:33 DocScrutinizer05: "font still under construction" ? :) 2013-10-27 13:33 the question is why any of you even touch displays with dpi lower than 170 or better 200-230 2013-10-27 13:33 lol, yeah 2013-10-27 13:33 (I don't think PCs with >170dpi exist) 2013-10-27 13:34 whitequark: no touch displays here 2013-10-27 13:34 lol 2013-10-27 13:34 seriously tho, those retina guys have a point 2013-10-27 13:35 meh, that's what I discussed with Raster like 6 years ago 2013-10-27 13:35 too bad linux is pretty horrible at supporting nonstandard dpis. 2013-10-27 13:35 well, they have a GHz race going on. they had to find something to do with all that performance ... 2013-10-27 13:36 wpwrak: meh, simple integrated intel video is all you need. I think it can do up to 4096x4096 2013-10-27 13:37 my 1920*1080 screen is absolutely perfect for me 2013-10-27 13:37 DocScrutinizer05: same here, except it's 13" 2013-10-27 13:37 diagonale is irrelevant, first approach 2013-10-27 13:37 also I suggest you to actually look at some hi-dpi displays :) 2013-10-27 13:38 it's the viewing angle 2013-10-27 13:38 viewing angle? IPS displays are 180° :) 2013-10-27 13:38 well, 178° 2013-10-27 13:38 * DocScrutinizer05 headdesks 2013-10-27 13:38 hm? 2013-10-27 13:38 I mean the segment of your view the screen takes 2013-10-27 13:39 a 3" display can give you same viewing ange as a 40", when you're close enough 2013-10-27 13:39 ah, that. yes, if you're sitting further from it, you would see it the same way even with lower dpi 2013-10-27 13:40 and I fon't need a viewing angle of >60° 2013-10-27 13:40 (or is it farther?) 2013-10-27 13:40 actually I think 45° are fine for me 2013-10-27 13:42 everything beyond that angle is best served by a second (and third) screen 2013-10-27 13:42 btw those 45° are horizontal 2013-10-27 13:42 vertical it's even way less 2013-10-27 13:42 hmm, some 120 deg here (three screens) 2013-10-27 13:42 and no need at all for extending it 2013-10-27 13:43 three screens are for quick "switching" to another screen. You can't watch all three same time 2013-10-27 13:43 wish the one in the middle was bigger (at same dpi, of course), and the side ones then vertical 2013-10-27 13:44 this for sure is convenient, but you definutely don't need the high resolution of main screen on the auxliary ones 2013-10-27 13:45 arhuaco has joined #qi-hardware 2013-10-27 13:45 they are just for a cursory glance. If anything fgoes on there, you rather need a conventient method to pull it to main screen quickly 2013-10-27 13:48 DocScrutinizer05: you can actually easily see all three at the same time: http://en.wikipedia.org/wiki/Human_eye#Field_of_view 2013-10-27 13:48 * DocScrutinizer05 wonders if the brain could train to use eye-blinking for command purposes. Like in: look at sth on aux xscreen, blink to move it to main screen 2013-10-27 13:49 that's field of reception, not field of clear sight 2013-10-27 13:50 for everythin more than 50° off the straight-ahead line horizontally, a mere blinking LED would suffice 2013-10-27 13:50 even color doesn't matter there 2013-10-27 13:50 and i work mainly on two of them. the third is for more static content. center screen for main activity (editing), left screen for auxiliary activity (main, chat), right screen for reference material or low-activity sessions 2013-10-27 13:51 DocScrutinizer05: I think not, you normally do not notice blinking as brain edits it out of the input stream 2013-10-27 13:51 perhaps you could use sequences which have a low probability of random occurence, ie double or triple blinking 2013-10-27 13:51 yep, that's why I wondered if you could train it so it comes naturally 2013-10-27 13:51 whitequark: your ancestors must have procreated at a very young age 2013-10-27 13:51 whitequark: for they couldn't see the predators approach from the side ... 2013-10-27 13:52 DocScrutinizer05: I believe it's reflextive, so probably no 2013-10-27 13:52 wpwrak: what? 2013-10-27 13:52 ETAB 2013-10-27 13:52 ah, you meant blinking of the eye, not of a light source. sorry, then it makes sense :) 2013-10-27 13:52 he meant me, I guess 2013-10-27 13:53 aah 2013-10-27 13:53 actually blinking of the light source is all you can notice at peripheral filed of sight 2013-10-27 13:54 even better than in center of sharp view 2013-10-27 13:54 but you can't tell about color or shape 2013-10-27 13:54 fun fact: you have sharp, color vision only in a really really small angle, slightly more than 1° 2013-10-27 13:54 exactly my words 2013-10-27 13:54 everything else is grayscale and with really low resolution BUT also really low latency 2013-10-27 13:55 thus for aux xscreens basically 10dpi would suffice ;-P 2013-10-27 13:55 also, both optical nerve and visual cortex do a lot of preprocessing 2013-10-27 13:55 movement and evenm simply shapes are done in retina 2013-10-27 13:56 simple* 2013-10-27 13:56 yeah 2013-10-27 13:56 you seem to actually have nerves that signal "moving right" from eye to brain 2013-10-27 13:57 and other elementary tokens 2013-10-27 13:57 like "corner" 2013-10-27 13:57 end of line 2013-10-27 13:58 also, unlike most other nerves in the body, which basically use a form of PWM--two levels, frequency-based signal transmission--color receptors actually use level-based transmission 2013-10-27 13:58 insects are happy with just that, they don't need more optical processing at all 2013-10-27 13:58 wow 2013-10-27 13:59 didn't know that 2013-10-27 13:59 I think it's recent research 2013-10-27 13:59 fun 2013-10-27 13:59 a smal "analog" section in the "digital" brain 2013-10-27 13:59 err CNS 2013-10-27 14:00 it's not very digital I'd say. the range of duty cycles is still continous 2013-10-27 14:00 same for inhibition/excitation points 2013-10-27 14:01 did you know that venus flie catcher plant needs exactly 2 signals on 2 of the sensor hairs to close their leaf and catch the fly. (or maybe it been 2 signals or 1 on three hairs) 2013-10-27 14:01 btw, did you know that walking is ultimately controlled by exactly two neurons in your spine? 2013-10-27 14:02 whitequark, did you know that in europe carrot is officially a fruit? 2013-10-27 14:02 each one, when excited, excites another one and inhibits itself 2013-10-27 14:02 yeah, heard sth like that. and it seems those two neurons get nervous now 2013-10-27 14:02 DocScrutinizer05: (venus fly catcher) does it even have nerves? 2013-10-27 14:02 or any electrical signal transmission at all 2013-10-27 14:02 that's the funny point, they don't 2013-10-27 14:02 cde: @RealCarrotFacts ? :D 2013-10-27 14:03 still they can count, and do logical OR 2013-10-27 14:03 even latch 2013-10-27 14:03 and timeout 2013-10-27 14:03 DocScrutinizer05: sounds not very complex, there's a *lot* of processes like that in the cell which are purely chemical 2013-10-27 14:03 for example gene expression 2013-10-27 14:03 sure 2013-10-27 14:04 you can have AND, OR, latches, delays there 2013-10-27 14:04 still amazing 2013-10-27 14:04 NOT too 2013-10-27 14:04 I think it'll be mere years until you can synthesize verilog into DNA :D 2013-10-27 14:04 yeah, that will be fun 2013-10-27 14:05 I#m already planning for my time as IT psychologist 2013-10-27 14:05 once neuronal networks become too complex to service them any other way 2013-10-27 14:05 "do you want to talk about it?" *the bacteria indifferently spins his flagellum" 2013-10-27 14:06 hehe 2013-10-27 14:06 whitequark: I'm not kidding. that's because portugal make carrot jam, and since jams can only be made from fruit, the lawmakers declared carrot to be a fruit 2013-10-27 14:07 cde: what the hell 2013-10-27 14:07 my thoughts exactly 2013-10-27 14:07 they should have triaged that as an upstream issue to whoever wrote the definition of "jam" 2013-10-27 14:07 the distinction beween fruit and vegetable is pretty fuzzy in daily speech 2013-10-27 14:07 and absolutely not in line with biology 2013-10-27 14:08 peppers and cucumber are usually called vegetables 2013-10-27 14:09 while indian figs are fruits (hmm maybe they really are) 2013-10-27 14:09 watermelon is a berry, strawberry is a multinutlet 2013-10-27 14:10 yeah 2013-10-27 14:10 (there's apparently no word equivalent to "многоорешек" in english, I translated it as "multinutlet") 2013-10-27 14:10 let's just call them pre-meat 2013-10-27 14:10 ROTFL 2013-10-27 14:10 I'm looking forward to the sequel of Cloudy with a Chance of Meatballs. lots of puns of fruit names 2013-10-27 14:10 also there's this weird thing in english where you call everything a fruit 2013-10-27 14:10 a nut is a fruit, wtf 2013-10-27 14:11 Rhubarb is no fruit either 2013-10-27 14:11 in russian, a fruit is something bright and juicy. apples, oranges, ... 2013-10-27 14:11 can you make a jam of the fruits of your labour ? 2013-10-27 14:12 they can make jam of me though 2013-10-27 14:12 wpwrak: "jam" is a borrowed word in russian 2013-10-27 14:12 so no proverbs with it 2013-10-27 14:12 pcercuei has quit [Ping timeout: 248 seconds] 2013-10-27 14:12 or good puns 2013-10-27 14:14 btw I'm working on a bright theme for irclogger 2013-10-27 14:14 well, EU is mad again. While jam in german is "Marmelade", it officially only applies to lemon and orange jam, everything else is "Konfituere" officially 2013-10-27 14:15 duh, somebody switching off lights outside 2013-10-27 14:16 throw a grenade there. that'll teach 'em 2013-10-27 14:16 I have to throw that pretty high 2013-10-27 14:17 looks and feels like thunderstorm approaching 2013-10-27 14:17 o/ 2013-10-27 14:21 wow I just found out my server's reverse is within poneytelecom.eu 2013-10-27 14:21 insanity rules 2013-10-27 14:29 http://stuff.whitequark.org/f/479eaf611a/ 2013-10-27 14:30 whatcha think about it? 2013-10-27 14:32 good contrast 2013-10-27 14:33 yeah, I thought I would hate a bright version--nope, it's pretty good 2013-10-27 14:33 pcercuei has joined #qi-hardware 2013-10-27 14:34 wpwrak: you should get a few rtl2832 dvb-t sticks 2013-10-27 14:36 particularly with a decent up-converter to give you that range fro virually 0 to 100MHz as well, they are extremely convenient lab tools 2013-10-27 14:37 receiver tuning from 0..2GHz, with just one switch-over at 100MHz. I couldn't even dream of this a few years back 2013-10-27 14:38 they're missing the most interesting band, 2.4 GHz 2013-10-27 14:38 only limitation is max bandwidth 2013-10-27 14:38 should also work with a down-converter 2013-10-27 14:38 ring-modulator and a 2GHz LXO 2013-10-27 14:39 but i have an usrp2. was pricy, of course, but works. not that i used it all that extensively, mainly as spectrum analyzer 2013-10-27 14:39 2GHz highpass at the input, lowpass at the output 2013-10-27 14:39 LOL 2013-10-27 14:39 preaching to the choir 2013-10-27 14:42 pk, so look at this: http://talk.maemo.org/showthread.php?t=91182 2013-10-27 14:44 s/pk/ok 2013-10-27 14:44 DocScrutinizer05 meant: "ok, so look at this: http://talk.maemo.org/showthread.php?t=91182" 2013-10-27 14:44 fU qi-bot, your syntax checker is broken 2013-10-27 14:45 yeah, nice if you're interested in those frequency bands 2013-10-27 14:53 FrankBlues has joined #qi-hardware 2013-10-27 14:53 aaaand deployed: http://irclog.whitequark.org/ 2013-10-27 14:57 jekhor has joined #qi-hardware 2013-10-27 14:59 we had no activity at all on the 21st ? interesting 2013-10-27 15:12 rtl_fm -N -l 6 -E -f 173240000 -s 22050 -o 4 -g 100 – | tee >(./multimon -a POCSAG1200 -f alpha -t raw /dev/stdin) >(play -t raw -r 22050 -e signed-integer -b 16 -c1 -V1 -q -) > /dev/null 2013-10-27 15:12 can somebody explain to me the " >(... " syntax? I never seen that 2013-10-27 15:15 DocScrutinizer05: it launches a subprocess with a pipe as input, and is substituted for that pipe 2013-10-27 15:15 try echo >(cat) 2013-10-27 15:15 and echo foo > >(cat) 2013-10-27 15:15 lol, that's exaxtly what I just did 2013-10-27 15:16 hivemind! 2013-10-27 15:16 I still don't get it 2013-10-27 15:16 jr@saturn:~> echo >(cat) 2013-10-27 15:16 /dev/fd/63 2013-10-27 15:17 jr@saturn:~> date >(cat) 2013-10-27 15:17 date: ungültiges Datum „/dev/fd/63“ 2013-10-27 15:17 the more common variant is <(command) 2013-10-27 15:18 yeah, that is kinda semi-sensible 2013-10-27 15:18 e.g., cat <(do-this) <(do-that) 2013-10-27 15:18 parent shell does r,w=pipe(). suppose r=62,w=63. it launches cat with stdin replaced with 62, then substitutes the >... construct with /dev/fd/63 2013-10-27 15:19 you need to add another > before it to redirect 2013-10-27 15:19 so date > >(cat) 2013-10-27 15:19 whoever came up with that syntax is a horrible person 2013-10-27 15:19 * DocScrutinizer05 tries in vain to wrap his head around it 2013-10-27 15:20 just use | instead 2013-10-27 15:20 * whitequark . o O ( whoever came up with unix... oh, we know who. ) 2013-10-27 15:20 larsc: won't work 2013-10-27 15:20 here tee is used as a stream duplicator 2013-10-27 15:21 ah I see 2013-10-27 15:21 there's obviously a huge difference between date|cat and date >(cat) 2013-10-27 15:21 >() is evaluated as a file? 2013-10-27 15:21 DocScrutinizer05: date|cat ~ date > >(cat) 2013-10-27 15:21 date >(cat) cannot be reproduced with | syntax 2013-10-27 15:21 pcercuei: kind of 2013-10-27 15:22 * DocScrutinizer05 gives up and ponders having some real food 2013-10-27 15:22 you could do the same using pipes 2013-10-27 15:23 larsc: could you? 2013-10-27 15:23 s/pipes/named pipes 2013-10-27 15:23 larsc meant: "you could do the same using named pipes" 2013-10-27 15:23 oh. yeah. it's even more contrived to use though 2013-10-27 15:24 I'd like to see it using named pipes. Might help me to grok it 2013-10-27 15:24 mkfifo pipe1; mkfifo pipe2; rtl_fm ... | tee pipe1 pipe2& ./multimon ... pipe1& play ... pipe2 2013-10-27 15:25 could you use the simpler date >(cat) ? 2013-10-27 15:25 no 2013-10-27 15:25 that doesn;t make any sense 2013-10-27 15:25 ohwell 2013-10-27 15:25 cya 2013-10-27 15:26 arhuaco has quit [Ping timeout: 240 seconds] 2013-10-27 15:26 date > >(cat) on the other hand would be. mkfifo pipe; data > pipe& cat pipe 2013-10-27 15:27 FrankBlues has quit [Remote host closed the connection] 2013-10-27 15:32 ok, now I think I got it 2013-10-27 15:32 pcercuei: btw. did you already had your phone call? 2013-10-27 15:33 >(...) returns a FD that's used as stdin in the command inside () 2013-10-27 15:33 exactly 2013-10-27 15:33 but the fd looks like a regular file to the other program 2013-10-27 15:33 yep 2013-10-27 15:35 wget -O >(less) 2013-10-27 15:36 wget -O >(less -) even 2013-10-27 15:36 with which shell does this work? 2013-10-27 15:36 bash 2013-10-27 15:36 and zsh 2013-10-27 15:36 weird, I cannot get it to work 2013-10-27 15:37 I tried uname -a >(cat) 2013-10-27 15:37 uname -a > >(cat) 2013-10-27 15:37 ah, right 2013-10-27 15:37 works:) 2013-10-27 15:39 but it is a bash-ism I think 2013-10-27 15:39 so be careful when you use it in scripts 2013-10-27 15:39 ok =) 2013-10-27 15:39 I gather <() returns a FD pointing to stdout then 2013-10-27 15:40 DocScrutinizer05: but when it is useful as compared to a regular pipe? 2013-10-27 15:40 when your outer command needs a filename as argument 2013-10-27 15:40 as in wget -O >(less -) 2013-10-27 15:40 or even strace 2013-10-27 15:42 or tee ;-) 2013-10-27 15:42 oh, clever 2013-10-27 15:47 porchao has joined #qi-hardware 2013-10-27 15:49 porchaso0 has quit [Ping timeout: 245 seconds] 2013-10-27 16:09 wget -O - http://example.com/dvd.iso | tee >(sha1sum > dvd.sha1) > dvd.iso 2013-10-27 16:09 That makes `tee' write not just to the expected output file, but also to a pipe running `sha1sum' and saving the final checksum in a file named `dvd.sha1'. 2013-10-27 16:09 Note, however, that this example relies on a feature of modern shells called "process substitution" (the `>(command)' syntax, above; Process Substitution (bashref)Process Substitution.), so it works with `zsh', `bash', and `ksh', but not with `/bin/sh'. 2013-10-27 16:09 from tee info page 2013-10-27 16:11 http://anticache.img9.joyreactor.com/pics/post/comics-optipess-jurassic-park-unix-638659.png 2013-10-27 16:14 * DocScrutinizer05 tries to recall where his humor got lost 2013-10-27 16:16 a joke with bash in it would have been better, I admit 2013-10-27 16:17 You can extend this example to make `tee' write to two processes, computing MD5 and SHA1 checksums in parallel. In this case, process substitution is required: 2013-10-27 16:17 wget -O - http://example.com/dvd.iso | tee >(sha1sum > dvd.sha1) >(md5sum > dvd.md5) > dvd.iso 2013-10-27 16:19 arhuaco has joined #qi-hardware 2013-10-27 16:21 that's severly awesome 2013-10-27 16:29 now, how do you check for wget failing ? :) 2013-10-27 16:30 it's hard to try not to remember such tricks 2013-10-27 16:30 because you cannot use it for production or important scripts 2013-10-27 16:30 since it's not portable :( 2013-10-27 16:31 bash is pretty much everywhere. sometimes you don't really have a choice. 2013-10-27 16:31 just make sure your script begins with #!/bin/bash, and your makefile with SHELL = /bin/bash 2013-10-27 16:34 99% of the time indeed that works 2013-10-27 16:35 having scripts that expect bash to be installed are really annoying, in my opinion 2013-10-27 16:36 especially if they do #!/bin/sh 2013-10-27 16:36 hehe sure this is a serious mistake 2013-10-27 16:36 using bashism and putting /bin/sh in shebang 2013-10-27 16:39 you need bash for things that use pipes and need to check the success of the first command, i.e., with $PIPESTATUS 2013-10-27 16:39 not if you use named pipes 2013-10-27 16:39 or subshells 2013-10-27 16:40 yeah, but that's messy 2013-10-27 16:40 coproc is also nice 2013-10-27 16:42 best is to not use shell scripts at all, I guess ;) 2013-10-27 16:43 and I wouldn't want to miss array envs 2013-10-27 16:43 yes. better write the stuff in erlang .net ;-) 2013-10-27 16:43 *cough+ 2013-10-27 16:49 kilae_ has joined #qi-hardware 2013-10-27 16:51 kilae has quit [Ping timeout: 248 seconds] 2013-10-27 16:53 wpwrak: Errorlevels? (wget failing) 2013-10-27 16:54 the return code, yes 2013-10-27 16:57 If you are already using >(...) you could slightly rewrite things though: wget -O >(tee >(sha1sum > dvd.sha1) >(md5sum > dvd.md5) > dvd.iso) http://example.com/dvd.iso 2013-10-27 16:59 you could still get errors from the subshells ... 2013-10-27 16:59 e.g., after mkdir dvd.sha1 2013-10-27 16:59 (to not use the standard example of the disk getting full) 2013-10-27 17:01 ( wget -O - http://example.com/dvd.iso; echo $? >&2) | tee >(sha1sum > dvd.sha1) >(md5sum > dvd.md5) > dvd.iso 2013-10-27 17:01 ? 2013-10-27 17:01 ;-)) 2013-10-27 17:01 you want to use $PIPESTATUS 2013-10-27 17:02 but i guess all the subshells will be a problem you can't even solve with that 2013-10-27 17:06 well, you got &3 to &9 iirc 2013-10-27 17:08 but yeah, no globals suck 2013-10-27 17:08 I never really got it why unix shells don't support global envs 2013-10-27 17:09 at least common var space for a parent and its childs 2013-10-27 19:38 lekernel has quit [Ping timeout: 245 seconds] 2013-10-27 19:51 lekernel has joined #qi-hardware 2013-10-27 20:20 * arhuaco is scared of bash 2013-10-27 20:38 rz2k has quit [] 2013-10-27 20:44 DocScrutinizer05: you can use any fd for &. echo foo >&63 2013-10-27 21:08 kilae_ has quit [Quit: ChatZilla 0.9.90.1 [Firefox 24.0/20130910160258]] 2013-10-27 21:39 hmm, somewhere I've read about a limitation, build-dependent 2013-10-27 21:40 maybe not for >& 2013-10-27 21:49 wolfspraul has quit [Quit: leaving] 2013-10-27 22:15 jekhor has quit [Ping timeout: 272 seconds] 2013-10-27 22:31 lekernel has quit [Quit: Leaving]