Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 .. 5022
— 502,152 events total
2021-03-27 17:21:19 <maerwald> yeah, I want to avoid exactly those assumptions
2021-03-27 17:21:43 × frozenErebus quits (~frozenEre@37.231.244.249) (Quit: leaving)
2021-03-27 17:21:47 <ephemient> you will have to handle unexpected failures later, regardless of what you assume up front
2021-03-27 17:21:55 <maerwald> the question now is how to unify over windows and unix (given that we assume all ByteStrings are valid on unix)
2021-03-27 17:22:15 <maerwald> when converting FROM ByteString
2021-03-27 17:22:19 ← jakalx parts (~jakalx@base.jakalx.net) ("Error from remote client")
2021-03-27 17:23:22 <ephemient> I would leave it as a bytestring until open() time, which can fail due to encoding *or* system-specific reasons
2021-03-27 17:23:48 → ddellacosta joins (~ddellacos@86.106.143.92)
2021-03-27 17:24:00 <maerwald> this is what I've done: https://git.io/JYsjV
2021-03-27 17:25:34 × gentauro quits (~gentauro@unaffiliated/gentauro) (Quit: leaving)
2021-03-27 17:25:37 × zebrag quits (~inkbottle@aaubervilliers-654-1-98-245.w86-212.abo.wanadoo.fr) (Quit: Konversation terminated!)
2021-03-27 17:25:46 × MrMobius quits (~MrMobius@208.58.206.154) (Ping timeout: 245 seconds)
2021-03-27 17:25:53 <maerwald> FileSystem support is up to an application to worry about imo. E.g. some applications don't support AUFS, because it breaks `rename` sometimes (you need portable file copy).
2021-03-27 17:26:00 → zebrag joins (~inkbottle@aaubervilliers-654-1-98-245.w86-212.abo.wanadoo.fr)
2021-03-27 17:26:41 <ephemient> I don't see why you need detect invalid Windows filenames earlier than `open` time
2021-03-27 17:26:44 → MrMobius joins (~MrMobius@208.58.206.154)
2021-03-27 17:26:54 <maerwald> ephemient: that's the invariant for the type
2021-03-27 17:26:56 <ephemient> you can't detect many cases of invalid filenames without trying to `open` them, across all OSes
2021-03-27 17:27:09 × gxt quits (~gxt@gateway/tor-sasl/gxt) (Ping timeout: 240 seconds)
2021-03-27 17:27:13 → wroathe joins (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net)
2021-03-27 17:27:18 <maerwald> did you read the abstract filepath proposal?
2021-03-27 17:27:20 → pera joins (~pera@unaffiliated/pera)
2021-03-27 17:28:33 × ddellacosta quits (~ddellacos@86.106.143.92) (Ping timeout: 265 seconds)
2021-03-27 17:28:33 → gentauro joins (~gentauro@unaffiliated/gentauro)
2021-03-27 17:29:13 → gxt joins (~gxt@gateway/tor-sasl/gxt)
2021-03-27 17:29:22 <ephemient> `data WindowsFilePath = WFP ByteArray# -- UTF16 data` doesn't seem like the right thing to do, to me
2021-03-27 17:29:42 <maerwald> well, I'm implementing it, so this is assumed as correct
2021-03-27 17:31:54 <maerwald> I'm less worried about the internal representation decisions, but how the abstraction will play out (or if there's no reasonable one). Converting from ByteString is one of those things that can't be properly abstracted over. Another option is to just refer ppl to the internals if they need that kind of thing
2021-03-27 17:32:59 → Vadrigar_ joins (~Vadrigar@ip5b417208.dynamic.kabel-deutschland.de)
2021-03-27 17:33:16 <maerwald> that's why there are 2 constructors
2021-03-27 17:33:38 <maerwald> so you don't accidentially write one codepath if you start matching on the raw data
2021-03-27 17:34:14 → jakalx joins (~jakalx@base.jakalx.net)
2021-03-27 17:35:02 <ephemient> if I were targeting W10+ only, I'd use UTF-8 across all platforms, period
2021-03-27 17:36:14 → vicfred joins (vicfred@gateway/vpn/mullvad/vicfred)
2021-03-27 17:36:30 <ephemient> with older Windows in the mix... `data WindowsFilePath = WFPW {- UTF-16 -} | WFPA {- UTF-8 -}` I guess. a bit unfortunate in that it would require a custom Eq that uses the current codepage for conversion
2021-03-27 17:37:49 → d34df00d joins (~d34df00d@104-14-27-213.lightspeed.austtx.sbcglobal.net)
2021-03-27 17:37:52 <d34df00d> Hi!
2021-03-27 17:38:11 × Vadrigar_ quits (~Vadrigar@ip5b417208.dynamic.kabel-deutschland.de) (Ping timeout: 240 seconds)
2021-03-27 17:40:31 × madjest14 quits (~Android@2a02-a420-47-b3b2-8077-257-ad0b-f19d.mobile6.kpn.net) (Ping timeout: 276 seconds)
2021-03-27 17:41:24 <maerwald> ephemient: rust does a very similar thing to the abstract filepath proposal, see OsString. It's Vec<u8> on unix and WTF-8 (more lax utf-8) on windows
2021-03-27 17:41:51 × wroathe quits (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net) (Quit: Lost terminal)
2021-03-27 17:42:29 <maerwald> but it is more than about filepaths, so it's not a clean abstraction
2021-03-27 17:43:07 <maerwald> https://doc.rust-lang.org/std/ffi/index.html#conversions is also interesting
2021-03-27 17:43:38 <maerwald> following that, they would only expose `fromByteString` on unix via CPP ifdef
2021-03-27 17:44:09 × paddymahoney quits (~paddymaho@cpe9050ca207f83-cm9050ca207f80.cpe.net.cable.rogers.com) (Remote host closed the connection)
2021-03-27 17:45:16 → wroathe joins (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net)
2021-03-27 17:45:47 → curl joins (5ce955fb@gateway/web/cgi-irc/kiwiirc.com/ip.92.233.85.251)
2021-03-27 17:46:22 × geekosaur quits (82650c7a@130.101.12.122) (Ping timeout: 240 seconds)
2021-03-27 17:47:11 × Pickchea quits (~private@unaffiliated/pickchea) (Ping timeout: 240 seconds)
2021-03-27 17:47:41 → ezrakilty joins (~ezrakilty@97-113-58-224.tukw.qwest.net)
2021-03-27 17:47:56 <curl> is there a name for the first argument of (a->(o,a)) -> a -> [o]
2021-03-27 17:48:42 <curl> or of that function?
2021-03-27 17:49:18 → pupuupup joins (~pupuupup@node-bv.pool-125-24.dynamic.totinternet.net)
2021-03-27 17:49:56 <maerwald> https://hackage.haskell.org/package/fixed-vector-1.2.0.0/docs/Data-Vector-Fixed.html#v:unfoldr
2021-03-27 17:49:58 <maerwald> unfold?
2021-03-27 17:50:46 hackage zenacy-unicode 1.0.1 - Unicode utilities for Haskell https://hackage.haskell.org/package/zenacy-unicode-1.0.1 (mlcfp)
2021-03-27 17:50:47 <curl> :t unfoldr
2021-03-27 17:50:48 <lambdabot> (b -> Maybe (a, b)) -> b -> [a]
2021-03-27 17:51:01 <curl> yeah i see
2021-03-27 17:51:03 <curl> thanks
2021-03-27 17:51:46 hackage zenacy-html 2.0.3 - A standard compliant HTML parsing library https://hackage.haskell.org/package/zenacy-html-2.0.3 (mlcfp)
2021-03-27 17:52:12 <curl> i guess i can always use Just to make the maybe version
2021-03-27 17:53:47 → heatsink joins (~heatsink@2600:1700:bef1:5e10:1cfc:5265:7008:dc13)
2021-03-27 17:54:15 <curl> how do i compose things like that together? i guess i need a fold inbetween to use fusion
2021-03-27 17:55:39 → paddymahoney joins (~paddymaho@cpe9050ca207f83-cm9050ca207f80.cpe.net.cable.rogers.com)
2021-03-27 17:55:55 × plakband quits (~plakband@softbank126227040096.bbtec.net) (Quit: WeeChat 2.9)
2021-03-27 17:55:57 <curl> probably just successive traversals and mappings, but i want the data to feed through one element at a time through some kind of sequence of processes each of which has its own memory state
2021-03-27 17:56:15 <curl> at least notionally
2021-03-27 17:56:36 <curl> or, because that can be a way of looking at it, that maybe there is a way of coding that is like that
2021-03-27 17:56:45 × gioyik quits (~gioyik@gateway/tor-sasl/gioyik) (Ping timeout: 240 seconds)
2021-03-27 17:57:23 <curl> does that make sense? feeding the data one element at a time through some kind of stationary bunch of functions
2021-03-27 17:57:28 <maerwald> > unfoldr (\b -> if b < 5 then Just (b, b + 1) else Nothing) 1
2021-03-27 17:57:31 <lambdabot> [1,2,3,4]
2021-03-27 17:58:02 <maerwald> b is your state
2021-03-27 17:58:18 <maerwald> `Nothing` aborts the loop
2021-03-27 17:58:34 <curl> :t mapAccumr undefined undefined $ unfoldr (\b -> if b < 5 then Just (b, b + 1) else Nothing) 1
2021-03-27 17:58:35 <lambdabot> error:
2021-03-27 17:58:36 <lambdabot> • Variable not in scope: mapAccumr :: t0 -> t1 -> [a0] -> t
2021-03-27 17:58:36 <lambdabot> • Perhaps you meant one of these:
2021-03-27 17:59:22 × heatsink quits (~heatsink@2600:1700:bef1:5e10:1cfc:5265:7008:dc13) (Ping timeout: 276 seconds)
2021-03-27 18:02:22 → ddellacosta joins (~ddellacos@86.106.143.66)
2021-03-27 18:02:27 <curl> > mapAccumR (\x y -> (x+1,x+y)) 0 $ unfoldr (\b -> if b < 5 then Just (b, b + 1) else Nothing) 1
2021-03-27 18:02:29 <lambdabot> (4,[4,4,4,4])
2021-03-27 18:02:56 → geekosaur joins (82650c7a@130.101.12.122)
2021-03-27 18:03:16 × borne quits (~fritjof@200116b8640fd800f7ed9fd86a2491f0.dip.versatel-1u1.de) (Ping timeout: 240 seconds)
2021-03-27 18:03:24 → supersven[m] joins (supersvenm@gateway/shell/matrix.org/x-wgclishhljzvmcme)
2021-03-27 18:03:38 <curl> > mapAccumL (\x y -> (x+1,x+y)) 0 $ unfoldr (\b -> if b < 5 then Just (b, b + 1) else Nothing) 1
2021-03-27 18:03:41 <lambdabot> (4,[1,3,5,7])
2021-03-27 18:04:11 <curl> see how the 0 enters and acts as a state undergoing +1
2021-03-27 18:04:52 × tromp quits (~tromp@dhcp-077-249-230-040.chello.nl) (Remote host closed the connection)
2021-03-27 18:05:24 <curl> so you see the idea of "composing" these things together, with them aech having their own state?
2021-03-27 18:06:10 → geowiesnot joins (~user@i15-les02-ix2-87-89-181-157.sfr.lns.abo.bbox.fr)
2021-03-27 18:06:55 → tromp joins (~tromp@dhcp-077-249-230-040.chello.nl)
2021-03-27 18:06:56 × ddellacosta quits (~ddellacos@86.106.143.66) (Ping timeout: 260 seconds)
2021-03-27 18:07:19 × e quits (e@freenode/staff/spy.edk) (Quit: edk)
2021-03-27 18:07:27 <curl> i guess there is some condition of the first argument of mapAccumL where it can factorise over 2 mapAccumL's in series
2021-03-27 18:08:24 <curl> giving some rewrite rule that mashed them all into one call to mapAccumL
2021-03-27 18:08:34 <curl> or even by fusion right into the unfold
2021-03-27 18:09:01 <curl> mashes*
2021-03-27 18:09:07 → ddellacosta joins (~ddellacos@86.106.143.27)
←Prev  Next→
Page 1 .. 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 .. 5022

All times are in UTC.