| 00:27:57 | NimBot | Araq/Nimrod a8ce01d Billingsly Wetherfordshire [+0 ±1 -0]: rename json.existsKey to hasKey, deprecates existsKey |
| 00:27:57 | NimBot | Araq/Nimrod 1d32ea1 Araq [+0 ±1 -0]: Merge pull request #455 from fowlmouth/patch-7... 3 more lines |
| 00:29:57 | NimBot | Araq/Nimrod 19529a6 Araq [+0 ±4 -0]: better error message for subtle nil errors |
| 00:29:57 | NimBot | Araq/Nimrod a2610fe Araq [+0 ±1 -0]: another attempt to implement #448 |
| 00:29:57 | NimBot | Araq/Nimrod dc2b3eb Araq [+0 ±2 -0]: fixes #458 |
| 00:29:57 | NimBot | Araq/Nimrod 55c4c0a Araq [+0 ±1 -0]: Merge branch 'master' of github.com:Araq/Nimrod |
| 00:31:18 | * | SirSkidmore quit (Ping timeout: 264 seconds) |
| 01:50:08 | * | q66 quit (Remote host closed the connection) |
| 01:50:11 | reactormonk | what's nimRTL? |
| 01:56:21 | fowl | rtl = runtime library |
| 02:00:49 | reactormonk | kk |
| 02:07:54 | * | GuidovanPossum joined #nimrod |
| 02:09:41 | * | GuidovanPossum quit (Client Quit) |
| 02:10:09 | * | GuidovanPossum joined #nimrod |
| 02:20:21 | * | GuidovanPossum quit (Quit: GuidovanPossum) |
| 02:21:35 | * | GuidovanPossum joined #nimrod |
| 02:23:01 | * | GuidovanPossum quit (Client Quit) |
| 02:32:10 | tangentstorm | we got the brainf*** thing working. even wrote a little visual debugger for it. :) |
| 02:32:16 | tangentstorm | https://github.com/tangentstorm/grimspike/blob/master/bfrun.nim |
| 02:51:47 | * | Associ8or joined #nimrod |
| 02:54:02 | * | SchalaZeal joined #nimrod |
| 02:54:20 | SchalaZeal | Getting a very weird error from the compiler. |
| 02:57:05 | SchalaZeal | http://pastebin.com/tbhh0s8v |
| 02:57:38 | SchalaZeal | According to the compiler, TNamedBinaryTag is marked as final when it isn't. |
| 02:58:02 | SchalaZeal | getting "Error: inheritance only works with non-final objects" on every inheriting object declaration line |
| 03:00:35 | fowl | hey SchalaZeal |
| 03:00:41 | SchalaZeal | hey |
| 03:00:57 | SchalaZeal | I made an AUR package for that nifty nimlibs project of yours. |
| 03:01:26 | fowl | object's aren't inheritable by default, so tnamedbinarytag needs to be object of TObject or {.inheritable.} |
| 03:01:41 | fowl | no need, its in babel |
| 03:01:44 | SchalaZeal | then why have {.final.}? |
| 03:02:14 | fowl | im not sure :D |
| 03:02:21 | fowl | i think the behavior used to be different |
| 03:02:30 | SchalaZeal | seems like more work for me now |
| 03:02:35 | SchalaZeal | it did |
| 03:02:49 | SchalaZeal | default was inheritable unless {.final.} was specified |
| 03:03:17 | SchalaZeal | seems annoying to reverse it |
| 03:03:56 | fowl | its been this way for a while now |
| 03:04:14 | SchalaZeal | well I haven't used Nimrod much in said while |
| 03:04:39 | SchalaZeal | trying to make a modular Minecraft server in it |
| 03:04:47 | fowl | cool |
| 03:05:33 | SchalaZeal | the .dat files use the "named binary tag" format and are g-zipped, so I'm building it according to the specs for the format |
| 03:06:51 | SchalaZeal | had to go out of my way to make floating point endian conversion procs though |
| 03:07:25 | fowl | you cant use the endians module? |
| 03:07:34 | SchalaZeal | it only does integers |
| 03:07:37 | SchalaZeal | not floats |
| 03:08:21 | SchalaZeal | for Minecraft data storage I need to work with integer and floats in big endian |
| 03:08:32 | SchalaZeal | and strings in UTF8 |
| 03:08:33 | fowl | yea |
| 03:09:10 | fowl | this seems to work: https://gist.github.com/fowlmouth/6206634695a55e223b80 |
| 03:10:03 | SchalaZeal | ah |
| 03:10:15 | SchalaZeal | *scraps the silly nbcommon module I have* |
| 03:10:45 | fowl | check this out i wrote it to read/write in bigendian https://github.com/fowlmouth/nimrod-enet/blob/master/enet_pkt_utils.nim |
| 03:11:46 | SchalaZeal | I can nail the NBT format, provided there's gzip capability in Nimrod's libraries. The region files are going to be a tad tricky though |
| 03:14:23 | SchalaZeal | is there any pragmas for bit/byte packing and also for abstract objects? |
| 03:14:42 | SchalaZeal | {.align: #.} no longer exists |
| 03:15:24 | fowl | not to my knowledge |
| 03:15:30 | fowl | what do you mean by abstract object |
| 03:16:05 | SchalaZeal | an object that shouldn't be instantiated, that should be inherited by a derived class |
| 03:16:48 | SchalaZeal | similar to abstract classes in C#/Java or virtual foo() = 0; in C++ |
| 03:20:02 | fowl | no you could make such an abstract class but you cant keep it from being instantiated |
| 03:20:22 | fowl | er abstract type |
| 03:20:41 | SchalaZeal | eh well...should anyone reuse my code, they're likely to get shot in the foot should they use it like that |
| 03:23:31 | SchalaZeal | byte/bit packed objects/tuples are a necessity though I think |
| 03:25:31 | fowl | you'd have to ask araq about that |
| 03:25:38 | SchalaZeal | hmm ok |
| 03:25:41 | fowl | what you been up to lately i havent seen you around |
| 03:25:56 | SchalaZeal | Minecraft, and seeking a career in 3D modeling |
| 03:26:26 | fowl | cool |
| 03:26:37 | SchalaZeal | yep |
| 03:27:45 | SchalaZeal | I guess I made an AUR package because I'm OCD about having installable packages of everything rather than rearranging files myself |
| 03:28:17 | fowl | you should make one for babel if there isnt already |
| 03:28:36 | SchalaZeal | I did but orphaned it when I personally didn't want it anymore |
| 03:30:32 | SchalaZeal | feel free to grab it, babel-git |
| 03:30:57 | fowl | im not on arch |
| 03:31:19 | fowl | it probably doesnt ever need to be updated, the install should just be git pull && nimrod c babel && mv babel /to/bin |
| 03:31:40 | SchalaZeal | I prefer install instead of mv |
| 03:31:49 | SchalaZeal | mv doesn't set the access |
| 03:32:04 | fowl | ah |
| 03:32:28 | SchalaZeal | install -Dm755 babel $pkgdir/usr/bin/babel |
| 03:33:41 | SchalaZeal | hmm... do generics have constraints? |
| 03:33:53 | reactormonk | SchalaZeal, sure you can have them, T: string|int |
| 03:34:19 | SchalaZeal | what about like T: TFoo being TFoo and descendants? |
| 03:34:38 | reactormonk | there is no covariance iirc |
| 03:34:48 | reactormonk | Araq, how's covariance? |
| 03:38:33 | SchalaZeal | I have this so far |
| 03:38:40 | SchalaZeal | proc `data`[T: TNamedBinaryTag]*(p: var TListTag[T]): seq[ref T]{.inline.} = |
| 03:38:42 | SchalaZeal | return p.fArray |
| 03:38:49 | reactormonk | SchalaZeal, covariance rings a bell? |
| 03:39:03 | SchalaZeal | with C++ I suppose |
| 03:39:18 | reactormonk | kk |
| 03:40:43 | SchalaZeal | template<Foo T> class<T> { }; T would be Foo or its descendants |
| 03:41:23 | reactormonk | not sure about covariance in nimrod, you could try |
| 03:46:13 | SchalaZeal | it compiles |
| 03:47:41 | SchalaZeal | and it runs! |
| 03:47:50 | SchalaZeal | covariance is indeed in Nimrod |
| 03:47:59 | fowl | doesnt work for me: http://pastebin.com/m95YcG3Z |
| 03:48:31 | SchalaZeal | mine: http://pastebin.com/BFu60pjV |
| 03:53:04 | SchalaZeal | ........hmm |
| 03:53:10 | SchalaZeal | this isn't right |
| 03:53:58 | fowl | so that part seems to work but no covariance for matching generic procs |
| 03:54:00 | SchalaZeal | tried to cause a compiler error by chancing b: TBlah[TFoo2] to TBlah[int] ...... still successful |
| 03:54:14 | fowl | ah |
| 03:54:54 | SchalaZeal | ah... but it fails to run |
| 03:55:15 | SchalaZeal | hmm actually... |
| 03:56:02 | SchalaZeal | yeah... getting a segfault if it's any type other than TFoo |
| 03:56:10 | SchalaZeal | or derived classes |
| 03:56:16 | SchalaZeal | instead of a compiler error |
| 04:02:59 | * | SchalaZeal quit (Quit: Konversation terminated!) |
| 04:05:56 | * | ___ joined #nimrod |
| 04:10:16 | * | OrionPK quit (Read error: Connection reset by peer) |
| 04:12:04 | * | ___ quit (Quit: Page closed) |
| 04:55:37 | * | ARCADIVS quit (Quit: WeeChat 0.3.8) |
| 05:35:48 | * | Associ8or quit (Read error: Connection reset by peer) |
| 06:52:26 | * | zahary_ joined #nimrod |
| 06:54:12 | zahary_ | dom96, nice work on the webframeworks benchmark, but why are you putting nginx in front of jester? just go straight to the metal like the onion guys: |
| 06:54:13 | zahary_ | https://github.com/TechEmpower/FrameworkBenchmarks/tree/master/onion |
| 06:54:25 | zahary_ | onion is a C framewok |
| 07:13:41 | * | fowl quit (Ping timeout: 248 seconds) |
| 07:17:42 | * | fowl joined #nimrod |
| 07:18:22 | * | fowl quit (Read error: Connection reset by peer) |
| 07:32:15 | * | comex` joined #nimrod |
| 07:34:58 | * | comex quit (*.net *.split) |
| 08:15:22 | * | fowl joined #nimrod |
| 08:21:58 | * | fowl quit (Read error: Connection reset by peer) |
| 08:22:02 | * | Endeg quit (Read error: Connection reset by peer) |
| 08:22:46 | * | fowl joined #nimrod |
| 08:23:08 | * | fowl quit (Read error: Connection reset by peer) |
| 08:24:08 | * | fowl joined #nimrod |
| 08:24:23 | * | fowl quit (Read error: Connection reset by peer) |
| 08:25:14 | * | fowl joined #nimrod |
| 08:27:58 | * | fowl quit (Read error: Connection reset by peer) |
| 08:28:37 | * | fowl joined #nimrod |
| 08:28:43 | * | fowl quit (Read error: Connection reset by peer) |
| 08:29:27 | * | Endeg joined #nimrod |
| 08:32:11 | * | fowl joined #nimrod |
| 08:32:12 | * | fowl quit (Read error: Connection reset by peer) |
| 09:17:00 | Araq | zahary_: actually I prefer it this way, nginx has optimized static file serving plus lots of other features, so it's a more realistic scenario this way |
| 09:17:46 | Araq | eventually we'll have all these features too but until then we use nginx |
| 09:17:52 | zahary_ | I would put nginx in production in front of nimrod to deal with security risks too, but this is benchmark |
| 09:18:11 | zahary_ | and others may be getting unfair advantage |
| 09:19:26 | Araq | I haven't looked at the details but yeah |
| 09:19:54 | Araq | fair benchmarks would put nginx in front of everything :-) |
| 09:49:47 | * | q66 joined #nimrod |
| 10:07:25 | * | q66 quit (Ping timeout: 240 seconds) |
| 10:07:32 | * | q66 joined #nimrod |
| 10:49:14 | dom96 | zahary_: Our httpserver is still not optimized, scgi is. We would lose, badly, if we went bare metal. httpserver would just lock up with the amount of requests that they put it through. |
| 10:49:43 | zahary_ | hmm, scgi does open/close connection on each request, right? |
| 10:50:15 | dom96 | yes |
| 10:50:28 | zahary_ | how come it's more optimal than HTTP? you mean the HTTP request parsing business will be slower? |
| 10:51:34 | dom96 | It's by no means more optimal. What I mean is that before 0.9.2 I optimized the scgi module, but I did not have enough time to do so for the httpserver. The current async implementation of httpserver isn't async enough, it still uses blocking sockets. |
| 10:52:02 | zahary_ | I see, so you have implemented an epoll loop somewhere? |
| 10:52:15 | dom96 | lol, not even. |
| 10:52:28 | dom96 | That's another optimization we could do. |
| 10:53:16 | dom96 | but results were promising even with select. |
| 10:54:37 | dom96 | The method in which I made scgi fully async is a bit ugly, and hopefully the 'await' stuff will help a lot: https://github.com/Araq/Nimrod/blob/master/lib/pure/scgi.nim#L179 |
| 10:55:24 | NimBot | Araq/Nimrod c0b1c07 Araq [+0 ±2 -0]: more accurate effects for times.nim; bugfix for the JS codegen |
| 10:58:26 | zahary_ | hmm, how does recvAsync work actually? I see the results are ready on the next line. where is the async part? |
| 10:59:48 | dom96 | zahary_: if you look closely you can see that I check its return value. `-1` means we got no data, so we simply return and the callback will be called again when data is available. |
| 11:05:55 | Araq | dom96: we need a "diff" like feature for nimbuild |
| 11:05:57 | dom96 | diff for what? |
| 11:05:58 | Araq | manually comparing the test results sucks |
| 11:05:58 | dom96 | the testresults? |
| 11:05:58 | dom96 | yeah, sure. |
| 11:06:12 | Araq | in fact, the test results suck too |
| 11:06:19 | Araq | I mean their presentation |
| 11:06:23 | dom96 | would perhaps be better if the test results were generated as JSON |
| 11:06:32 | dom96 | Then nimbuild could do something with them |
| 11:06:33 | Araq | it should be grouped by topic |
| 11:06:41 | dom96 | As well as NimBot |
| 11:07:41 | Araq | well things would be different if we were close to 100% green tests |
| 11:07:49 | Araq | but we aren't |
| 11:09:24 | Araq | and manually changing the tests to "ignore" doesn't work either, too laborsome |
| 11:09:42 | Araq | plus ignored tests stay on "Ignore" forever |
| 11:16:58 | dom96 | why is that relevant? |
| 11:19:12 | dom96 | what does this have to do with this 'diff' feature? |
| 11:19:35 | Araq | as I said, the diff feature wouldn't be necessary if we always had 100% green/yellow tests |
| 11:19:46 | Araq | you can easily spot the new red tests then |
| 11:19:53 | dom96 | i see |
| 11:21:35 | zahary_ | dom96, redis contains one of the smallest epoll/kqueue implementations. if we just convert/port the code to nimrod, it could serve as a good starting point: |
| 11:21:45 | zahary_ | https://github.com/antirez/redis/blob/unstable/src/ae.c |
| 11:21:46 | zahary_ | https://github.com/antirez/redis/blob/unstable/src/ae_epoll.c |
| 11:22:06 | zahary_ | https://github.com/antirez/redis/blob/unstable/src/ae_kqueue.c |
| 11:22:40 | zahary_ | here is an example usage from wrk: the server testing tool used by the benchmark guys |
| 11:22:41 | zahary_ | https://github.com/wg/wrk/blob/master/src/wrk.c |
| 11:25:48 | dom96 | interesting. |
| 11:28:11 | dom96 | indeed, looks like a good place to start. However I must admit my C is a bit rusty. |
| 11:32:55 | zahary_ | was c2nim supposed to be able to convert whole C files btw? |
| 11:46:37 | NimBot | Araq/Nimrod 03c9a02 Dominik Picheta [+0 ±1 -0]: Added unittest module to docgen. |
| 11:48:12 | Araq | indeed c2nim can convert whole C files (function implementations/control flow) |
| 11:48:46 | Araq | needs quite some manual tweaking afterwards of course |
| 11:50:06 | Araq | bbl |
| 15:12:10 | * | erlnoob joined #nimrod |
| 15:13:03 | erlnoob | nice site redesign :) |
| 15:13:43 | dom96 | hello erlnoob |
| 15:13:48 | dom96 | indeed it is :) |
| 15:13:56 | erlnoob | hello |
| 15:14:31 | erlnoob | and 0.9.2 is out! |
| 15:15:13 | erlnoob | but nimrod definitely needs more marketing |
| 15:16:07 | dom96 | erlnoob: What else do you think we could do to make it more popular? |
| 15:16:50 | erlnoob | hmm not sure.. needs lots of link-bait blog posts to HN hehehe |
| 15:17:38 | dom96 | yeah, that's what I think also. Problem is getting people to write those blog posts. Trust me I myself am trying to write some. |
| 15:17:45 | erlnoob | great language little known, but maybe it's a blessing! |
| 15:18:29 | erlnoob | dom96: care to share a link to your blog? |
| 15:18:47 | dom96 | It's still a work in progress so it's not up yet. |
| 15:21:10 | dom96 | The URL will probably be: http://picheta.me/blog/ |
| 15:21:23 | erlnoob | cool thanks |
| 15:22:20 | erlnoob | will keep the url for future :) |
| 15:22:27 | dom96 | cool :) |
| 15:29:00 | * | Endy joined #nimrod |
| 15:31:51 | * | [1]Endy joined #nimrod |
| 15:34:49 | * | Endy quit (Ping timeout: 276 seconds) |
| 15:37:13 | * | _ponce quit (Ping timeout: 246 seconds) |
| 15:56:41 | * | fowl joined #nimrod |
| 16:50:37 | Araq | erlnoob: we still have 'nil' in the language which is a "deal breaker" |
| 16:51:31 | Araq | never mind that all it accomplishes for complex code is making me write "if n.typ == nil: InternalError(n.info, "shouldn't happen") " in the compiler everywhere |
| 17:02:08 | * | erlnoob quit (Quit: erlnoob) |
| 17:37:56 | fowl | am i the only one that believes that nil is a value |
| 17:38:18 | fowl | even if it is the value of lacking value, its still a value |
| 17:42:56 | Araq | the difference between 0 and nil is that even if 0 is an invalid value, it doesn't cause a crash as often as nil as pointer deref is way more common than integer division |
| 17:58:48 | * | erlnoob joined #nimrod |
| 18:00:04 | * | erlnoob quit (Client Quit) |
| 18:12:18 | fowl | i dont see what the solution is |
| 18:12:35 | fowl | id rather var x: string NOT allocate a new string until i say to |
| 18:15:12 | dom96 | The solution is a TMaybe type. |
| 18:18:47 | reactormonk | dom96, haskell-style? |
| 18:19:04 | dom96 | yes |
| 18:29:37 | * | tymat joined #nimrod |
| 18:29:50 | tymat | Hello! |
| 18:30:28 | Araq | hi tymat |
| 18:35:18 | fowl | i think my TMaybe is pretty useful >_> |
| 18:35:23 | fowl | definitely moreso than dom's :p |
| 18:35:41 | Araq | where is the difference? |
| 18:36:38 | fowl | mine checks isNil(v) if it can |
| 18:37:12 | fowl | and doms is immutable |
| 18:51:26 | * | [1]Endy quit (Ping timeout: 252 seconds) |
| 18:54:32 | * | Endy joined #nimrod |
| 19:21:03 | dom96 | fowl: Mine was just a proof of concept, I've never actually used it. |
| 19:38:53 | Araq | fyi: http://www.info.ucl.ac.be/~pvr/paradigmsDIAGRAMeng108.pdf |
| 19:40:31 | dom96 | Araq: I was once again thinking about babel, and again I think it should work as follows: instead of having the compiler find the installed babel libs, the compiler should simply be called by babel and babel should pass the location of the libs. |
| 19:40:56 | * | Endy quit (Ping timeout: 252 seconds) |
| 19:46:51 | dom96 | This would give full control to babel, but would disadvantage quick prototyping where you don't necessarily want to write a babel file and call babel when writing some small app. |
| 19:47:13 | dom96 | But in the end you would need a babel file anyway. |
| 19:48:55 | Araq | babel could do that already, it could pass things via --path |
| 19:49:02 | Araq | so I can't really follow |
| 19:52:58 | Araq | we only need to make "babel install" also invoke the compiler |
| 19:53:08 | Araq | or perhaps make that "babel build" |
| 19:53:30 | fowl | i like the way it is now, if im working inside nimlibs i just have path = ".." in nimrod.cfg and it finds the working fowltek before ~/.babel/libs/fowltek |
| 19:54:27 | dom96 | I suppose we can do both. --path overrides --babelPath right? |
| 19:54:40 | Araq | I'm not sure |
| 19:54:55 | Araq | I think command line overrides configuration file |
| 19:55:12 | Araq | so --path on the command line overrides --babelPath hidden in the config |
| 19:55:48 | dom96 | right, that would still be fine then. |
| 19:57:52 | dom96 | how does --babelPath choose package versions again? |
| 19:58:12 | Araq | 0.3.2 is better than 0.3.1 |
| 19:58:32 | * | comex` is now known as comex |
| 19:59:10 | Araq | this means 0.3.1 will never be in your search path unless you add it explicitly |
| 19:59:52 | Araq | but with babel listening all paths, we can simply get rid of --babelPath completely |
| 20:03:28 | dom96 | I was thinking about allowing dependency lists to specify git commit hashes as the version. |
| 20:06:40 | dom96 | Since most people don't feel like tagging their commits on git :P |
| 20:09:13 | Araq | yeah sounds like a useful feature |
| 20:12:00 | * | fowl quit (Quit: Leaving) |
| 20:14:03 | * | fowl joined #nimrod |
| 20:42:09 | dom96 | wow, that's a lot of deprecated messages for existsKey heh |
| 20:44:03 | NimBot | nimrod-code/nimbuild 7fd9035 Dominik Picheta [+0 ±2 -0]: irclogs: Each IRC message is serialized as a single line and appended to... 2 more lines |
| 20:58:17 | Araq | dom96: if babel invokes the compiler we can implement a very simple form of dependency analysis |
| 20:59:14 | Araq | the basic idea is that babel simply tries out versions and sees which ones make the project compile :P |
| 20:59:32 | Araq | this sounds like a hackish solution |
| 20:59:43 | Araq | and yet this is exactly what a human being does |
| 21:00:03 | dom96 | huh. Why not let the .babel file specify the version? |
| 21:00:14 | Araq | and it only fails if there are changes in the semantics which are not covered by static checking |
| 21:00:40 | Araq | well for a start we need a feature "babel check out the deps" anyway |
| 21:00:51 | Araq | so that a library author doesn't have to |
| 21:01:02 | dom96 | i'm not sure what you mean |
| 21:01:06 | Araq | and it's a nice fallback |
| 21:01:33 | Araq | ok, say I wrote blah-1.0 |
| 21:01:42 | Araq | and am creating a babel package for it |
| 21:01:45 | dom96 | You mean "babel get me the latest foobar package, I don't care about the version" |
| 21:01:49 | dom96 | ? |
| 21:01:51 | Araq | no |
| 21:01:59 | dom96 | ok, continue then |
| 21:02:16 | Araq | and I used the jester that happens to be installed on the machine for blah-1.0 |
| 21:02:34 | Araq | but I don't know which version of jester |
| 21:02:39 | Araq | I want to do: |
| 21:02:50 | Araq | babel createPackage blah-1.0 |
| 21:02:58 | Araq | and babel generates the babel file for me |
| 21:03:01 | dom96 | yeah, ok. |
| 21:03:03 | dom96 | That can come later. |
| 21:03:29 | dom96 | A good start is getting 'babel build' to work -- with dependencies. |
| 21:03:42 | Araq | yeah alright |
| 21:12:54 | * | _ponce joined #nimrod |
| 21:15:25 | * | ARCADIVS joined #nimrod |
| 21:23:22 | * | _ponce quit (Ping timeout: 246 seconds) |
| 21:29:05 | dom96 | Araq: Is there an easy way to set all fields of an object to a value which is "empty but non-nil". i.e. string -> "" and seq -> @[] |
| 21:29:28 | dom96 | I suppose I could write a proc to do it myself, but there is nothing in the stdlib is there? |
| 21:30:58 | * | _ponce joined #nimrod |
| 21:31:39 | Araq | dom96: indeed there isn't |
| 21:32:16 | fowl | write an initialize(var T) for each type ie initialize(x: var string) = if x.isNil: x = "" |
| 21:33:16 | dom96 | meh, i can just do it manually. |
| 21:33:30 | fowl | what do you need that for |
| 21:33:53 | fowl | dom96, is there a difference in the .babel whether you use key = val or key: val |
| 21:34:10 | fowl | the example .babel does both |
| 21:34:19 | dom96 | there is not |
| 21:34:39 | fowl | ok |
| 21:34:51 | dom96 | it's how parsecfg works; '=' and ':' are the same. |
| 21:41:31 | dom96 | I should really patch parsecfg so that quotes are not required for values without spaces. |
| 21:41:47 | Araq | dom96: yeah ... |
| 21:41:49 | dom96 | or in fact even with spaces. |
| 21:41:52 | Araq | no |
| 21:42:00 | Araq | that breaks stuff |
| 21:42:33 | dom96 | alright |
| 21:45:50 | dom96 | hrm, which do you think it more proper: |
| 21:46:02 | dom96 | *is |
| 21:46:13 | dom96 | let babelDir = getHomeDir() / ".babel" |
| 21:46:15 | dom96 | or |
| 21:46:29 | dom96 | proc getBabelDir: string = getHomeDir() / ".babel" |
| 21:46:38 | dom96 | and by proper I guess I mean idiomatic |
| 21:47:38 | Araq | I'd use the 'let' |
| 21:48:37 | dom96 | yep, that's what I was thinking |
| 22:16:45 | * | NimBot joined #nimrod |
| 22:19:15 | dom96 | !lag |
| 22:19:15 | NimBot | 1500ms between me and the server. |
| 22:19:31 | dom96 | And we are now using new logs. |
| 22:19:58 | * | tangentstorm joined #nimrod |
| 22:28:57 | * | comex quit (Remote host closed the connection) |
| 22:32:55 | * | comex joined #nimrod |
| 23:08:45 | dom96 | Araq: Why does pegs.find match whereas pegs.findAll does not, what's the difference? |
| 23:10:24 | dom96 | Araq: Also, shouldn't pegs.find/match work with a seq too? |
| 23:19:51 | * | fowl quit (Ping timeout: 256 seconds) |
| 23:31:29 | * | fowl joined #nimrod |
| 23:46:25 | reactormonk | Araq, so raise without heap, right? |