|
hackbunny posted:the discussion was iirc about the many frustratingly similar but incompatible re syntaxes, and it's 100% true: sed has one, grep another, javascript, icu etc. someone proposed, in the usual hyperbolic yospos style, pcre as "the" standard. the word "pcre" triggered you and the rest is history ah, right right, i rather leapt from the ml module system for composing regex syntax/semantics implementations into the suggestion that pcre was the default. the default *syntax* i agree with entirely, but then neither re2, nor anything else really in the conversation since, is stupid enough to deviate from perl syntax, so i was a bit thrown by the conversation doubling back to syntax composition
|
#
?
Sep 21, 2017 05:32
|
|
|
|
| # ? Jun 4, 2024 22:25 |
|
|
the (oca)ml module system is loving great and anyone who disagrees doesn't understand or appreciate its power
|
|
#
?
Sep 21, 2017 06:31
|
|
|
gonadic io posted:counterpoint: if you want to search for .*.*.*.*.* then do that on your own cpu. software should be running on the user's CPU servers are for storing data eschaton fucked around with this message at 06:52 on Sep 21, 2017 |
|
#
?
Sep 21, 2017 06:48
|
|
|
leper khan posted:pretty sure email doesn’t need a . after the @ it also doesn't need a @
|
|
#
?
Sep 21, 2017 06:49
|
|
|
ICU is the standard regular expression syntax because ICU is a standard sort of enough
|
|
#
?
Sep 21, 2017 06:51
|
|
|
JewKiller 3000 posted:the (oca)ml module system is loving great and anyone who disagrees doesn't understand or appreciate its power me i don't. but then i don't really have any idea as to what makes a good or bad module system
|
|
#
?
Sep 21, 2017 07:11
|
|
|
eschaton posted:it also doesn't need a @ this is one of those 'it doesnt matter a library will handle validation' things
|
|
#
?
Sep 21, 2017 10:46
|
|
|
mostly regexes tend to be pretty natural where a human can broadly describe the format. "this field should always start with xyz and then follows the number which we need to extract, terminated by a comma, except when..." i.e. often stuff that is either just slightly denormalized data fields, or stuff that has been formatted with humans in mind but now need to be reversed a bit, like log files and textual file formats. email addresses are the product of a bunch of specifications which don't encode well at all. let the mailer figure it out. also ml modules are neato but one is basically out there with haskell type classes at that point, so practical relevance falls under the haskell curse
|
|
#
?
Sep 21, 2017 11:08
|
|
|
haskell typeclasses are mainstream now. haven't you heard about rust?
|
|
#
?
Sep 21, 2017 11:41
|
|
|
eschaton posted:it also doesn't need a @ Seriously? Since when?
|
|
#
?
Sep 23, 2017 00:08
|
|
|
Doom Mathematic posted:Seriously? Since when? since the beginning iirc, in some early email systems they used an ! instead of an @
|
|
#
?
Sep 23, 2017 00:23
|
|
|
Huh. No specification for email address syntax I've ever seen has been structured like that, the "@" in the middle was always the most obvious unchangeable feature, but I guess I never went back far enough. Do any of those still work?
|
|
#
?
Sep 23, 2017 00:26
|
|
|
probably not, this was back in arpanet times
|
|
#
?
Sep 23, 2017 00:28
|
|
|
check this: In the above path, the local email address of the user who sent the email (Greg Guthrie) was "ihux!grg" https://www.livinginternet.com/e/ew_addr.htm
|
|
#
?
Sep 23, 2017 00:30
|
|
|
fart simpson posted:haskell typeclasses are mainstream now. haven't you heard about rust? last i remember rust didnt have full typeclasses. is this a new development?
|
|
#
?
Sep 23, 2017 01:50
|
|
|
Shinku ABOOKEN posted:last i remember rust didnt have full typeclasses. is this a new development?
|
|
#
?
Sep 23, 2017 03:01
|
|
|
idk about rust (not part of the evangelism strikeforce) but type classes aren't magic, the haskell compiler is just passing a dictionary around for you. if you're willing to do that yourself, you can implement them pretty simply, even in a language that technically "doesn't have typeclasses" like ocaml: http://okmij.org/ftp/Computation/typeclass.html
|
|
#
?
Sep 23, 2017 03:16
|
|
|
Shinku ABOOKEN posted:last i remember rust didnt have full typeclasses. is this a new development? i havent actually looked at rust in a while but from what i remember rust traits are pretty much just haskell typeclasses with different syntax. by "full typeclasses" i assume you mean that in rust you can't implement something like ｍｏｎａｄｓ by using a rust trait? because if that's what you mean it's not because rust traits themselves are lacking, it's because the rust type system doesn't support higher kinded types.
|
|
#
?
Sep 23, 2017 03:56
|
|
|
JewKiller 3000 posted:idk about rust (not part of the evangelism strikeforce) but type classes aren't magic, the haskell compiler is just passing a dictionary around for you. if you're willing to do that yourself, you can implement them pretty simply, even in a language that technically "doesn't have typeclasses" like ocaml: http://okmij.org/ftp/Computation/typeclass.html see also: http://www.haskellforall.com/2012/05/scrap-your-type-classes.html this article gets thrown around quite a bit in the haskell community and it's basically just talking about a way you could do "typeclasses" by just using the normal data structures already built in to haskell
|
|
#
?
Sep 23, 2017 04:00
|
|
|
c# may be getting typeclasses as early as c# 8
|
|
#
?
Sep 23, 2017 04:23
|
|
|
redleader posted:c# may be getting typeclasses as early as c# 8 how does pl thread feel about c# and possibly .net core? c# seems nice but also like There's A Lot Going On.
|
|
#
?
Sep 23, 2017 04:55
|
|
|
C# is a great language with a dumpster fire of a standard library.
|
|
#
?
Sep 23, 2017 05:01
|
|
|
Powaqoatse posted:an ! how is this grammatical or pronounceable ever in english ofc
|
|
#
?
Sep 23, 2017 05:02
|
|
|
Lutha Mahtin posted:how is this grammatical or pronounceable ever an exclamation point?
|
|
#
?
Sep 23, 2017 05:07
|
|
|
Lutha Mahtin posted:how is this grammatical or pronounceable ever a bang (something no YOSPOSter will ever know about, so I get the confusion)
|
|
#
?
Sep 23, 2017 05:08
|
|
|
a witch posted:C# is a great language with a dumpster fire of a standard library. Wrong!
|
|
#
?
Sep 23, 2017 05:11
|
|
|
Sinestro posted:a bang (something no YOSPOSter will ever know about, so I get the confusion) yea bangs are one thing but i said "an ___" so its gotta start w a vowel yo Carthag Tuek fucked around with this message at 05:15 on Sep 23, 2017 |
|
#
?
Sep 23, 2017 05:12
|
|
|
AWWNAW posted:Wrong! which of the six built in json decoders did you settle on?
|
|
#
?
Sep 23, 2017 05:15
|
|
|
a witch posted:C# is a great language with a dumpster fire of a standard library. c# is an ok language with a dumpster fire of a standard library. though it did invent async/await which is now an essential feature in any language that sees serious use in the 2010s.
|
|
#
?
Sep 23, 2017 05:48
|
|
|
whats wrong w/ the c# std lib?
|
|
#
?
Sep 23, 2017 05:50
|
|
|
Shaggar posted:whats wrong w/ the c# std lib? you like it
|
|
#
?
Sep 23, 2017 05:56
|
|
|
fart simpson posted:i havent actually looked at rust in a while but from what i remember rust traits are pretty much just haskell typeclasses with different syntax. by "full typeclasses" i assume you mean that in rust you can't implement something like ｍｏｎａｄｓ by using a rust trait? because if that's what you mean it's not because rust traits themselves are lacking, it's because the rust type system doesn't support higher kinded types.
|
|
#
?
Sep 23, 2017 05:59
|
|
|
a witch posted:which of the six built in json decoders did you settle on? programming is all about tradeoffs. right tool for the job. you can't anticipate your users' every future need. we asked that team to add just one little method we needed for our entire product but they refused! fuckers! so we wrote our own realtalk the obvious correct decision is the one that encodes dates in the most insane way possible, ideally in a way that minimizes interoperability. congrats JavaScriptSerializer you win the prize
|
|
#
?
Sep 23, 2017 06:27
|
|
|
c# is real good
|
|
#
?
Sep 23, 2017 06:35
|
|
|
raminasi posted:c# is real good it is. it's getting pretty complicated with all the new syntax sugar, though.
|
|
#
?
Sep 23, 2017 06:54
|
|
|
c# is good and would be even better if they weren't loving up .net core/standard/dnx (are these the same thing? who can tell) so badly
|
|
#
?
Sep 23, 2017 08:43
|
|
|
Doom Mathematic posted:Seriously? Since when? if you're sending mail from one local user to another there's no need for anything more than a username and as someone else pointed out there was also bang path addressing, common with UUCP back in the day
|
|
#
?
Sep 23, 2017 10:23
|
|
|
Plus the user part can be a quoted string containing an @ so naively just scanning for an @ doesn't guarantee it's a valid email address.
|
|
#
?
Sep 23, 2017 11:12
|
|
only
|
Edit: I might be in the wrong thread
|
|
#
?
Sep 23, 2017 11:16
|
|
|
|
| # ? Jun 4, 2024 22:25 |
|
|
c# is decent but yeah they've badly damaged the platform with the nonsensical overlapping versioning systems for the platform, runtime, libraries, compiler, rest of the toolchain and the ide also failing to give any support to f# outside of research is disappointing
|
|
#
?
Sep 23, 2017 12:15
|
|
