Logs: freenode/#haskell
| 2021-03-16 17:06:25 | → | polxy joins (~polxy@2001:b07:a15:ec0c:14d1:ef73:1822:ceed) |
| 2021-03-16 17:06:27 | <koz_> | merijn: What's the flag for turning off parallel GC again? |
| 2021-03-16 17:08:40 | <merijn> | -qg |
| 2021-03-16 17:08:49 | <koz_> | Thanks! |
| 2021-03-16 17:09:05 | <merijn> | or -gq one of the two :p |
| 2021-03-16 17:09:20 | <geekosaur> | -qg |
| 2021-03-16 17:09:24 | <curiousgay> | parallel GC - slower, but on by default |
| 2021-03-16 17:09:34 | <geekosaur> | for the moment |
| 2021-03-16 17:09:50 | <geekosaur> | it's getting turned off by default in an upcoming ghc release |
| 2021-03-16 17:10:09 | × | polxy quits (~polxy@2001:b07:a15:ec0c:14d1:ef73:1822:ceed) (Client Quit) |
| 2021-03-16 17:10:16 | <curiousgay> | btw, I was taking a look at haddock's dependencies |
| 2021-03-16 17:10:34 | <curiousgay> | it always depends on specific GHC x series |
| 2021-03-16 17:10:38 | × | mouseghost quits (~draco@wikipedia/desperek) (Quit: mew wew) |
| 2021-03-16 17:10:40 | <merijn> | Well, yes |
| 2021-03-16 17:11:09 | <curiousgay> | does that mean GHC frequently breaks compatibility? |
| 2021-03-16 17:11:13 | <merijn> | It uses GHC to parse code, how could it not |
| 2021-03-16 17:11:17 | <koz_> | geekosaur: Is that change gonna get backported? |
| 2021-03-16 17:11:30 | <merijn> | curiousgay: Define "compatibility"\ |
| 2021-03-16 17:12:50 | → | conal joins (~conal@64.71.133.70) |
| 2021-03-16 17:13:37 | <curiousgay> | merijn: haddock 2.25.0 strictly depends on GHC 9.0.x, so it won't build with GHC 9.1.x, haddock 2.24.x depend on GHC 8.10.x and won't build on any GHC 9 |
| 2021-03-16 17:13:38 | → | Lycurgus joins (~niemand@98.4.116.165) |
| 2021-03-16 17:13:42 | × | evanjs quits (~evanjs@075-129-098-007.res.spectrum.com) (Quit: ZNC 1.8.2 - https://znc.in) |
| 2021-03-16 17:13:49 | <curiousgay> | s/on/with/ |
| 2021-03-16 17:13:56 | × | banjiewen__ quits (sid115913@gateway/web/irccloud.com/x-rsmfwpptfrqcdppn) () |
| 2021-03-16 17:13:59 | → | evanjs joins (~evanjs@075-129-098-007.res.spectrum.com) |
| 2021-03-16 17:14:12 | → | b20n joins (sid115913@gateway/web/irccloud.com/x-lczkliukjfykdgvl) |
| 2021-03-16 17:14:19 | → | Morrow_ joins (~MorrowM_@147.161.13.246) |
| 2021-03-16 17:14:40 | <merijn> | curiousgay: Because haddock uses GHC as library and therefore depends on GHC internals |
| 2021-03-16 17:14:45 | → | Sorna joins (~Sornaensi@79.142.232.102.static.router4.bolignet.dk) |
| 2021-03-16 17:15:10 | × | Sornaensis quits (~Sornaensi@79.142.232.102.static.router4.bolignet.dk) (Read error: Connection reset by peer) |
| 2021-03-16 17:15:40 | → | MorrowM joins (~MorrowM_@147.161.13.246) |
| 2021-03-16 17:15:54 | × | jlamothe quits (~jlamothe@198.251.55.207) (Ping timeout: 246 seconds) |
| 2021-03-16 17:17:42 | <geekosaur> | sorry, I don't know |
| 2021-03-16 17:18:08 | → | marinelli joins (~marinelli@gateway/tor-sasl/marinelli) |
| 2021-03-16 17:18:31 | × | royal_screwup21 quits (52254809@gateway/web/cgi-irc/kiwiirc.com/ip.82.37.72.9) (Quit: Connection closed) |
| 2021-03-16 17:18:51 | → | royal_screwup21 joins (52254809@gateway/web/cgi-irc/kiwiirc.com/ip.82.37.72.9) |
| 2021-03-16 17:19:11 | × | stree quits (~stree@68.36.8.116) (Quit: Caught exception) |
| 2021-03-16 17:19:38 | → | stree joins (~stree@68.36.8.116) |
| 2021-03-16 17:19:44 | <monochrom> | haddock comes with GHC anyway. You are not supposed to build haddock yourself. (Apart from building GHC yourself.) |
| 2021-03-16 17:20:09 | × | nhs_ quits (~nhs@70.113.67.118) (Read error: Connection reset by peer) |
| 2021-03-16 17:20:45 | × | cfricke quits (~cfricke@unaffiliated/cfricke) (Quit: WeeChat 3.0.1) |
| 2021-03-16 17:22:51 | → | jlamothe joins (~jlamothe@198.251.55.207) |
| 2021-03-16 17:23:50 | × | royal_screwup21 quits (52254809@gateway/web/cgi-irc/kiwiirc.com/ip.82.37.72.9) (Ping timeout: 265 seconds) |
| 2021-03-16 17:25:36 | → | Sorny joins (~Sornaensi@077213203030.dynamic.telenor.dk) |
| 2021-03-16 17:25:52 | <tomsmeding> | (which you're also not supposed to do if you're not hacking on GHC) |
| 2021-03-16 17:26:41 | <dolio> | I build every version of GHC from source, but I don't hack on it. |
| 2021-03-16 17:26:59 | <tomsmeding> | okay sure, and also unless you want it for a specific reason that most people don't have :p |
| 2021-03-16 17:27:38 | <tomsmeding> | dolio: for my curiosity, why do you build ghc yourself? Custom optimisation flags? |
| 2021-03-16 17:27:54 | <merijn> | masochism :p |
| 2021-03-16 17:28:01 | × | rj_ quits (~x@gateway/tor-sasl/rj) (Ping timeout: 268 seconds) |
| 2021-03-16 17:28:03 | landonf_ | is now known as landonf |
| 2021-03-16 17:28:19 | <dolio> | No, I have many versions of GHC installed in /opt/ghc. |
| 2021-03-16 17:28:40 | <merijn> | dolio: That seems unrelated to building GHC from source, though? |
| 2021-03-16 17:28:48 | × | Sorna quits (~Sornaensi@79.142.232.102.static.router4.bolignet.dk) (Ping timeout: 245 seconds) |
| 2021-03-16 17:28:56 | <dolio> | Well, I use the previous version to build the next one. |
| 2021-03-16 17:29:20 | <tomsmeding> | to help prevent the Thompson problem? |
| 2021-03-16 17:30:29 | → | rj_ joins (~x@gateway/tor-sasl/rj) |
| 2021-03-16 17:30:31 | <dolio> | I guess another reason is that it ensures they're built against the right library versions on my system, although it's uncommon for that to be a problem. |
| 2021-03-16 17:30:42 | <dolio> | I think it was a problem around the time I started. |
| 2021-03-16 17:32:17 | <dolio> | I needed to use the Fedora-distributed ghc to begin the process, because the ghc homepage one was built on Debian or something, and they disagreed on glibc or something. |
| 2021-03-16 17:32:55 | <monochrom> | I did that for a little while, around the time of 7.6. It was the infamous libgmp versioning problem. |
| 2021-03-16 17:33:11 | <dolio> | Oh, maybe that's what itw as. |
| 2021-03-16 17:33:43 | <dolio> | 7.6 is the earliest version I have, I think. |
| 2021-03-16 17:34:07 | → | electricityZZZZ joins (~electrici@108-216-157-17.lightspeed.sntcca.sbcglobal.net) |
| 2021-03-16 17:34:35 | × | Lycurgus quits (~niemand@98.4.116.165) (Quit: Exeunt) |
| 2021-03-16 17:35:16 | <monochrom> | http://www.vex.net/~trebla/haskell/I-built-GHC.xhtml |
| 2021-03-16 17:35:29 | <monochrom> | "Give a man a fire, and he is warm for a day. Show him how to build GHC, and he is warm for life." |
| 2021-03-16 17:35:31 | <edwardk> | i was staring at https://hackage.haskell.org/package/recover-rtti and wishing i could register custom handlers for my own types. then i realized there might be a way to make an extensible version of that scheme, but its super baroque. |
| 2021-03-16 17:35:47 | <LKoen> | are you saying GHC will overheat the computer? |
| 2021-03-16 17:36:56 | <dolio> | Anyhow, it doesn't take a significant amount of effort at this point. Just time. |
| 2021-03-16 17:37:03 | <edwardk> | haskell doesn't really give me hooks that let me in separate files tell a function it should also consider another case.. but c++ does. i can register a top level definition of a class and have it hook itself on a global list of such objects in its static initializer. libraries that parse command lines, etc. tend to do this in c++ all the time. |
| 2021-03-16 17:37:07 | × | Yumasi quits (~guillaume@2a01:e0a:5cb:4430:17e3:41cc:739d:497d) (Ping timeout: 260 seconds) |
| 2021-03-16 17:37:24 | <edwardk> | and we have inline-c-cpp which can use template haskell to spew out a bit of c++ and link it into the program you are writing |
| 2021-03-16 17:38:07 | <dolio> | I think the biggest snag I've hit was that I unrelatedly installed some of the dependencies needed to build the docs on the latest version, so GHC thought it should try to build them, but wasn't able to. |
| 2021-03-16 17:38:18 | <edwardk> | so in theory one could install a c++-side list of all the handlers you want, then haskell side use a template haskell splice to register a new handler for a new type, and then extend the scheme offered by something like recover-rtti to new types you made up after the library was coined. |
| 2021-03-16 17:39:11 | <edwardk> | the main difference would be that classifier would become a data family and some other things like that would change, but overall i think you could keep the feel of the library |
| 2021-03-16 17:39:36 | → | royal_screwup21 joins (52254809@gateway/web/cgi-irc/kiwiirc.com/ip.82.37.72.9) |
| 2021-03-16 17:47:37 | → | seven_three joins (~user@2601:18d:c180:4060::2829) |
| 2021-03-16 17:48:30 | × | evanjs quits (~evanjs@075-129-098-007.res.spectrum.com) (Quit: ZNC 1.8.2 - https://znc.in) |
| 2021-03-16 17:48:49 | → | shad0w_ joins (67573b43@103.87.59.67) |
| 2021-03-16 17:49:38 | × | average quits (uid473595@gateway/web/irccloud.com/x-ywsfnraofofbenfj) (Quit: Connection closed for inactivity) |
| 2021-03-16 17:50:19 | → | evanjs joins (~evanjs@075-129-098-007.res.spectrum.com) |
| 2021-03-16 17:50:41 | <hyiltiz> | What's the lens for [[Maybe Bool]], i.e. I'd like to view, (also set and over) the Bool nested inside the Maybe inside the list inside a list, and |
| 2021-03-16 17:50:41 | <seven_three> | would you say that parser combinators have an advantage over regexps in terms of creating recursive expression parsers? For example data with a structure like `(4 + (4 + (4 * (2 + 1))))`? |
| 2021-03-16 17:50:50 | × | denisse quits (~spaceCat@gateway/tor-sasl/alephzer0) (Ping timeout: 268 seconds) |
| 2021-03-16 17:51:27 | → | denisse joins (~spaceCat@gateway/tor-sasl/alephzer0) |
| 2021-03-16 17:51:27 | × | conal quits (~conal@64.71.133.70) (Read error: Connection reset by peer) |
| 2021-03-16 17:51:53 | <hyiltiz> | I could do `fromJust $ x !! m !! n` but returning the whole stuff with only a single element changed (aka set) is more clunky |
| 2021-03-16 17:52:03 | <dolio> | Yes, regular expressions are only suitable for very simple languages. |
| 2021-03-16 17:52:53 | hackage | haskoin-store 0.50.2 - Storage and index for Bitcoin and Bitcoin Cash https://hackage.haskell.org/package/haskoin-store-0.50.2 (jprupp) |
| 2021-03-16 17:52:58 | <dolio> | Like, tokenizing at most, really. |
| 2021-03-16 17:54:22 | <edmundnoble> | Yeah regular expressions are really not about "nesting", they're more "linear" |
| 2021-03-16 17:54:29 | <seven_three> | dolio: Yes I am finding it unnatural currently |
| 2021-03-16 17:54:48 | <dminuoso> | % [[Just True, Just False], [Just True]] ^.. each . each . _Just -- hyiltiz |
| 2021-03-16 17:54:48 | <yahb> | dminuoso: [True,False,True] |
| 2021-03-16 17:56:38 | <seven_three> | I see how that works. You could quickly get into making really big parsers that you can just call like that. And then easily use those blocks to make a grammar. |
| 2021-03-16 17:57:46 | → | conal joins (~conal@64.71.133.70) |
| 2021-03-16 17:58:05 | <hyiltiz> | domenkozar[m]: thx! Now can I define ^..each.each._Just myself (it is fine if it only works for [[Maybe Bool]] so I do not have to pull in a dependency? |
| 2021-03-16 17:58:12 | × | merijn quits (~merijn@83-160-49-249.ip.xs4all.nl) (Ping timeout: 256 seconds) |
| 2021-03-16 17:58:29 | <seven_three> | Thats what I am trying to do now with my regexs but it is more about program structure then the regexs themselves |
All times are in UTC.