Logs: freenode/#haskell
| 2020-11-25 20:03:18 | <monochrom> | Let's up the game. System calls are faster if the caller is in ring 0, too. |
| 2020-11-25 20:03:19 | <int-e> | it's true for gettimeofday, and I'm generalizing from there |
| 2020-11-25 20:03:31 | <dminuoso> | I know of at least one syscall that blocks |
| 2020-11-25 20:03:33 | → | skiold joins (~skiold@ns378912.ip-5-196-69.eu) |
| 2020-11-25 20:03:47 | × | skiold quits (~skiold@ns378912.ip-5-196-69.eu) (Client Quit) |
| 2020-11-25 20:03:49 | <int-e> | dminuoso: sleep? |
| 2020-11-25 20:03:53 | <dminuoso> | ioctl |
| 2020-11-25 20:03:59 | → | skiold joins (~skiold@ns378912.ip-5-196-69.eu) |
| 2020-11-25 20:04:10 | <int-e> | (which is probably implemented using select I guess) |
| 2020-11-25 20:04:29 | <dminuoso> | (well ioctl *can* block, depending on which ioctl you invoke) |
| 2020-11-25 20:04:35 | <int-e> | I think the list will be shorter if we list system calls that don't block |
| 2020-11-25 20:04:38 | × | skiold quits (~skiold@ns378912.ip-5-196-69.eu) (Client Quit) |
| 2020-11-25 20:05:31 | × | jonatanb quits (jonatanb@gateway/vpn/protonvpn/jonatanb) (Remote host closed the connection) |
| 2020-11-25 20:06:03 | <frdg> | when you initially compile a program from scratch will stack use more ram then it would when you later make small changes to the application? |
| 2020-11-25 20:06:33 | <dminuoso> | yes. no. perhaps |
| 2020-11-25 20:06:46 | <int-e> | frdg: I don't know. ghc tends to use less RAM when it (re-)compiles fewer modules |
| 2020-11-25 20:06:51 | <monochrom> | Most likely yes. Because the first time it builds aeson, the second time it doesn't, for example. |
| 2020-11-25 20:06:51 | <dminuoso> | frdg: It doesnt really matter at the end. GHC has no parallel module compilatoin |
| 2020-11-25 20:07:19 | <ezzieyguywuf> | is there a generalized "concat", i.e. something that would do `foldl (<>) "" ["A", "B", "C"]` |
| 2020-11-25 20:07:22 | <frdg> | ok thanks |
| 2020-11-25 20:07:43 | <int-e> | dminuoso: Yes it does, because everybody uses ghc --make and GHC has internal caches that are not flushed between module compilations. |
| 2020-11-25 20:07:54 | <int-e> | (everybody except ghc's own build system, incidentally) |
| 2020-11-25 20:08:11 | → | skiold joins (~skiold@gateway/tor-sasl/skiold) |
| 2020-11-25 20:08:12 | <dminuoso> | oh |
| 2020-11-25 20:08:18 | <geekosaur> | mconcat? |
| 2020-11-25 20:08:36 | × | Franciman quits (~francesco@host-82-54-193-143.retail.telecomitalia.it) (Ping timeout: 240 seconds) |
| 2020-11-25 20:08:51 | <geekosaur> | or sconcat if it's a Semigroup |
| 2020-11-25 20:09:18 | <dminuoso> | int-e: well, recently Ive been annoyed by all the caching. There's subtle bugs in either cabal or ghc, I almost weekly have an issue where I have to nuke dist-newstyle before compilation can succeed |
| 2020-11-25 20:10:03 | <merijn> | dminuoso: You must be doing *something* I haven't encountered that in years |
| 2020-11-25 20:10:05 | × | avdb quits (~avdb@ip-81-11-215-86.dsl.scarlet.be) (Quit: avdb) |
| 2020-11-25 20:10:07 | <int-e> | Hmm. I'm still on v1 sandboxes... but I tend to nuke those a lot too. |
| 2020-11-25 20:10:09 | × | wonko7 quits (~wonko7@2a01:e35:2ffb:7040:55f1:c3a3:cdbe:bf52) (Ping timeout: 272 seconds) |
| 2020-11-25 20:10:18 | <merijn> | int-e: ... why? o.O |
| 2020-11-25 20:10:21 | <monochrom> | "Please wash your hands and run 'cabal clean' often." >:) |
| 2020-11-25 20:10:24 | <dminuoso> | merijn: I think its either this particular cabal or ghc version |
| 2020-11-25 20:10:28 | <dminuoso> | or combination |
| 2020-11-25 20:10:30 | <ezzieyguywuf> | geekosaur: I think one of those may be what I'm looking for thanks. |
| 2020-11-25 20:10:33 | <int-e> | merijn: why not? |
| 2020-11-25 20:11:03 | <monochrom> | Heh, I don't push everyone to go v2, I was a bit reluctant too. |
| 2020-11-25 20:11:08 | <int-e> | They work, they have a neat cleanup story, and I'm used to them. |
| 2020-11-25 20:11:17 | <merijn> | because v2-build is so much more robust and reliable >.> It already was years ago >.> |
| 2020-11-25 20:11:18 | <sclv> | i've certainly noticed that when i use nix to swap out entire package envs then the cached stuff goes stale on v1 |
| 2020-11-25 20:11:25 | <int-e> | Well I'll be forced to switch eventually. |
| 2020-11-25 20:11:28 | → | f-a joins (~f-a@151.34.188.246) |
| 2020-11-25 20:11:44 | <sclv> | because the hashing relies on version numbers and the like, so if i'm swapping between different local versions of deps it gets confused |
| 2020-11-25 20:11:46 | <int-e> | But this channel should understand the desire to not do so eagerly :P |
| 2020-11-25 20:12:15 | → | Tario joins (~Tario@200.119.185.214) |
| 2020-11-25 20:12:18 | <int-e> | If you want a rational reason: The later I switch the more well-rounded the v2 commands will be. |
| 2020-11-25 20:12:33 | ← | f-a parts (~f-a@151.34.188.246) () |
| 2020-11-25 20:12:48 | <sclv> | this is for $WORK$ which is pervasively nixified so i stick to v1 |
| 2020-11-25 20:13:02 | <sclv> | for personal dev i use v2 and its pretty damn seamless at this point |
| 2020-11-25 20:13:07 | <merijn> | I switched to v2-build in 2016 or something, breaking all my editor tooling (like hdevtools in the progress) because it was less painful than fighting v1 >.> |
| 2020-11-25 20:13:10 | <sclv> | given that i sort of know what stuff doesn't work |
| 2020-11-25 20:13:38 | → | treotmnor joins (~treotmnor@8.210.183.149) |
| 2020-11-25 20:13:55 | <int-e> | merijn: that's the thing though... ever since I switched to using sandboxes I never had to fight |
| 2020-11-25 20:14:26 | <merijn> | int-e: I had to reinit sandboxes *a lot*, though |
| 2020-11-25 20:14:31 | <merijn> | Because that's what I used before |
| 2020-11-25 20:14:36 | <int-e> | sure. |
| 2020-11-25 20:14:37 | → | alp joins (~alp@2a01:e0a:58b:4920:3dd7:733:8dcb:bbbc) |
| 2020-11-25 20:14:44 | <int-e> | I have a shell script for that |
| 2020-11-25 20:14:53 | → | Ariakenom__ joins (~Ariakenom@h-98-128-229-104.NA.cust.bahnhof.se) |
| 2020-11-25 20:14:53 | <merijn> | If dependencies are in flux nuking and rebuilding the world every time is a *massive* time sync |
| 2020-11-25 20:15:06 | <geekosaur> | "sink" |
| 2020-11-25 20:15:08 | <int-e> | sink |
| 2020-11-25 20:15:29 | <merijn> | I got stuff to do that doesn't involve waiting for the kmettiverse to compile for the 15th time |
| 2020-11-25 20:15:43 | <int-e> | merijn: also, honestly, I don't really have big projects |
| 2020-11-25 20:15:53 | <merijn> | geekosaur: Fixing my brain to no output a random homophone when I type is a lost cause :p |
| 2020-11-25 20:15:56 | <monochrom> | If dependencies are in flux, v2 rebuilds enough of the world to be a time sink, too. |
| 2020-11-25 20:16:07 | <merijn> | monochrom: Depends of how much |
| 2020-11-25 20:16:16 | <int-e> | make all the dependencies that are in flux source dependencies :P |
| 2020-11-25 20:16:21 | <merijn> | monochrom: Like, adding 1 or 2 packages won't rebuild the entire kmettiverse :p |
| 2020-11-25 20:16:41 | <koz_> | merijn: Lol@'kmettiverse'. The Kmett Extended Universe. |
| 2020-11-25 20:16:52 | <merijn> | Actually |
| 2020-11-25 20:17:10 | <monochrom> | Well I should get out of this debate. |
| 2020-11-25 20:17:12 | <merijn> | Maybe we should start calling it the KCU Kmett Cinematic Universe or something :p |
| 2020-11-25 20:17:21 | <koz_> | I'm down for KCU. |
| 2020-11-25 20:17:31 | <koz_> | Avengers: Rise of the Profunctor. |
| 2020-11-25 20:17:56 | <monochrom> | I was going to say, the disease of the lot of you is debating until other people do things your way. This is already the second time within an hour, even half an hour. |
| 2020-11-25 20:18:05 | <monochrom> | Today is compiling day and bikeshedding day. |
| 2020-11-25 20:18:20 | × | LKoen quits (~LKoen@169.244.88.92.rev.sfr.net) (Ping timeout: 260 seconds) |
| 2020-11-25 20:18:34 | <merijn> | monochrom: It *is* compiling day, that's why I have nothing better to do :p |
| 2020-11-25 20:18:36 | <monochrom> | Kmett Categorical Universe |
| 2020-11-25 20:18:43 | <merijn> | monochrom: Rats |
| 2020-11-25 20:18:48 | <monochrom> | Oh haha good cause |
| 2020-11-25 20:18:48 | × | Ariakenom_ quits (~Ariakenom@h-158-174-186-195.NA.cust.bahnhof.se) (Ping timeout: 260 seconds) |
| 2020-11-25 20:19:03 | <merijn> | monochrom: I spent 5 minutes pondering better a better word for C and somehow completely missed Category... |
| 2020-11-25 20:19:33 | geekosaur | resembles that too (although now the compiling is mostly done but my poor little netbook is still recovering from it) |
| 2020-11-25 20:19:50 | <koz_> | Instead of GHC2021, we should have 'KCU' as a language option. :P |
| 2020-11-25 20:19:51 | <merijn> | monochrom: I'm messing with the database schema of my code, which is waaaaaaaay at the bottom of my module hierarchy, so it triggers nearly full rebuild that takes like 10 minutes each time >.> |
| 2020-11-25 20:20:03 | <monochrom> | You are in front of the guy who thought up how to expand "tl;dr" to "type-level instance diversification and resolution", "sicp" to "structure and identification of cabalized packages". You really stand no chance. :) |
| 2020-11-25 20:20:08 | <merijn> | monochrom: So lots of time to spread the gospel :p |
| 2020-11-25 20:20:26 | <koz_> | monochrom: Wordplay is how you flex in #haskell (and the ecosystem). |
| 2020-11-25 20:20:31 | → | Fractalis joins (~Fractalis@c-174-54-165-158.hsd1.pa.comcast.net) |
| 2020-11-25 20:20:50 | → | kritzefitz joins (~kritzefit@212.86.56.80) |
| 2020-11-25 20:21:16 | <tomjaguarpaw> | Yes, we've already maxed out on pointfree flexing |
| 2020-11-25 20:21:24 | <monochrom> | Hey, Kmett Compiling Universe is appropriate, too. |
| 2020-11-25 20:21:47 | <koz_> | Lol. |
| 2020-11-25 20:22:17 | <koz_> | Steps to putting a package on Hackage: 1. Write it; 2. Agonize for hours over suitable worlplay-ey name. |
| 2020-11-25 20:22:18 | × | kelsey quits (~keteskyl@2600:6c64:7b7f:fa42:e18f:e78b:7854:f771) (Quit: WeeChat 2.8) |
| 2020-11-25 20:22:39 | → | kelsey joins (~keteskyl@2600:6c64:7b7f:fa42:e18f:e78b:7854:f771) |
All times are in UTC.