Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 97 98 99 100 101 102 103 104 105 106 107 .. 5022
— 502,152 events total
2020-09-20 18:42:51 <Cale> Our keyboards need more keys
2020-09-20 18:43:02 × ukari quits (~ukari@unaffiliated/ukari) (Remote host closed the connection)
2020-09-20 18:43:24 → wroathe joins (~wroathe@c-73-24-27-54.hsd1.mn.comcast.net)
2020-09-20 18:43:36 <hyperisco> one idea is to decouple the notion of overloading
2020-09-20 18:43:49 → ukari joins (~ukari@unaffiliated/ukari)
2020-09-20 18:44:29 → xerox_ joins (~xerox@unaffiliated/xerox)
2020-09-20 18:44:33 <Cale> Add a subclass of Num called Ring, with no additional operations, but more laws
2020-09-20 18:44:38 <hyperisco> basically what Num does now, at least the report version
2020-09-20 18:44:53 × mpereira quits (~mpereira@2a02:810d:f40:d96:b46b:1e98:8653:4550) (Remote host closed the connection)
2020-09-20 18:44:59 <Cale> and then things which truly care that they have a proper Ring can demand it
2020-09-20 18:45:47 <phadej> Ring is wrong for Natural
2020-09-20 18:46:15 <phadej> even without thinking about Double :)
2020-09-20 18:46:33 <hyperisco> a trouble being that Num defines a lot of syntax, more than say a semiring would want
2020-09-20 18:47:02 <hyperisco> I guess this is the Idris lawful/lawless duality all over again
2020-09-20 18:47:30 <hyperisco> one is just syntax, the other adds semantics
2020-09-20 18:47:56 <hyperisco> maybe just a different mechanism is warranted, than type classes
2020-09-20 18:48:09 <hyperisco> C++ has macros, by preprocessor or template
2020-09-20 18:48:15 <hyperisco> it isn't a shining example but it is more truthful
2020-09-20 18:48:24 × juuandyy quits (~juuandyy@90.166.144.65) (Ping timeout: 260 seconds)
2020-09-20 18:49:01 <phadej> I also don't know how much Ocaml people hate the difference between + and .+
2020-09-20 18:49:39 <phadej> whether the dotted operators are driving people crazy and into Haskell
2020-09-20 18:49:52 <hyperisco> I suspect the best way to solve this is with the editor / code viewer
2020-09-20 18:49:53 <[exa]> is there a squiggly variant of + in utf8 that could mark the approximate + on doubles etc ?
2020-09-20 18:50:14 <[exa]> because this is really just ascii problem
2020-09-20 18:50:14 <hyperisco> but that upends probably the most entrenched aspect of programming
2020-09-20 18:51:06 <hyperisco> I already use compose sequences to type things like →
2020-09-20 18:51:08 <ski> ⍭
2020-09-20 18:51:10 <ski> maybe ?
2020-09-20 18:51:31 <[exa]> looks too precise
2020-09-20 18:52:26 <ski> ⨤
2020-09-20 18:52:32 <hyperisco> if the same symbol was used in two ways then all I'd ask is there be some subtle indication of overloading, be it by colour or an extra mark (especially for the colourblind)
2020-09-20 18:52:42 <hyperisco> and the editor affords a way to unmask the overload
2020-09-20 18:53:01 <[exa]> ski: that's a strong candidate!
2020-09-20 18:53:18 <hyperisco> so then we don't have to play the unicode game either
2020-09-20 18:53:19 × shad0w_ quits (~shad0w_@160.202.36.27) (Remote host closed the connection)
2020-09-20 18:53:56 <ski> (there's ⌜⨦⌝, too)
2020-09-20 18:54:04 <hyperisco> in other words, there is a difference of syntax representation in a way somewhat similar to the compose sequences I use
2020-09-20 18:54:23 <hyperisco> I have multiple ways to type → say, but they present the same
2020-09-20 18:55:06 <hyperisco> and all I'd want to know is that 1) → represents more than one sequence and 2) which sequence it represents, but I only need to know (2) by explicitly asking
2020-09-20 18:55:23 → orion joins (~orion@c-76-19-238-5.hsd1.nh.comcast.net)
2020-09-20 18:55:23 × orion quits (~orion@c-76-19-238-5.hsd1.nh.comcast.net) (Changing host)
2020-09-20 18:55:23 → orion joins (~orion@unaffiliated/orion)
2020-09-20 18:55:31 <hyperisco> so then we can have + and .+ and whatever, and these present both as +
2020-09-20 18:55:56 <hyperisco> plus some signal it is overloaded
2020-09-20 18:56:51 <hyperisco> and all this without having to go off the deep end of projectional editing
2020-09-20 18:57:40 × mariatsji quits (~mariatsji@2a01:79d:53aa:c66c:342b:5324:f59b:5b0e) (Remote host closed the connection)
2020-09-20 18:58:10 → kenran joins (~maier@b2b-37-24-119-190.unitymedia.biz)
2020-09-20 18:58:17 → mariatsji joins (~mariatsji@2a01:79d:53aa:c66c:342b:5324:f59b:5b0e)
2020-09-20 19:00:27 → sfvm joins (~sfvm@37.228.215.148)
2020-09-20 19:00:41 → adam_wespiser joins (~adam_wesp@209.6.42.110)
2020-09-20 19:00:45 × abhixec quits (~abhixec@c-67-169-141-95.hsd1.ca.comcast.net) (Ping timeout: 260 seconds)
2020-09-20 19:01:18 → abhixec joins (~abhixec@c-67-169-141-95.hsd1.ca.comcast.net)
2020-09-20 19:01:20 <sm[m]> I don't suppose there's any easy way to tell GHCI to use the extensions configured for a particular module ?
2020-09-20 19:01:48 adam_wespiser is now known as adamwespiser
2020-09-20 19:02:33 <adamwespiser> On load, you should be able to load a specific module, if I recall correctly, and in that module can be your extensions
2020-09-20 19:02:55 <geekosaur> but that doesn't set them for interactive use
2020-09-20 19:02:57 × mariatsji quits (~mariatsji@2a01:79d:53aa:c66c:342b:5324:f59b:5b0e) (Ping timeout: 260 seconds)
2020-09-20 19:03:08 <sm[m]> right
2020-09-20 19:03:10 × kenran quits (~maier@b2b-37-24-119-190.unitymedia.biz) (Ping timeout: 258 seconds)
2020-09-20 19:03:12 × Dolly quits (585fd1fd@ti0203q160-5312.bb.online.no) (Remote host closed the connection)
2020-09-20 19:03:18 <geekosaur> nor does having multiple modules loaded and switching between them
2020-09-20 19:03:31 <sm[m]> I'm always being thwarted by mismatches between the module and ghci
2020-09-20 19:03:33 × Lycurgus quits (~niemand@98.4.96.130) (Quit: Exeunt)
2020-09-20 19:04:11 <geekosaur> (and this is often a feature, consider ExtendedDefaultRules which you probably don't want turned off by loading a module)
2020-09-20 19:04:16 <sm[m]> I guess I can make some shorter aliases to turn them on/off
2020-09-20 19:05:32 <ddellacosta> kind of a nix question, but can anyone tell me (or point me at docs for) how to override a dependency in a nix Haskell project with a local version of that package which I've compiled myself? Basically I just want to install my copy of the library just for that Haskell project
2020-09-20 19:06:35 <sm[m]> on another note, if anyone here has experience with sdl2, I have questions. Eg, why does destroyWindow not get rid of the window when running in GHCI. And, why does kill -TERM not kill a sdl app, it must be kill -KILL
2020-09-20 19:08:35 → bennofs joins (~quassel@dslb-178-000-064-245.178.000.pools.vodafone-ip.de)
2020-09-20 19:09:04 → thir joins (~thir@p200300f27f0fc600ed2222922a5678d5.dip0.t-ipconnect.de)
2020-09-20 19:10:47 × shafox quits (~shafox@106.51.234.111) (Remote host closed the connection)
2020-09-20 19:12:34 × howdoi quits (uid224@gateway/web/irccloud.com/x-dtashatgyjqumlie) (Quit: Connection closed for inactivity)
2020-09-20 19:12:44 × bennofs_ quits (~quassel@dslb-188-106-250-199.188.106.pools.vodafone-ip.de) (Ping timeout: 272 seconds)
2020-09-20 19:13:04 → notzmv`` joins (~user@177.103.86.92)
2020-09-20 19:14:02 × thir quits (~thir@p200300f27f0fc600ed2222922a5678d5.dip0.t-ipconnect.de) (Ping timeout: 260 seconds)
2020-09-20 19:15:24 × josh quits (~josh@c-67-164-104-206.hsd1.ca.comcast.net) (Remote host closed the connection)
2020-09-20 19:16:00 → josh joins (~josh@c-67-164-104-206.hsd1.ca.comcast.net)
2020-09-20 19:16:05 × notzmv` quits (~user@177.103.86.92) (Ping timeout: 240 seconds)
2020-09-20 19:17:41 → cosimone joins (~cosimone@2001:b07:ae5:db26:b248:7aff:feea:34b6)
2020-09-20 19:18:44 <koz_> Is there a good way to generate a Text randomly which _is_ valid UTF-8, but _not_ valid Latin-1?
2020-09-20 19:19:24 <geekosaur> I don't think there's an invalid Latin-1
2020-09-20 19:19:52 <geekosaur> unless you mean embedded control codes or similar, but that would also apply to UTF-8
2020-09-20 19:19:55 <koz_> I think I'm being imprecise, sorry. I want to generate Text which contains at least one grapheme requiring a multi-byte encoding in UTF-8.
2020-09-20 19:20:05 <koz_> What's a good way to do this?
2020-09-20 19:20:16 × josh quits (~josh@c-67-164-104-206.hsd1.ca.comcast.net) (Ping timeout: 246 seconds)
2020-09-20 19:20:28 <phadej> arbitrary ++ [multibytechar] ++ arbitrary
2020-09-20 19:20:51 <adamwespiser> If you have a bytestring, I might use Data.Text.Encoding for that task
2020-09-20 19:20:53 <geekosaur> include at least one grapheme with a codepoint > 255
2020-09-20 19:21:04 <phadej> geekosaur: >=128
2020-09-20 19:21:16 <phadej> but >255 would work too
2020-09-20 19:21:24 <geekosaur> no, because 128..255 are valid Latin-1
2020-09-20 19:21:36 <phadej> at least one grapheme requiring a multi-byte encoding in UTF-8.
2020-09-20 19:21:39 <koz_> phadej: Yeah, that's a good point.
2020-09-20 19:21:44 <koz_> Thanks!
2020-09-20 19:21:52 <phadej> codepoint 128 is encoded as two bytes
2020-09-20 19:22:04 <phadej> up to koz_ to refine the specification
2020-09-20 19:22:19 <koz_> phadej: The basic idea works well though.
2020-09-20 19:22:28 <koz_> Thanks for the suggestion.
2020-09-20 19:24:58 hackage stm-queue 0.1.2.0 - An implementation of a real-time concurrent queue https://hackage.haskell.org/package/stm-queue-0.1.2.0 (sgschlesinger)
2020-09-20 19:26:09 → bahamas joins (~lucian@unaffiliated/bahamas)
2020-09-20 19:30:27 hackage Win32 2.10.0.0 - A binding to Windows Win32 API. https://hackage.haskell.org/package/Win32-2.10.0.0 (TamarChristina)
←Prev  Next→
Page 1 .. 97 98 99 100 101 102 103 104 105 106 107 .. 5022

All times are in UTC.