Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 .. 5022
— 502,152 events total
2020-11-04 12:00:02 × BjarniRunar1 quits (~BjarniRun@84.39.116.180) ()
2020-11-04 12:00:12 → GUEST1604491197 joins (~GUEST1604@124.123.107.91)
2020-11-04 12:00:28 <tomsmeding> though perhaps you can attempt to see head/fromMaybe/toEnum as a different "class" of partial functions than the case I described, but that's getting firmly and deeply in the subjective realm :p
2020-11-04 12:00:52 → ahmr88 joins (~ahmr88@cpc85006-haye22-2-0-cust131.17-4.cable.virginm.net)
2020-11-04 12:01:08 <bqv> i actually hit the problem you described earlier
2020-11-04 12:01:26 → chaosmasttter joins (~chaosmast@p200300c4a7117c01215e340256877fbe.dip0.t-ipconnect.de)
2020-11-04 12:01:27 × GUEST1604491197 quits (~GUEST1604@124.123.107.91) (Read error: Connection reset by peer)
2020-11-04 12:01:28 <bqv> my solution was undefined, which is... kinda worse
2020-11-04 12:01:29 <merijn> Enum is a terrible class anyway :p
2020-11-04 12:01:37 <merijn> If anything I'd just recommend avoiding Enum :p
2020-11-04 12:01:43 → GUEST1604491290 joins (~GUEST1604@124.123.107.91)
2020-11-04 12:02:00 <bqv> as usual, next to impossible if you FFI at all
2020-11-04 12:02:11 → kish joins (~oracle@unaffiliated/oracle)
2020-11-04 12:02:27 <merijn> How so?
2020-11-04 12:03:07 <bqv> C enums would have to be sent back and forth somehow
2020-11-04 12:03:24 <bqv> any haskell sum type eventually ends up as a number
2020-11-04 12:03:57 <bqv> you could i guess have checks at every marshal in, but that's no better than Enum with safeToEnum
2020-11-04 12:04:02 <merijn> I fail to see how that requires the Enum class, though?
2020-11-04 12:04:25 × GUEST1604491290 quits (~GUEST1604@124.123.107.91) (Read error: Connection reset by peer)
2020-11-04 12:04:35 <bqv> well what would be an alternative?
2020-11-04 12:05:52 <dminuoso> bqv: I tend to have my own Enumerable typeclass that adds a MonadFail constraint to toEnum
2020-11-04 12:05:53 → GUEST1604491541 joins (~GUEST1604@124.123.107.91)
2020-11-04 12:06:10 <dminuoso> For things where want to keep a mapping to numbers (usually I use this for protocol de/serializatoin)
2020-11-04 12:06:16 <merijn> Just writing a single function to convert, safeToEnum only works for bounded types and makes a bunch of assumptions about the underlying Enum instance (and I'd be war to make any assumptions of those)
2020-11-04 12:06:31 <dminuoso> (The reason for using a typeclass is mostly convenience, it could just be plain functions)
2020-11-04 12:06:43 → thir joins (~thir@p200300f27f0b7e00f4e9381c2bf90854.dip0.t-ipconnect.de)
2020-11-04 12:07:02 <dminuoso> And that typeclass is not about *enumerating* from/to, it's just for mapping into something
2020-11-04 12:07:11 <bqv> oh yeah, just, generally i considered any function to be of vastly the same nature as Enum anyway
2020-11-04 12:07:47 × GUEST1604491541 quits (~GUEST1604@124.123.107.91) (Remote host closed the connection)
2020-11-04 12:07:51 → Tops2 joins (~Tobias@dyndsl-095-033-016-001.ewe-ip-backbone.de)
2020-11-04 12:08:43 <dminuoso> bqv: class (t ~ MappedInto f, Show t) => EnumMapping f where type MappedInto t; mappingTo :: f -> t; mappingFrom :: MonadFail m => t -> m f
2020-11-04 12:08:46 <dminuoso> Is roughly what I use
2020-11-04 12:09:52 <bqv> pretty general, fair enough
2020-11-04 12:09:54 <dminuoso> oh heh, that hsould be `type MappedInto s`
2020-11-04 12:10:09 <dminuoso> or f, rather. bah, my code is not copy pastable for other reasons
2020-11-04 12:10:20 → GUEST91807 joins (~GUEST9180@124.123.107.91)
2020-11-04 12:10:30 <bqv> curious why the Show constraint at that level
2020-11-04 12:11:02 × ahmr88 quits (~ahmr88@cpc85006-haye22-2-0-cust131.17-4.cable.virginm.net) (Remote host closed the connection)
2020-11-04 12:11:04 × thir quits (~thir@p200300f27f0b7e00f4e9381c2bf90854.dip0.t-ipconnect.de) (Ping timeout: 244 seconds)
2020-11-04 12:11:52 → Chi1thangoo joins (~Chi1thang@87.112.60.168)
2020-11-04 12:11:57 <merijn> For the error to MonadFail, presumably
2020-11-04 12:12:04 <dminuoso> Mmm. Good question, I guess there is no need for it there
2020-11-04 12:12:15 → bitmagie joins (~Thunderbi@200116b80654c700253602a817d4606f.dip.versatel-1u1.de)
2020-11-04 12:12:44 <dminuoso> bqv: The reason I have it that general, is because Im writing a protocol library here, so some enums are mapped into Word8, others into Word32, etc..
2020-11-04 12:12:53 <dminuoso> (So that helper typeclass is then used in my Get/Put code)
2020-11-04 12:13:07 <bqv> ah fair
2020-11-04 12:13:36 <dminuoso> It was either that, or sprinkling fromIntegral left and right, not knowing if I accidentally narrow data.
2020-11-04 12:13:50 <merijn> Or...just write individual Get/Put operations? >.>
2020-11-04 12:14:07 <dminuoso> merijn: That would have worked too, of course. The typeclass was really just for convenience.
2020-11-04 12:14:37 <dminuoso> It reduces some boilerplate in some areas. :)
2020-11-04 12:14:37 × GUEST91807 quits (~GUEST9180@124.123.107.91) (Ping timeout: 260 seconds)
2020-11-04 12:14:38 <merijn> dminuoso: The convenience of being unable to have multiple versions in your codebase? >.>
2020-11-04 12:14:45 <dminuoso> merijn: That's perfectly fine for me.
2020-11-04 12:14:51 <dminuoso> It's an internal typeclass I dont ever expose.
2020-11-04 12:15:07 <dminuoso> And these things do *not* ever change
2020-11-04 12:15:24 <bqv> merijn: as a fan of a strictly typed language, surely you appreciate restriction :D
2020-11-04 12:15:43 <dminuoso> bqv: No, it's an annoying restriction since if you ever have multiple protocol versions, you need to start using newtype wrappers
2020-11-04 12:15:56 <dminuoso> But in my case, the protocol is standardized and hasnt changed in 20 years. :p
2020-11-04 12:16:02 <merijn> You fool! You fell victim to one of the classic blunders - the most famous of which is "never get involved in a land war in Asia" and "Never go in against a Sicilian when death is on the line", but only slightly less well-known is this: "Never use typeclasses for (de)serialisation"! Ha ha ha ha ha ha ha! Ha ha ha ha ha ha ha!
2020-11-04 12:16:03 <dminuoso> (It was only ever ammended)
2020-11-04 12:16:07 <bqv> heh
2020-11-04 12:16:19 <dminuoso> merijn: I dont use typeclasses for get/put itself.
2020-11-04 12:16:23 <dminuoso> Or actually, that's not true
2020-11-04 12:16:28 <dminuoso> I have my own Get/Put typeclasses!
2020-11-04 12:16:30 <dminuoso> Again, for convenience
2020-11-04 12:16:46 <dminuoso> But I know what price I pay, the cost is lower than the extra effort and annoyance from writing functions instead.
2020-11-04 12:16:51 → yoyo joins (50a46f47@80-164-111-71-dynamic.dk.customer.tdc.net)
2020-11-04 12:17:08 <dminuoso> There is ever only a single protocol encoding of `Foo
2020-11-04 12:17:17 <dminuoso> And, should it ever be different, I can turn it into functions easily
2020-11-04 12:18:23 × zfnmxt quits (~zfnmxt@unaffiliated/zfnmxt) (Quit: Bye!)
2020-11-04 12:18:49 × todda7 quits (~torstein@ppp-2-84-30-242.home.otenet.gr) (Ping timeout: 260 seconds)
2020-11-04 12:20:01 hackage servant-multipart 0.12 - multipart/form-data (e.g file upload) support for servant https://hackage.haskell.org/package/servant-multipart-0.12 (maksbotan)
2020-11-04 12:20:13 → zfnmxt joins (~zfnmxt@unaffiliated/zfnmxt)
2020-11-04 12:20:29 × avdb quits (~avdb@ip-62-235-73-30.dsl.scarlet.be) (Quit: WeeChat 2.9)
2020-11-04 12:21:39 × bitmagie quits (~Thunderbi@200116b80654c700253602a817d4606f.dip.versatel-1u1.de) (Quit: bitmagie)
2020-11-04 12:21:55 → vfaronov joins (~vfaronov@broadband-95-84-210-78.ip.moscow.rt.ru)
2020-11-04 12:23:26 × chaosmasttter quits (~chaosmast@p200300c4a7117c01215e340256877fbe.dip0.t-ipconnect.de) (Ping timeout: 264 seconds)
2020-11-04 12:24:30 → asthasr joins (~asthasr@162.210.29.120)
2020-11-04 12:28:46 → knupfer joins (~Thunderbi@dynamic-046-114-148-101.46.114.pool.telefonica.de)
2020-11-04 12:30:01 → bitmagie joins (~Thunderbi@200116b80654c700253602a817d4606f.dip.versatel-1u1.de)
2020-11-04 12:30:17 → elfets joins (~elfets@ip-37-201-23-96.hsi13.unitymediagroup.de)
2020-11-04 12:31:05 → asthasr_ joins (~asthasr@162.210.29.120)
2020-11-04 12:31:20 × asthasr_ quits (~asthasr@162.210.29.120) (Client Quit)
2020-11-04 12:36:40 → avdb joins (~avdb@ip-62-235-73-30.dsl.scarlet.be)
2020-11-04 12:36:40 × knupfer quits (~Thunderbi@dynamic-046-114-148-101.46.114.pool.telefonica.de) (Read error: Connection reset by peer)
2020-11-04 12:37:18 <idnar> merijn: hahaha
2020-11-04 12:37:26 → ahmr88 joins (~ahmr88@cpc85006-haye22-2-0-cust131.17-4.cable.virginm.net)
2020-11-04 12:38:18 × bitmagie quits (~Thunderbi@200116b80654c700253602a817d4606f.dip.versatel-1u1.de) (Quit: bitmagie)
2020-11-04 12:38:45 <idnar> @type \f x -> coerce (f (coerce x))
2020-11-04 12:38:46 <lambdabot> error:
2020-11-04 12:38:46 <lambdabot> • Variable not in scope: coerce :: t1 -> t3
2020-11-04 12:38:47 <lambdabot> • Perhaps you meant ‘coerced’ (imported from Control.Lens)
2020-11-04 12:38:49 × nbloomf quits (~nbloomf@2600:1700:ad14:3020:fc37:ff3f:9c12:2922) (Quit: My MacBook has gone to sleep. ZZZzzz…)
2020-11-04 12:39:07 → GUEST93534 joins (~GUEST9353@124.123.107.91)
2020-11-04 12:39:11 <bqv> unsafeCoerce?
2020-11-04 12:39:21 → bitmagie joins (~Thunderbi@200116b80654c700253602a817d4606f.dip.versatel-1u1.de)
2020-11-04 12:39:28 <merijn> unsafeCoerce is different and, well, unsafe :p
2020-11-04 12:39:35 <idnar> no, I want the safe one
2020-11-04 12:39:37 <bqv> thought that was his plan
2020-11-04 12:39:39 <bqv> fair
←Prev  Next→
Page 1 .. 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 .. 5022

All times are in UTC.