Logs: freenode/#haskell
| 2021-03-02 07:17:19 | → | nineonine joins (~nineonine@50.216.62.2) |
| 2021-03-02 07:17:33 | × | merijn quits (~merijn@83-160-49-249.ip.xs4all.nl) (Ping timeout: 245 seconds) |
| 2021-03-02 07:18:04 | → | Graf_Blutwurst joins (~grafblutw@2001:171b:226e:adc0:30e7:f9ef:2218:63e0) |
| 2021-03-02 07:21:14 | <edwardk> | rebuilding, it appears 20 minutes was a low estimate |
| 2021-03-02 07:21:44 | × | nineonine quits (~nineonine@50.216.62.2) (Ping timeout: 260 seconds) |
| 2021-03-02 07:21:48 | <edwardk> | does anyone know off hand the time limit on github-ci actions? |
| 2021-03-02 07:21:48 | × | superstar64 quits (6ccefa7c@108-206-250-124.lightspeed.miamfl.sbcglobal.net) (Quit: Connection closed) |
| 2021-03-02 07:22:05 | → | nineonine joins (~nineonine@2604:3d08:7785:9600:1c8e:e5ce:d317:3075) |
| 2021-03-02 07:22:23 | <edwardk> | /Users/ekmett/.ghcup/bin/ghc: createProcess: runInteractiveProcess: exec: resource exhausted (Argument list too long) |
| 2021-03-02 07:22:25 | <edwardk> | oof |
| 2021-03-02 07:22:36 | <edwardk> | that may be more problematic than anything else |
| 2021-03-02 07:23:21 | <koz_> | edwardk: 72 hours. |
| 2021-03-02 07:23:36 | <koz_> | Oh wait, never mind, not 72. |
| 2021-03-02 07:23:44 | <koz_> | 6 hours. |
| 2021-03-02 07:23:55 | × | xff0x quits (~xff0x@2001:1a81:53bf:4d00:4425:e0a9:a29b:345e) (Ping timeout: 240 seconds) |
| 2021-03-02 07:24:13 | × | nineonine quits (~nineonine@2604:3d08:7785:9600:1c8e:e5ce:d317:3075) (Remote host closed the connection) |
| 2021-03-02 07:24:23 | <edwardk> | it finished building everything then went to put it together and died. |
| 2021-03-02 07:24:44 | × | Tops21 quits (~Tobias@dyndsl-095-033-025-007.ewe-ip-backbone.de) (Read error: Connection reset by peer) |
| 2021-03-02 07:24:54 | → | nineonine joins (~nineonine@2604:3d08:7785:9600:1c8e:e5ce:d317:3075) |
| 2021-03-02 07:24:58 | → | xff0x joins (~xff0x@2001:1a81:53bf:4d00:3684:b677:6699:2fd9) |
| 2021-03-02 07:25:46 | → | LKoen joins (~LKoen@252.248.88.92.rev.sfr.net) |
| 2021-03-02 07:26:01 | <mniip> | edwardk, imagine the hackage docs |
| 2021-03-02 07:26:15 | <edwardk> | that was why i was building it =) |
| 2021-03-02 07:26:26 | <edwardk> | i mean its not _much_ worse here than my gl package |
| 2021-03-02 07:26:31 | <edwardk> | and that still builds |
| 2021-03-02 07:29:18 | × | nineonine quits (~nineonine@2604:3d08:7785:9600:1c8e:e5ce:d317:3075) (Ping timeout: 260 seconds) |
| 2021-03-02 07:29:45 | × | Graf_Blutwurst quits (~grafblutw@2001:171b:226e:adc0:30e7:f9ef:2218:63e0) (Quit: WeeChat 3.0) |
| 2021-03-02 07:30:12 | → | graf_blutwurst joins (~user@2001:171b:226e:adc0:30e7:f9ef:2218:63e0) |
| 2021-03-02 07:30:27 | <mniip> | edwardk, how fixed is the list of reps? |
| 2021-03-02 07:30:48 | <mniip> | could you perhaps collapse one of the layers of the hierarchy |
| 2021-03-02 07:31:48 | <mniip> | hmm, but that loses you that backpack orphan instance handling quirk |
| 2021-03-02 07:31:52 | <edwardk> | yeah |
| 2021-03-02 07:31:55 | <edwardk> | exactly |
| 2021-03-02 07:32:03 | → | coot joins (~coot@37.30.55.141.nat.umts.dynamic.t-mobile.pl) |
| 2021-03-02 07:32:16 | × | coot quits (~coot@37.30.55.141.nat.umts.dynamic.t-mobile.pl) (Remote host closed the connection) |
| 2021-03-02 07:32:22 | <edwardk> | and i really would _like_ to handle 2-tuples properly because then i can handle Maybe# and the like |
| 2021-03-02 07:32:41 | → | coot joins (~coot@37.30.55.141.nat.umts.dynamic.t-mobile.pl) |
| 2021-03-02 07:32:57 | <edwardk> | and (# State# s, a #) -- for a :: TYPE r |
| 2021-03-02 07:33:33 | <edwardk> | but losing 30+ minutes to mostly compile then crashing at the end because of command line length to ghc is kinda frustrating |
| 2021-03-02 07:33:57 | <edwardk> | i thought we'd fixed the ghc <-> cabal command line length issue back when i shipped 'gl' but apparently not |
| 2021-03-02 07:34:28 | <mniip> | what's your `getconf ARG_MAX` say |
| 2021-03-02 07:34:52 | → | Yumasi joins (~guillaume@2a01:e0a:5cb:4430:ac92:f3cb:a8a3:4136) |
| 2021-03-02 07:35:35 | → | nineonine joins (~nineonine@2604:3d08:7785:9600:1c8e:e5ce:d317:3075) |
| 2021-03-02 07:35:39 | <mniip> | it's 2MB on my system and if cabal is trying to pass in 2000 full pathnames I could see that being exhausted |
| 2021-03-02 07:36:07 | → | royal_screwup21 joins (52254809@gateway/web/cgi-irc/kiwiirc.com/ip.82.37.72.9) |
| 2021-03-02 07:36:36 | <mniip> | solution: build in / |
| 2021-03-02 07:37:21 | × | nineonine quits (~nineonine@2604:3d08:7785:9600:1c8e:e5ce:d317:3075) (Remote host closed the connection) |
| 2021-03-02 07:37:25 | × | abhixec quits (~abhixec@c-67-169-139-16.hsd1.ca.comcast.net) (Quit: leaving) |
| 2021-03-02 07:38:01 | <olligobber> | hmm, haskell's type inference isn't as powerful as I thought |
| 2021-03-02 07:38:23 | → | nineonine joins (~nineonine@2604:3d08:7785:9600:1c8e:e5ce:d317:3075) |
| 2021-03-02 07:38:31 | <[exa]> | olligobber: with great power comes great undecidability |
| 2021-03-02 07:38:49 | <olligobber> | I can just add in type annotations I guess |
| 2021-03-02 07:40:09 | <[exa]> | olligobber: any specific problem btw? |
| 2021-03-02 07:40:34 | <nshepperd2> | 'building all the 2-ary representations' sounds concerning, what does that mean |
| 2021-03-02 07:40:55 | × | royal_screwup21 quits (52254809@gateway/web/cgi-irc/kiwiirc.com/ip.82.37.72.9) (Ping timeout: 240 seconds) |
| 2021-03-02 07:41:58 | <edwardk> | mniip: 262144 |
| 2021-03-02 07:42:09 | <mniip> | why is it so small :( |
| 2021-03-02 07:42:13 | <edwardk> | osx |
| 2021-03-02 07:43:18 | × | nineonine quits (~nineonine@2604:3d08:7785:9600:1c8e:e5ce:d317:3075) (Ping timeout: 260 seconds) |
| 2021-03-02 07:43:43 | <mniip> | nshepperd2, [TupleRep [r1, r2] | r1 <- primitiveReps, r2 <- primitiveReps] |
| 2021-03-02 07:43:46 | <mniip> | except backpack |
| 2021-03-02 07:44:45 | <edwardk> | nshepperd2: https://github.com/ekmett/unboxed/blob/main/unboxed.cabal#L191 |
| 2021-03-02 07:44:58 | → | sammuel86 joins (~sammuel86@217.146.82.202) |
| 2021-03-02 07:45:21 | <edwardk> | the code from there to the end is what breaks |
| 2021-03-02 07:46:34 | ← | feuerbach_ parts (~feuerbach@2a03:b0c0:1:d0::408:6001) ("Leaving") |
| 2021-03-02 07:46:59 | <edwardk> | building in the root doesn't help, btw =) |
| 2021-03-02 07:47:11 | <edwardk> | /Users/ekmett/unboxed/dist-newstyle/build/x86_64-osx/ghc-9.0.1/unboxed-0.2/l/consdef/unboxed-0.2-inplace-consdef+3tP2tTLTg5qEHov4661Voa/build/unboxed-0.2-inplace-consdef+3tP2tTLTg5qEHov4661Voa/autogen/H.hsig -- most of that is under dist-newstyle |
| 2021-03-02 07:47:16 | <boxscape> | Is the only way to get a type with a kind like `TYPE (TupleRep [LiftedRep, IntRep])` to define a custom type with that kind? |
| 2021-03-02 07:47:19 | <nshepperd2> | so, defining a sum and product datatype for every combination of two unlifted typereps? |
| 2021-03-02 07:47:35 | <edwardk> | :k (# 4, 3# #) |
| 2021-03-02 07:47:37 | <lambdabot> | error: Operator applied to too few arguments: # |
| 2021-03-02 07:47:49 | <mniip> | % :k (# 4, 3# #) |
| 2021-03-02 07:47:49 | <yahb> | mniip: ; <interactive>:1:7: error: parse error on input `3#' |
| 2021-03-02 07:48:05 | <mniip> | % :set -XMagicHash -XUnboxedTuples |
| 2021-03-02 07:48:05 | <yahb> | mniip: |
| 2021-03-02 07:48:06 | <mniip> | % :k (# 4, 3# #) |
| 2021-03-02 07:48:07 | <yahb> | mniip: ; <interactive>:1:7: error: parse error on input `3#' |
| 2021-03-02 07:48:12 | <mniip> | er |
| 2021-03-02 07:48:15 | <mniip> | of course, that's a term |
| 2021-03-02 07:48:24 | <mniip> | % :k (# Int, Int# #) |
| 2021-03-02 07:48:25 | <yahb> | mniip: TYPE ('TupleRep '[ 'LiftedRep, 'GHC.Exts.IntRep]) |
| 2021-03-02 07:48:25 | <edwardk> | anyways that lives in that type |
| 2021-03-02 07:48:31 | <boxscape> | ah, thanks |
| 2021-03-02 07:48:33 | <edwardk> | % :t (# 4, 3# #) |
| 2021-03-02 07:48:33 | <yahb> | edwardk: Num a => (# a, Int# #) |
| 2021-03-02 07:48:39 | <edwardk> | % :t (# 4 :: Int, 3# #) |
| 2021-03-02 07:48:39 | <yahb> | edwardk: (# Int, Int# #) |
| 2021-03-02 07:48:54 | <edwardk> | % :k (# Int, Int# #) |
| 2021-03-02 07:48:54 | <yahb> | edwardk: TYPE ('TupleRep '[ 'LiftedRep, 'GHC.Exts.IntRep]) |
| 2021-03-02 07:48:57 | <edwardk> | thank you =) |
| 2021-03-02 07:49:06 | <boxscape> | -.- right I tried :k forall a . Num a => (# a, Int# #) but of course that's * because it's a function |
| 2021-03-02 07:49:15 | <edwardk> | yeah |
| 2021-03-02 07:49:15 | → | michalz joins (~user@185.246.204.42) |
| 2021-03-02 07:49:20 | <edwardk> | thats the trick that backs my Lev trick |
| 2021-03-02 07:49:31 | → | mananamenos joins (~mananamen@193.red-88-11-66.dynamicip.rima-tde.net) |
| 2021-03-02 07:49:36 | <edwardk> | type Lev (a :: TYPE r) = ()~() => a -- lives in Type even when a lives in TYPE r |
| 2021-03-02 07:49:48 | <boxscape> | Oh, Isee |
| 2021-03-02 07:50:12 | <boxscape> | I saw Iceland_Jack posting that on twitter |
| 2021-03-02 07:50:34 | <mniip> | that does involve boxing a pretty much |
| 2021-03-02 07:50:47 | <mniip> | except not as a data closure |
All times are in UTC.