←2015-03-27 2015-03-28 2015-03-29→ ↑2015 ↑all
00:00:26 <int-e> "MD5 takes an arbitrary string of characters, passes it through a deterministic blender and produces an unnecessarily long 128-bit value."
00:11:29 <FreeFull> Unnecessarily long?
00:12:42 <FireFly> Yes. MD5/4 compresses it into 32 bits
00:13:18 <int-e> FreeFull: it might help to know that the proceedings are dated April 1st 2015.
00:13:31 -!- Phantom__Hoover has quit (Ping timeout: 250 seconds).
00:13:46 <FreeFull> int-e: Hmm, so they're from the future
00:14:00 <int-e> But not for long :)
00:14:15 -!- Phantom__Hoover has joined.
00:15:04 <fizzie> "Or, more likely, you're looking at them on a screen because you're too cheap to actually buy the proceedings." They know me so well.
00:15:13 <fizzie> I wonder if that's on the paper copy at all.
00:15:18 <FreeFull> int-e: I wasn't looking at the proceedings, and that sentence is mostly accurate
00:15:23 -!- clog has joined.
00:15:50 <FreeFull> fizzie: Oh jeez, the last page is filled with lots of tiny text
00:16:07 <int-e> FreeFull: At the same time, no self-respecting cryptographer would write that in a scientific publication.
00:17:43 <FreeFull> "the solid foundation of the MD5/4 cryptographic hash function"
00:18:20 <FreeFull> "the RSA public key cryptosystem signing procedure [10] with a key length of 32 bits"
00:18:25 <orin> wait, so 96 of the bits are not needed?
00:18:35 <FreeFull> Heh
00:19:02 <int-e> FreeFull: wait until you get to the footnote explaining why 2048 is a more popular choice
00:19:27 <FreeFull> I'll read this in-depth later
00:20:04 <FreeFull> "Copyright © 2015 The Association of Computational Heresy, your leading source for cutting-edge research on Science, Computational Archaeolinguistics, and Artificial Stupidity. "
00:20:54 <int-e> owie. "Kaliningrad, Russia (colloq.); better known in the Formal Literature as Königsberg, Prussia...
00:20:57 <int-e> "
00:21:26 <orin> Oh, they aren't saying that 96 of the bits can be deduced from 32, they're just saying THEY only need 32 bits
00:22:39 <orin> wait is this entire thing an elaborate joke or not?
00:23:24 <int-e> it's a valid cryptographic operation, assuming your attacker doesn't have access to modern (post 1960?) computers.
00:23:42 <int-e> orin: you're catching on fast
00:23:57 -!- callforjudgement has joined.
00:24:08 -!- ais523 has quit (Disconnected by services).
00:24:11 -!- callforjudgement has changed nick to ais523.
00:24:29 <orin> Unfortunaely in the internet age it is hard to tell whether anything is a joke
00:26:20 <fizzie> If the website is to be believed, it's "joke realizations of joke ideaas", "joke realizations of serious ideas", and "serious realizations of joke ideas", all of which /joke/.
00:27:59 <ais523> that's a similar set as this channel, isn't it?
00:28:03 <ais523> at least when it's ontopic?
00:30:47 <fizzie> Yes, although this channel also has the "related to programming languages" bit set.
00:31:06 <fizzie> Allegedly, anyway.
00:31:20 <zzo38> Yes, I suppose is mostly the topic anyways
00:32:38 -!- copumpkin has joined.
00:37:27 <zzo38> `danddreclist 64
00:37:28 <HackEgo> danddreclist 64: shachaf nooodl boily \ http://zzo38computer.org/dnd/recording/level20.tex
00:46:35 -!- contrapumpkin has joined.
00:47:03 -!- Phantom__Hoover has quit (Ping timeout: 244 seconds).
00:48:19 <orin> ancient greeks, including Aristotle and a mortal man [1] named Socrates.
00:48:43 <zzo38> What is that about the Greeks?
00:49:40 <orin> zzo38: it's a sentence from the artivle on "artisanal type theory"
00:49:55 -!- copumpkin has quit (Ping timeout: 250 seconds).
01:02:05 <zzo38> My RDF namespace now has 19 leaf nodes defined.
01:08:30 <FireFly> "By he way, the images are always 256x256, because numbers that are a power of two are faster.²" "²This is true on computers, because computers count in binary. In the human eye, powers of ten are faster, because humans have ten fingers."
01:08:34 <FireFly> the*
01:09:35 <ais523> that is one of those statements that's /just/ accurate enough that you can't outright call it false, while nonetheless being completely misleading
01:12:10 <zzo38> Eyes aren't fingers though, isn't it?
01:12:16 <ais523> indeed
01:13:20 <HackEgo> [wiki] [[Batch file]] N http://esolangs.org/w/index.php?oldid=42237 * Esowiki201529A * (+76) Created page with "== examples == === Hello, World === <pre> @echo off echo Hello, World </pre>"
01:14:13 <HackEgo> [wiki] [[Batch file]] http://esolangs.org/w/index.php?diff=42238&oldid=42237 * Esowiki201529A * (+24)
01:14:36 -!- augur_ has joined.
01:14:38 <orin> god damn it these earbuds have too much bass
01:14:55 -!- augur has quit (Ping timeout: 265 seconds).
01:15:42 <orin> I _knew_ I should have bought cheaper ones
01:20:04 <HackEgo> [wiki] [[Batch file]] http://esolangs.org/w/index.php?diff=42239&oldid=42238 * Esowiki201529A * (+6)
01:20:54 <zzo38> Why is PDJSON larger than my RDF Turtle parser when both are compiled with -s -O2 options?
01:23:14 -!- contrapumpkin has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…).
01:24:59 <zzo38> Isn't RDF Turtle more complicated than JSON?
01:26:08 <orin> How does the error handling compare in each piece of code?
01:27:54 <zzo38> I don't know; maybe you should look. Although, both do check for several errors (there are probably more errors to check for in RDF).
01:28:14 <zzo38> Maybe you can look though, and see if that help
01:31:18 -!- Phantom_Hoover has quit (Remote host closed the connection).
01:36:17 -!- Phantom__Hoover has joined.
01:37:20 <orin> Hmm... maybe it's strerror?
01:38:30 <orin> which would implicitly include a bunch of strings, most of which are never used,
01:39:35 <zzo38> O, maybe
01:39:58 <int-e> ^rerere loop unrolling
01:40:37 <int-e> fungot: aww
01:40:37 <fungot> int-e: fnord by pikhq already exists and doesn't require tcness from that. therefore, a language with card decks. yarr.
01:46:35 -!- copumpkin has joined.
01:46:50 <zzo38> No, strerror is linked externally; the compiled DLL doesn't include all of those strings
01:46:56 <zzo38> I checked
01:47:29 <zzo38> Despite that, sqlext_pdjson.dll is 46K and sqlext_turtle.dll is 22K
01:54:19 <zzo38> PDJSON does have smaller source-codes though.
02:02:47 -!- callforjudgement has joined.
02:05:08 -!- ais523 has quit (Ping timeout: 248 seconds).
02:06:09 -!- ^v has quit (Read error: Connection reset by peer).
02:06:27 -!- ^v has joined.
02:09:51 -!- copumpkin has quit (Ping timeout: 246 seconds).
02:12:58 -!- copumpkin has joined.
02:14:32 -!- callforjudgement has changed nick to ais523.
02:27:35 -!- GeekDude has quit (Quit: {{{}}{{{}}{{}}}{{}}} (www.adiirc.com)).
02:33:15 -!- copumpkin has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…).
02:39:26 -!- Phantom__Hoover has quit (Read error: Connection reset by peer).
03:34:40 -!- vodkode has joined.
03:38:55 -!- adu has joined.
03:47:16 -!- adu has quit (Quit: adu).
03:53:45 -!- Patashu has quit (Ping timeout: 246 seconds).
03:58:10 -!- vodkode has quit (Ping timeout: 256 seconds).
04:16:10 -!- Patashu has joined.
04:59:40 -!- shachaf has joined.
05:29:30 <Sgeo_> https://www.reddit.com/r/programming/comments/30hk3l/github_may_be_inaccessible_today_due_to_a_ddos/
05:29:32 <Sgeo_> Gotta love JSONP
05:40:24 -!- oerjan has joined.
05:42:43 <Sgeo_> Given that other people are pointing out the obvious right way to fix the DDoS attack, is it that bad if I mention it too?
05:42:55 <Sgeo_> The first few things I said I avoided mentioning it, but now I haven't
05:43:11 <Sgeo_> But it's an obvious thing... but apparently not obvious to the Chinese gov
05:45:55 <oerjan> someone's ddos'ing china?
05:46:43 <Sgeo_> China is attempting to DDoS GitHub
05:47:00 <Sgeo_> By causing sites using Baidu analytics to make JSONP requests to GitHub
05:47:15 <Sgeo_> *to cause users to make etc
05:47:25 <Sgeo_> GitHub responded by making that URL return an alert()
06:08:31 -!- ais523 has quit (Read error: Connection reset by peer).
06:08:37 -!- callforjudgement has joined.
06:08:48 -!- callforjudgement has changed nick to ais523.
06:34:30 -!- ais523 has quit (Remote host closed the connection).
06:34:45 -!- ais523 has joined.
06:37:10 -!- ais523 has quit (Read error: Connection reset by peer).
06:37:14 -!- callforjudgement has joined.
06:37:21 -!- callforjudgement has changed nick to ais523.
06:44:29 <oerjan> <orin> god damn it these earbuds have too much bass <-- bass are inferior to babel fishes, anyhow
06:46:26 <oerjan> btw i suggest featuring Esme for april 1, on the reasoning that it will likely greatly speed up the next feature.
06:50:16 -!- ais523 has quit (Read error: Connection reset by peer).
06:50:23 -!- callforjudgement has joined.
06:50:48 <zzo38> If the new version of CLC-INTERCAL is released on April 1, then should you feature the new version of CLC-INTERCAL on April 1?
06:51:59 <Sgeo_> But Esme is better as a 'this should not be featured' language being featured
06:52:42 <callforjudgement> I wouldn't recommend featuring Esme
06:52:45 -!- callforjudgement has changed nick to ais523.
06:53:21 <Sgeo_> If VeeBeeWiki needs Esme to run, maybe there would be hints of what the language is actually supposed to be
06:54:46 <Sgeo_> Except only link apparently was dead
07:01:13 <zzo38> I looked on a TV guide and on the guide in the cable box with DVR to figure out when the show is on, and then I went into the basement to record it on the VCR; are you going to call it strange if I do such thing?
07:02:50 -!- zadock has quit (Read error: Connection reset by peer).
07:02:54 <Sgeo_> Slightly strange that you'd use VCR if a DVR is present, also wondering why you needed to check two sources
07:03:09 <orin> videotapes aren't used anymore in Canada, but they are used a lot in Japan, so it' relative
07:05:26 <orin> Anyway DVR can have limitations as to getting the data off the DVR
07:06:26 <zzo38> As it turns out they connected it to an input on the DVD recorder on the main floor, so that is one way to copy it I suppose
07:06:52 <zzo38> Although someone else is using the DVR and that TV set so I don't want to fill up the disk space or to interfere with their TV shows
07:07:39 <zzo38> The reason I checked two sources is that the TV guide doesn't list shows on at 3AM
07:08:00 <orin> it's 3AM now
07:08:41 <zzo38> Not in my timezone
07:18:40 <elliott> oerjan: agreed re: esme
07:18:46 <elliott> not even ironically
07:18:49 <elliott> I mean
07:18:50 <elliott> yes ironically
07:18:52 <elliott> but not on that level
07:32:54 -!- zzo38 has quit (Ping timeout: 265 seconds).
07:42:14 <oerjan> OKAY
07:42:59 <oerjan> the haskell wiki is so broken i cannot even view diffs :(
07:43:32 <oerjan> the fact no one else has seemingly noticed doesn't bode well.
07:43:44 <oerjan> because it means no one is checking edits.
07:44:10 <shachaf> #haskell-infrastructure hth
07:45:02 <oerjan> ok
07:50:59 <shachaf> i,i hats that help
07:56:29 <oerjan> i'm sure it did, just look at the response
07:57:40 <shachaf> it's p. late in haskellland
08:01:48 <oerjan> pesky western hemisphere
08:39:37 -!- hjulle has joined.
08:54:43 -!- hjulle has quit (Ping timeout: 264 seconds).
09:21:30 -!- zadock has joined.
09:38:39 <orin> struct mytest A = {3,5.6,"World",0};
09:38:42 <orin> struct mytest B = {5,3.1415,"Hello",&A};
09:38:45 <orin> prin7f("{%d,%f,%s,%R}",&B);
09:39:11 <orin> output is: {5,3.141500,Hello,{3,5.600000,World,}}
09:39:17 <orin> muhuhahahaha
09:40:28 <oerjan> charming
09:41:53 <orin> the prin7f family of functions works sort of like the printf family, except they take their data out of a given buffer instead of from dynamically typed arguments
09:42:37 <orin> The %R specifier recurses into a pointer with the same format
09:44:14 <oerjan> must it be the whole format? otherwise it seems a bit limited.
09:44:57 <oerjan> clearly we need the full power of BNF here
09:45:09 <orin> Hmmm....
09:46:29 <oerjan> solution: make the first argument a struct also >:)
09:46:59 <oerjan> ...might be a bit tricky.
09:47:22 <orin> Well, I'm just covering onw of the most common cases. Full BNF would use an array of formats that can refer to each other with "%{N}"
09:48:58 <orin> or something
09:50:21 <orin> Of course, I'd have to write code to detect looped pointers.... ugh
09:51:08 <orin> wait, I already need that
09:55:26 <shachaf> `olist 979
09:55:27 <HackEgo> olist 979: shachaf oerjan Sgeo FireFly boily nortti
09:55:48 <Sgeo_> ooh, list
09:56:23 <Sgeo_> huh my FB page hasn't said anything yet
09:57:59 <oerjan> ol' ist
10:01:04 <shachaf> Sgeo_ can just take your trademark like that?
10:01:15 <shachaf> next it'll have to be ølist
10:02:28 <oerjan> øl, øl, og mere øl
10:07:17 <Jafet> https://wiki.postgresql.org/wiki/Cyclic_Tag_System
10:08:39 <ais523> Jafet: oh wow
10:09:34 -!- shikhin has joined.
10:11:03 <Jafet> (The guilty party seems to be http://assets.en.oreilly.com/1/event/27/High%20Performance%20SQL%20with%20PostgreSQL%20Presentation.pdf)
10:14:48 <orin> https://gist.github.com/orenwatson/985cdf4a945ee80f756d
10:19:20 -!- MoALTz has joined.
10:20:44 -!- zadock has quit (Quit: Leaving).
10:38:53 -!- vodkode has joined.
10:53:00 -!- Patashu_ has joined.
10:53:00 -!- Patashu has quit (Disconnected by services).
10:58:23 -!- SopaXorzTaker has joined.
10:59:36 -!- MoALTz has quit (Quit: Leaving).
11:02:49 -!- MoALTz has joined.
11:09:23 -!- vodkode has quit (Ping timeout: 255 seconds).
11:15:50 -!- Phantom_Hoover has joined.
11:19:07 -!- oerjan has quit (Quit: More nagging later).
11:28:17 -!- shikhin has quit (Ping timeout: 265 seconds).
11:31:51 -!- shikhin has joined.
11:37:45 -!- shikhin has quit (Ping timeout: 256 seconds).
11:37:55 -!- boily has joined.
12:34:32 -!- shikhin has joined.
12:35:58 -!- Patashu_ has quit (Ping timeout: 246 seconds).
12:51:01 -!- shikhin has quit (Quit: leaving).
12:53:41 -!- nys has joined.
13:25:14 <boily> @metar CYUL
13:25:15 <lambdabot> CYUL 281300Z 35011KT 30SM FEW090 FEW240 M09/M17 A3006 RMK AC1CI1 AC TR CI TR SLP181
13:25:50 <boily> minus fungotting nine. and it was snowing yesterday...
13:25:50 <fungot> boily: and diamondie; if they'd still age at this speed and stayed in bed the last 100 years, i'd say.
13:26:30 <boily> fungot: I wouldn't mind cats or dogs or diamondies pouring down. as long as it isn't any more snow.
13:26:31 <fungot> boily: also pushing should be fnord?? weirdo! set-car! set-cdr! are functions, but f takes as an argument
13:27:04 <boily> fungot: it's not weird, it's common sense. you push and shovel and hurl the snow away.
13:27:04 <fungot> boily: tell me what " fnord" is longer than
13:27:15 <boily> fungot: fnord is longer than snow.
13:27:15 <fungot> boily: anyone here familiar with dr scheme? fnord!
13:27:27 <boily> fungot: perhaps. I think racket's more popular nowadays.
13:33:05 -!- ^v has quit (Read error: Connection reset by peer).
13:33:31 -!- ^v has joined.
13:33:41 <fizzie> @metar uh what was that code again
13:33:45 <fizzie> @metar EGLL
13:33:46 <lambdabot> EGLL 281320Z 23017KT 9999 OVC013 13/09 Q1010 TEMPO SCT012 BKN016
13:41:58 -!- jameseb has quit (Quit: ZNC - http://znc.in).
13:41:58 -!- Taneb has quit (Quit: ZNC - http://znc.in).
13:57:28 <int-e> fizzie: do you know why fungot forgot about ^rerere and ^rreree?
13:57:28 <fungot> int-e: i just want to dump the fnord leaves.
13:57:42 <int-e> fungot: brilliant excuse
13:57:42 <fungot> int-e: in your file makes it instantly clear. chopping it into at least two other links that don't work
13:59:12 <int-e> or perhaps a better question is, does fungot support remembering defined commands over restarts?
13:59:12 <fungot> int-e: that's a zip file stored in a hash-table? what's wrong with it just pointing out that a lot
14:00:40 -!- copumpkin has joined.
14:01:33 <FireFly> a zip file in a hash table seems a bit cumbersome for befunge processing
14:02:29 <boily> you're underestimating fungot's prowesses.
14:02:30 <fungot> boily: i'm afraid i haven't read it yet. also, saying the f word more than once ( per invocation).
14:02:54 <boily> fungot: I wouldn't know. I'm seriously lacking in the f-word department.
14:02:55 <fungot> boily: which is good. ( well, it seems...
14:03:01 <boily> fungot: it is.
14:03:02 <fungot> boily: and merging is great!" and was curious if scheme code were parsed textually, you'd never know what you'll find over the rainbow"? x_x
14:05:11 <FireFly> graphical parsing of scheme sounds very cumbersome, fungot
14:05:11 <fungot> FireFly: it's unnecessary, especially if one restricts it to working stuff and marks them appropriately)
14:05:20 <FireFly> fungot: why indeed
14:05:21 <fungot> FireFly: in c. " fnord"? when i load it?' and how will that help me tracing the prog? hardly
14:06:52 <fizzie> int-e: Nobody ^saved?
14:15:00 <boily> there's a ^save?
14:15:28 <int-e> fizzie: quite possibly.
14:17:00 <int-e> ^def rreree bf ,..>,[.<.>.>,]<.
14:17:00 <fungot> Defined.
14:17:11 <int-e> ^def rerere bf ,.>,.<.>>,[.<.<.>>>,]<.<.>.
14:17:11 <fungot> Defined.
14:17:13 <int-e> ^save
14:17:31 <int-e> (am I even allowed to do that?)
14:19:23 <boily> ŝave
14:19:26 <boily> hm.
14:19:29 <boily> ^save
14:19:38 <int-e> revive keys!
14:19:58 <boily> my keys are dead and I like it that way.
14:20:28 * boily is a certified necrokeyboardomancer :D
14:26:16 <ais523> int-e: I don't think it works for anyone but fizzie to ^save?
14:31:02 -!- zadock has joined.
14:53:05 -!- Koen_ has joined.
15:02:26 -!- Tritonio has joined.
15:04:49 -!- copumpkin has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…).
15:09:25 -!- Phantom_Hoover has quit (Ping timeout: 252 seconds).
15:24:27 -!- Tritonio has quit (Ping timeout: 252 seconds).
16:12:31 -!- vodkode has joined.
16:15:00 -!- ais523 has quit.
16:17:56 -!- copumpkin has joined.
16:19:59 -!- zzo38 has joined.
16:29:08 <fizzie> ^save
16:29:08 <fungot> OK.
16:29:20 <fizzie> Yes, it's one of these restricted ones.
16:35:28 <boily> ^unrestrict
16:35:52 <boily> fungot: please, unrestrict yourself. it's for the Greater Good®
16:35:52 <fungot> boily: i'm getting rather tired of incessant divergence from the topic has the url :) :) 3 :)
16:36:02 <boily> fungot: bleh to you too!
16:36:02 <fungot> boily: it's put onto amb-fail.
16:44:04 <olsner> ooh, fungot
16:44:04 <fungot> olsner: it was some other programming language, it's just too difficult in a statically typed macro system from a large variety of broken keymaps
16:59:09 -!- vodkode has quit (Read error: Connection reset by peer).
17:11:56 <FireFly> fungot: I could see that, yes
17:11:56 <fungot> FireFly: well at least it will be a variable name is really call-with-current-continuation. amb is somewhat more roundabout than i thought
17:19:32 -!- SopaXorzTaker has quit (Remote host closed the connection).
17:24:32 <newsham> https://github.com/rntz/rotten#rotten
17:24:40 <newsham> lisp compiler that compiles in a backdoor
17:24:52 <newsham> when self compiling
17:31:57 <int-e> newsham: you know "reflections on trusting trust", right? http://cm.bell-labs.com/who/ken/trust.html
17:34:03 <int-e> Oh, it's the first thing mentioned in the readme.md, I should've clicked the link before answering
17:38:35 <newsham> inte: nevah hoid of it!
17:38:42 <newsham> inte: you know about diverse double compilation, right?
17:38:52 <int-e> yes
17:45:52 -!- GeekDude has joined.
18:03:32 <Jafet> Indeed, lisp is a poor attack vector for this because we know that every lisp hacker can make their own implementation of lisp
18:04:26 <Jafet> (Large defense surface?)
18:06:03 <b_jonas> fungot, is lisp a poor attack vector?
18:06:03 <fungot> b_jonas: seriously i mean da scm monster. water walker.
18:06:25 <b_jonas> fungot, have you tried to make your own dialect of lisp yet?
18:06:25 <fungot> b_jonas: forget your religion is i am an image artist" fnord, i'm sorry, but i can't make
18:06:36 <b_jonas> from that parenthesis stuff you mentioned a few days ago I thought you would
18:07:17 -!- Nokaji has joined.
18:13:10 <izabera> do you guys know suckless?
18:15:52 -!- lambdabot has quit (Quit: brb).
18:17:21 <myname> of course
18:18:00 <izabera> good, because i need to rant
18:18:04 <izabera> THEIR NAME IS WRONG
18:18:08 <izabera> THEIR STUFF SUCKS
18:18:26 <izabera> WTF
18:18:29 <izabera> ok /rant
18:20:46 -!- lambdabot has joined.
18:35:53 <FireFly> fungot: does suckless suck less?
18:35:54 <fungot> FireFly: ' these people's c is horrible' for ' binary', though. i rather enjoy sounding the opinions of other schemers, but if they used python or scheme or cl
18:36:11 <FireFly> I think fungot agrees with you, izabera
18:36:11 <fungot> FireFly: i suppose doom iii on geforce 4 under the hurd might be an example of upwards only continuations.
18:36:22 <FireFly> uh. I don't think so
18:37:12 <int-e> izabera: maybe (turning an old joke about MS around) they tried to write firmware for a vacuum cleaner...
18:51:48 <elliott> orin: you'll need to handle padding
18:55:22 <zzo38> What is that? How does their sed breaks UTF-8?
18:56:08 -!- Taneb has joined.
18:58:54 -!- shikhin has joined.
19:00:41 <elliott> izabera: the suckless people rarely seem to bother being consistent in their philosophy...
19:01:11 <elliott> the Unicode support thing seems like kind of an outlier to me. I have a feeling that if Plan 9 didn't have such a focus on it, they'd consider it unnecessary bloat :p
19:08:36 <zzo38> If you want to add Unicode support into sed and stuff that using regular expression, is easy to do, is make up a regex modifier flag to enable UTF-8 in which case when it read the byte 0xC0-0xFF then read however many next bytes 0x80-0xBF are considered as part of the same character. Nothing else needs to be done, and now you can even turn it off by just not specifying this flag.
19:08:59 <zzo38> You shouldn't try to add in everything else like character classes and so on
19:10:52 <int-e> the meaning of [a-ä] is very different with and without the flag; you cannot encode the resulting unicode range conveniently in a 256 bit array.
19:13:10 <zzo38> That is true, but, such ranges are not likely what you would want anyways, whether or not the flag is set.
19:13:35 <int-e> (I have a DFA based regex engine around that uses arrays of size 256 for transitions.)
19:13:57 <int-e> (It would be ... interesting ... to try to make it work with utf-8)
19:14:54 <zzo38> You should instead use multiple ranges together if you really want to include both accented and unaccented letters in the same range like that.
19:16:40 <zzo38> int-e: Ah, well I suppose it may be possible to compile UTF-8-based regular expressions into byte-based regular expressions, like I explained how to compile the dot, for example
19:17:02 <int-e> right.
19:17:47 <int-e> I guess negating ranges is the tricky bit.
19:29:06 -!- copumpkin has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…).
19:29:46 <FireFly> I had an issue a while ago where a sed script worked differently depending on the locale, due to character ranges
19:30:30 <FireFly> it even happened when I used escapes. You'd think [\x30-\x3f] would unambiguously match the set of bytes 0x30..0x3f, but apparently it doesn't
19:30:40 <int-e> collating is the worst idea ever.
19:31:30 <elliott> it's almost like bolting unicode on in a kind of ad-hoc manner after the fact doesn't go to well :p
19:31:34 <int-e> No, [a-z] should not include B-Y and one of A or Z (I forgot which one), oh and äöüÄÖÜ while we're at it)
19:33:02 <elliott> character ranges just seem like a fundamentally poor idea for unicode to me...
19:37:36 <izabera> sorry i had to go
19:37:38 <izabera> $ suckless sed 's/[à]/x/' <<< è
19:37:41 <izabera>
19:37:46 <izabera> this is... broken
19:37:50 <zzo38> That's stupid it shouldn't use Unicode such thing
19:38:11 <zzo38> It shouldn't use the locale
19:40:33 <zzo38> In byte mode it should match all bytes in range, in UTF-8 mode it should match all sequences of bytes corresponding to the shortest UTF-8 encoding of the codepoints in the given range.
19:41:08 <izabera> zzo38: what do you mean with stupid? my example or their sed's behaviour?
19:41:30 <izabera> because it obviously should use unicode imho
19:41:50 <zzo38> No it shouldn't use Unicode, unless you set a flag to tell it to do so
19:42:12 <izabera> ??? why does that make any sense? the flag should be always active
19:42:37 <zzo38> No it shouldn't, you should use single byte encodings by default (the program doesn't have to care which)
19:42:49 <izabera> it's 2015
19:43:09 <elliott> not for zzo it isn't
19:43:22 <zzo38> The alternative is to provide a flag to turn it off instead
19:43:39 <izabera> your reasoning is maybe valid in a 1975 background
19:43:56 <int-e> `` LC_CTYPE=C sed 's/[à]/x/' <<< è
19:44:12 <HackEgo> x
19:44:23 <int-e> `` sed 's/[à]/x/' <<< è
19:44:25 <HackEgo> ​è
19:44:50 <izabera> yay gnu
19:45:20 <FireFly> Excellent
19:45:22 <int-e> `` locale
19:45:38 <FireFly> New zealand? whatever floats your boat
19:45:44 <zzo38> But I think it shouldn't use the locale, instead add a regex flag after the / to tell it to use UTF-8, if you don't specify then it doesn't, and either way doens't care about what language or anything like that
19:46:06 <izabera> did you happen to read the posix standard
19:46:28 <zzo38> Now you have to add another shortcut that will change the locale to C when executing the command
19:47:17 <zzo38> In UNIX systems I have access to I had to even change the locale to C in my login scripts, since they did it bad too
19:47:27 <elliott> technically unicode did not exist in 1975 :p
19:48:13 <elliott> some parts of the locale system *do* seem to mess things up more than they help, but... that's unix for you
19:48:32 <izabera> you mean [a-z] matching X ?
19:48:41 <izabera> that's what [[:lower:]] is for
19:48:51 <elliott> I think [a-z] is kind of unreasonable to start with.
19:49:04 <zzo38> I still think [a-z] matching X is stupid unless case-insensitive mode is selected
19:49:17 <zzo38> By default it shouldn't do that
19:49:19 <elliott> (but I do admit that I have no idea how to reliably match "ASCII characters a to z". what if you're grepping for some programming language's syntax?)
19:49:43 <elliott> I find, e.g. locale-based ordering annoying in "ls". it messes up SHOUTYFILES coming first, iirc
19:50:02 <pikhq> LC_CTYPE=C [a-z]
19:50:21 <elliott> well, that's the other thing
19:50:27 <pikhq> (if [a-z[ in the C locale matches X yell at your vendor)
19:50:42 <elliott> sadly unix is a mess and a whole bunch of software is never going to make sure the locale is right for what it does in its scripts or whatever
19:50:54 <zzo38> I agree it is stupid, I want to force the locale to be C which is why I set up the login script to force the locale to C
19:50:55 <int-e> I hate almost everything about locales, but I do have LC_CTYPE=en_US.UTF-8 (why the en_US?) set.
19:51:22 <zzo38> int-e: I too, but I prefer to set it to C instead of en_US.UTF-8
19:51:25 <elliott> (I do like localised software, and I do think all software should handle Unicode, no excuses.)
19:51:29 <pikhq> The C/Unix locale system isn't *that* bad, but very few people understand it well.
19:51:34 <elliott> (I just think the Unix locale system is about as much of a mess as Unix itself.)
19:51:45 <pikhq> elliott: You'd like musl's locale approach.
19:51:59 <pikhq> "Unicode is the only supported charset".
19:52:08 <int-e> (Actually I hate localised error messages, because I have to read them, rather than recognize them by pattern matching.)
19:52:09 <elliott> does musl even support locales to the point of letting you get error messages in your language from ls or whatever?
19:52:13 <pikhq> Yes.
19:52:20 <zzo38> I disagree; only documentation and GUIs should be localised.
19:52:40 <pikhq> Nobody's bothered doing much localization, but there's locale infrastructure in place just fine.
19:52:41 <elliott> I know you disagree with almost everything I say >_>
19:52:44 <zzo38> And, not all software needs to handle Unicode; it depends on the software.
19:52:57 <int-e> elliott: you're wrong!
19:53:06 <zzo38> Most software probably shouldn't need to handle Unicode.
19:53:14 <pikhq> I think there's a couple omissions?
19:53:24 <pikhq> ... Oh *right*, it doesn't support localized collation.
19:54:00 <elliott> we should all go back to the original Emacs-on-TECO ported to TOPS-20, imo, argument over
19:54:04 <FireFly> [a-z] matching X if /i is given is reasonable to me
19:54:21 <elliott> oh nice TECO even ran on TOPS-20!
19:54:26 <FireFly> without /i any range should match exactly that range of bytes IMO
19:54:30 <pikhq> And (not exactly the same thing as locales, but somewhat relevant to some localized users) musl's iconv doesn't support *extensions* to its charset support ATM.
19:54:55 <zzo38> Making the sort command to support user-defined collation sequences (independent of chaacter set) seem reasonable to me; most other programs don't need to, but there are some such as SQLite that you can add extensions for user-defined collation sequences already too.
19:55:01 <elliott> izabera: I apologise for any part i had in starting this argument -_-
19:55:10 <izabera> lol
19:55:56 <pikhq> Which means you're going to have some issues if the legacy charset for your languange is not sufficiently common to have been implemented in iconv.
19:58:08 <zzo38> SQLite has no SQL commands to implement collation sequences, but you can write them in C and then load them using the LOAD_EXTENSION function in SQL.
19:58:42 <pikhq> So, SOL if it's not: UTF-8, UCS2, UTF-16, UTF-32, ASCII, EUC-JP, Shift-JIS, GB18030, GBK, GB2312, Big5, EUC-KR, ISO-8859*, or one of the more popular Windows code pages.
19:59:37 <int-e> what about EBCDIC?
19:59:51 -!- Nokaji has left.
20:00:18 <pikhq> Nope.
20:00:18 <int-e> (IBM's proprietary encryption)
20:01:27 <pikhq> (fuck EBCDIC)
20:02:49 <zzo38> EBCDIC isn't very good; ASCII is better.
20:03:23 <zzo38> I usually use ASCII when possible though rather than any other characters sets; most are compatible with ASCII though, so won't be much problem
20:04:32 <int-e> Ah, from Wikipedia: Professor: "So the American government went to IBM to come up with an encryption standard, and they came up with—" Student: "EBCDIC!"
20:10:36 <Sgeo_> Is using cookies on GitHub Pages actually -safe-?
20:11:00 <Sgeo_> github.io is on the public suffix list
20:11:32 -!- copumpkin has joined.
20:14:10 <int-e> Sgeo_: "safe"?
20:14:46 <Sgeo_> As in, if I run a.github.io and use cookies for stuff, what chance does evil.github.io have of breaking stuff
20:16:25 <zzo38> The VCR I have has no zero-return function. You can reset the counter to zero, but doesn't have the mode to automatically fast-forward/rewind until reaching zero and then stop.
20:16:37 <int-e> it's a matter of setting the cookie's domain and path, isn't it?
20:17:52 <Sgeo_> What if evil overrides a cookie that a is using?
20:18:05 <Sgeo_> But public suffix list entry for github.io should block that I think
20:18:09 <Sgeo_> For browsers that use that list
20:18:16 <Sgeo_> Fuck cookies for needing a list like that
20:20:12 <int-e> Sgeo_: evil.github.io should not be able to set a cookie for a.github.io
20:20:13 <zzo38> Just mention on your webpage to add such an entry to the list
20:21:13 <int-e> (nor should the browser send a cookie for a.github.io in requests to evil.github.io)
20:22:13 <zzo38> Another thing you can use the client's IP address for additional security measures
20:24:01 <int-e> Sgeo_: I'd assume that the purpose of the "public suffix list" is that websites cannot set cookies with such suffixes; i.e. evil.github.io cannot set a cookie for the domain .github.io, which would then be sent along with all requests to github.io websites.
20:24:12 <Sgeo_> int-e: right
20:24:15 -!- boily has quit (Quit: MURAL CHICKEN).
20:24:25 <int-e> so I'm unsure what you're worried about here.
20:25:21 <Sgeo_> Browsers that don't use the Public Suffix List?
20:25:43 <b_jonas> int-e: no
20:25:46 <int-e> the cookies would still be for the wrong domains.
20:26:08 <b_jonas> int-e: the cookies contain the domains, so the servers can just ignore such cookies
20:26:24 <int-e> b_jonas: yes.
20:26:26 <zzo38> A user-defined public suffix list is probably a better idea, although you can still provide lists of default values too
20:26:36 <elliott> you don't need to guess what the public suffix list is for: https://publicsuffix.org/
20:26:40 <elliott> it says right there
20:26:51 <b_jonas> int-e: the goal of the public suffix list is to prevent a DOS attack and possibly information leakage where a website would make your browser send a dozen large cookies to every site with a *.com domain
20:26:55 <b_jonas> I think
20:27:23 <elliott> you could never set cookies for *.com, anyway
20:27:43 <int-e> elliott: "Avoid privacy-damaging "supercookies" being set for high-level domain name suffixes " is what I was thinking about
20:27:48 <elliott> right
20:28:14 <b_jonas> yep
20:28:19 <int-e> b_jonas: I'm not sure what exactly I wrote that you think is wrong.
20:30:13 <int-e> (Except that I didn't consider the DoS angle of attack.)
20:31:14 <b_jonas> int-e: sorry, nothing, it's fine
20:31:18 <b_jonas> you wrote the right thing
20:31:35 -!- adu has joined.
20:33:21 -!- Phantom_Hoover has joined.
20:33:35 -!- pdxleif has joined.
21:04:30 -!- copumpkin has quit (Read error: Connection reset by peer).
21:05:09 -!- Albert_ has joined.
21:11:57 -!- copumpkin has joined.
21:18:30 -!- Albert_ has quit (Quit: Page closed).
21:22:49 -!- AnotherTest has joined.
21:28:17 -!- tromp has quit (*.net *.split).
21:28:18 -!- sebbu has quit (*.net *.split).
21:28:18 -!- mitchs has quit (*.net *.split).
21:28:19 -!- b_jonas has quit (*.net *.split).
21:28:19 -!- ineiros_ has quit (*.net *.split).
21:28:19 -!- olsner has quit (*.net *.split).
21:28:21 -!- EgoBot has quit (*.net *.split).
21:28:30 -!- mitchs has joined.
21:28:32 -!- b_jonas has joined.
21:28:32 -!- ineiros has joined.
21:28:51 -!- EgoBot has joined.
21:28:52 -!- tromp has joined.
21:33:56 -!- olsner has joined.
21:38:12 <zzo38> Do you have Family Channel on your television?
21:39:16 -!- shikhin has quit (Quit: leaving).
21:39:20 <int-e> sigh ... how do I quote an @foo on github so it doesn't get treated as an @mention thing, but gets displayed as ordinary text?
21:39:40 <zzo38> I don't know?
21:39:53 <int-e> `@foo` disables the @mention but adds some markup...
21:39:54 <HackEgo> ​/home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: @foo`: not found
21:43:35 * int-e tries #github... sorry.
21:43:38 <Melvar> int-e: Does \@foo not do that?
21:44:07 <int-e> Melvar: unfortunately, no. The \ is displayed, and the @foo is treated as before.
21:44:25 <Melvar> That is quite broken.
21:44:45 -!- copumpkin has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…).
21:45:59 <int-e> (I keep running into this because @ is lambdabot's command prefix. Using `` for it is okay-ish, but not what I want.)
21:54:15 -!- Phantom__Hoover has joined.
22:00:39 <fizzie> int-e: Maybe @⁢foo?
22:14:09 -!- adu has quit (Quit: adu).
22:15:01 -!- Patashu has joined.
22:17:25 -!- AnotherTest has quit (Remote host closed the connection).
22:21:31 <int-e> wow, #github is dead.
22:52:16 -!- oerjan has joined.
23:00:35 -!- adu has joined.
23:01:52 -!- CADD has quit (Remote host closed the connection).
23:27:52 <oerjan> @tell boily <boily> fungot: I wouldn't mind cats or dogs or diamondies pouring down. as long as it isn't any more snow. <-- i suspect diamondies would be even harder on the car drivers. btw oslo was also unexpectedly hit by snow yesterday, complete chaos at the traditional start of easter vacation week.
23:27:52 <lambdabot> Consider it noted.
23:27:52 <fungot> oerjan: slime for scheme48: http://paste.lisp.org/ list
23:27:54 -!- sebbu has joined.
23:28:16 <oerjan> fungot: i already have more slime in my life than i need, thank you very much.
23:28:16 <fungot> oerjan: it wasn't a dream).
23:28:32 -!- sebbu has quit (Changing host).
23:28:32 -!- sebbu has joined.
23:28:49 -!- yorick has joined.
23:29:47 -!- scott has joined.
23:30:09 -!- yorick has quit (Remote host closed the connection).
23:31:55 <int-e> fungot: eww
23:31:55 <fungot> int-e: no not fnord :p.) appears to be writing about the implementation.
23:32:24 -!- yorick has joined.
23:32:24 -!- yorick has quit (Changing host).
23:32:24 -!- yorick has joined.
23:35:28 -!- Phantom__Hoover has quit (Read error: Connection reset by peer).
23:46:35 -!- adu has quit (Quit: adu).
23:47:54 <int-e> fizzie: putting an &#x2060; after the @ "works".
23:48:04 <int-e> `unidecode @⁠foo
23:48:25 <int-e> But may confuse people wanting to cut&paste. Oh well.
23:48:53 <oerjan> int-e: hey, i also use en_us.utf-8 for essentially the same reason.
23:49:16 <FireFly> I use en_GB.UTF-8 because why not
←2015-03-27 2015-03-28 2015-03-29→ ↑2015 ↑all