<< 15-05-2016 >>
00:06:14tautologicois the namespace for procs/templates the same as variable names? will a template named x be hidden by a local variable x?
00:21:05*Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
00:25:45*Demon_Fox quit (Ping timeout: 276 seconds)
00:39:38*yglukhov joined #nim
00:43:51*yglukhov quit (Ping timeout: 246 seconds)
00:51:05*ephja quit (Quit: WeeChat 1.5)
00:59:52*themagician quit (Ping timeout: 250 seconds)
01:03:07*Demon_Fox joined #nim
01:05:03*themagician joined #nim
01:09:17tautologicoseems to be a single namespace, so yes
01:12:47*lubos_cz quit (Ping timeout: 265 seconds)
01:15:09*yglukhov joined #nim
01:18:38Motanc2nim worked amazingly well, thanks araq_
01:19:47*yglukhov quit (Ping timeout: 260 seconds)
01:31:59Motanactually, maybe i spoke too fast, seems it just omitted generating lines for
01:32:00Motantypedef struct TF_Library TF_Library
01:32:13Motan(5 different output types), not entirely sure how to handle this
01:33:15Motanis there a way to say "this is an opaque type handled by the library"
01:36:53*arnetheduck joined #nim
01:53:35*lubos_cz joined #nim
02:02:32*yglukhov joined #nim
02:06:48*yglukhov quit (Ping timeout: 246 seconds)
02:14:39*lubos_cz quit (Ping timeout: 265 seconds)
02:29:08*yglukhov joined #nim
02:33:24*yglukhov quit (Ping timeout: 246 seconds)
02:53:41*yglukhov joined #nim
02:58:11*yglukhov quit (Ping timeout: 244 seconds)
03:08:52*aharlan joined #nim
03:18:38*yglukhov joined #nim
03:22:45*yglukhov quit (Ping timeout: 246 seconds)
03:33:14*brson quit (Quit: leaving)
03:54:51*yglukhov joined #nim
03:59:09*yglukhov quit (Ping timeout: 244 seconds)
04:10:22*mtj_ quit (Ping timeout: 252 seconds)
04:14:40*mtj_ joined #nim
04:19:31*yglukhov joined #nim
04:20:15*aharlan quit (Quit: Textual IRC Client: www.textualapp.com)
04:22:27*nsf quit (Quit: WeeChat 1.4)
04:23:39*yglukhov quit (Ping timeout: 246 seconds)
04:25:46*Jesin quit (Ping timeout: 252 seconds)
04:38:03IcyFoxyHow should one use constants exported by a header as functions to constant #define integers?
04:41:14*Jesin joined #nim
04:42:02*themagician quit (Ping timeout: 244 seconds)
04:42:12*themagician joined #nim
04:44:10*yglukhov joined #nim
04:48:30*yglukhov quit (Ping timeout: 246 seconds)
04:59:28*Motan quit (Quit: Page closed)
05:08:54*yglukhov joined #nim
05:13:42*yglukhov quit (Ping timeout: 260 seconds)
05:45:02*yglukhov joined #nim
05:49:24*yglukhov quit (Ping timeout: 246 seconds)
05:50:14*nsf joined #nim
05:51:08*zaquest joined #nim
06:09:39*yglukhov joined #nim
06:14:22*yglukhov quit (Ping timeout: 260 seconds)
06:38:24*yglukhov joined #nim
06:42:56*yglukhov quit (Ping timeout: 276 seconds)
07:02:34*yglukhov joined #nim
07:06:45*yglukhov quit (Ping timeout: 246 seconds)
07:17:00*endragor joined #nim
07:26:56*endragor quit (Remote host closed the connection)
07:27:10*yglukhov joined #nim
07:28:25*endragor joined #nim
07:31:04*yglukhov_ joined #nim
07:31:36*yglukhov quit (Ping timeout: 250 seconds)
07:36:21*rok joined #nim
07:58:04*kingofoz quit (Ping timeout: 252 seconds)
07:58:25*kingofoz joined #nim
08:20:15*Ven joined #nim
08:21:34*fredrik92 joined #nim
08:22:20*br3w5 joined #nim
08:23:07Araq_IcyFoxy: const Foo = 34 ? I don't understand your question.
08:23:37IcyFoxyAraq_: That doesn't get the constant from the header
08:23:58Araq_the better, so it can work with the LLVM backend.
08:24:17IcyFoxyI.e. If you want an array of length N. Where N is a #define N 32; with a function export of size_t export_N(void) { return N; }
08:24:25*fredrik92 left #nim (#nim)
08:24:25IcyFoxyHow would you get this N into a constant?
08:25:07IcyFoxyThe problem is hardcoding is bad; and if it's flexible to the library, then it may change without breaking.
08:25:20Araq_well if the library exposes it as a function, you have to import it as a function
08:25:40Araq_and if it's a function, you cannot use it to declare the array
08:25:47IcyFoxySure; and I cannot then set a constant to it's output?
08:26:42Araq_you can do something like const Foo = parseInt(staticExec("gcc ..."))
08:27:10IcyFoxyo.o
08:27:39Araq_but I'd say it's overkill for your situation.
08:27:58IcyFoxyAraq_: You think I should just hardcode it?
08:28:10Araq_tbh your situation makes no sense. even when I'm in C land there is no such thing as a constant/function hybrid
08:28:11*yglukhov_ quit (Remote host closed the connection)
08:28:31Araq_I need to rely on the fact it's a constant to declare the array
08:29:01Araq_the 'export_N' is a pointless exercise.
08:29:12IcyFoxyI agree.
08:29:39IcyFoxyBut how should I get the constants out of a C library that uses #declare instead of say.. static size_t N = 32; ?
08:29:54IcyFoxyOr s/static/const/
08:30:10Araq_#define becomes const
08:30:23IcyFoxyAraq_: Sure; but it isn't exported
08:30:45Araq_static size_t N = 32; is something like var N {.importc, header: "foo.h".}: int
08:31:02Araq_but then of course it's a 'var' since Nim doesn't know its value
08:31:47Araq_it doesn't matter if it's exported or not, you tell Nim the values
08:32:01IcyFoxyAlso isn't the header prgama just adding to a set of #include's and not explicitly forcing the header to be the source?
08:32:16IcyFoxy(Which would be nice for handling namespaces and I think could be done too)
08:32:25Araq_yes.
08:32:42Araq_well now it looks like your C lib wants its clients to us e export_N
08:32:52Araq_so import export_N as a proc
08:33:06Araq_that it's not a const doesn't matter then, since it's not a constant in C either ?
08:33:40IcyFoxyAraq_: Sure; but I need a const to use it for the arrays
08:33:49IcyFoxyAnd in C; you use the #define equivillent
08:34:08Araq_how can you if it's not exported?
08:34:39IcyFoxy<Araq_> #define becomes const
08:34:43Araq_seems like it's only an implementation detail you cannot rely on
08:34:55IcyFoxyDid you mean this as nim can use #define'd macros as constants?
08:35:01Araq_no
08:35:14IcyFoxyWell that #define N 32, N is used for the arrays.
08:35:35Araq_well either it's exported in the header or it isn't
08:35:55Araq_if it is exported, you can use 'const N = 32'
08:36:25Araq_or 'const N = parseInt(staticExec("gcc testprogramthatoutputsN.c"))'
08:36:51Araq_if you're so concerned that the value might change in the future without you noticing ...
08:37:08IcyFoxyExample https://github.com/jedisct1/libsodium/blob/master/src/libsodium/include/sodium/crypto_box.h
08:37:13IcyFoxyAnd this is the library in question btw
08:38:08IcyFoxyShould I instead send a PR for using const instead of #define? And would that be breaking ?
08:39:04Araq_what? you sound confused
08:39:13IcyFoxyMaybe
08:40:26Araq_I think you only the #defines and can ignore the dummy functions
08:40:31Araq_*only need
08:40:42IcyFoxyCorrect
08:41:10IcyFoxyconst N = {.emit "N".} maybe? (To use the #define macro?)
08:41:33Araq_as I tried to explain, this feature doesn't exist
08:41:43Araq_because *Nim* must know the value
08:42:09Araq_but the my staticExec solution will work fine
08:42:16IcyFoxyNot even with the emit?
08:42:36IcyFoxyI guess that makes sense actually
08:43:29IcyFoxyThe staticExec solution will require helper programs just to print the numbers and is very hacky...
08:44:34IcyFoxyI think I'll just go with hardcoded constants with test code to check that the constants == their function equivs.
08:45:39Araq_that's cool too.
08:45:58Araq_but I prefer staticExec, it's got a 3rd parameter to cache the results
08:46:54Araq_thinking about it ... 'importc' for 'const' could be done with staticExec under the hood and this would help the LLVM backend too. Hmmm.
08:47:39IcyFoxytemplate magic?
08:47:45IcyFoxyCustom pragma maybe?
08:48:08*br3w5 quit (Quit: br3w5)
08:48:08Araq_custom macro that looks like a pragma
08:48:19IcyFoxyCool
08:48:34*br3w5 joined #nim
08:50:13IcyFoxyLooking at the following rust; appears that there are const size_ts, laying around too. Although the math in #defines will need to be manually carried over. https://dnaq.github.io/sodiumoxide/src/sodiumoxide/src/crypto/box_/curve25519xsalsa20poly1305.rs.html#17
08:50:23IcyFoxyOr maybe not
08:51:02IcyFoxyhttps://github.com/dnaq/sodiumoxide/blob/master/libsodium-sys/src/crypto_box_curve25519xsalsa20poly1305.rs
08:51:06IcyFoxyHardcoded
08:51:22*nsf quit (Quit: WeeChat 1.4)
08:51:24IcyFoxyAlso with #[test] functions. Same idea :P
08:52:23IcyFoxyNim doesn't have any #[test] like infrastructure does it?
08:52:36IcyFoxyAt least builtin
08:53:10Araq_I use when isMainModule + doAssert
08:53:16Araq_both are builtin :P
08:54:38Araq_or use the popular unittest module. Not sure why so many people love it though ;-)
09:03:53*Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
09:08:48IcyFoxyAraq_: I was afk and also thought about isMainModule + asserts. :)
09:08:58IcyFoxyHave you tried out this unittest module?
09:09:19Araq_I fixed plenty of compiler bugs triggered by it
09:09:30Araq_so I'm biased.
09:10:05IcyFoxyWithout looking at it; it may do arbitrary inputs and expect good outputs; although with nim's affects system that is less necessary.
09:10:29IcyFoxyOr maybe not so less actually
09:10:36IcyFoxyStill plenty of possible problems :P
09:10:52Araq_it's nice that the bugs were found and fixed but I don't base my tests on a module that is so convoluted it triggers compiler bugs
09:11:23IcyFoxyOh the module itself triggered the bugs :P
09:11:42Araq_tests need to be as simple as possible.
09:11:43IcyFoxyI think I've recently triggered a compiler bug; but was busy trying to get things to work and didn't keep the buggy version...
09:11:48IcyFoxyShould have copied it out
09:11:56Araq_yup.
09:13:04IcyFoxySomething related to FFI I guess :P
09:14:19IcyFoxyThere we go; found in compile backlog (but not source) "SIGSEGV: Illegal storage access. (Attempt to read from nil?)"
09:15:49IcyFoxyReproduced!
09:22:45IcyFoxyAraq_: Odd
09:25:06*gokr joined #nim
09:25:46*rok quit (Remote host closed the connection)
09:26:01*rok joined #nim
09:26:22IcyFoxyAraq_: http://darkfox.id.au/ice.nim
09:26:43IcyFoxyAraq_: This causes the above error
09:27:06IcyFoxyI don't think it can reduce any further
09:27:25IcyFoxy(While keeping the bug at least)
09:30:51*lubos_cz joined #nim
09:36:14*Matthias247 joined #nim
09:41:45*br3w5 quit (Quit: br3w5)
09:42:53*nsf joined #nim
10:00:15*silven joined #nim
10:06:26*yglukhov joined #nim
10:32:03arnetheduckAraq_, regarding shared libraries, I would think that scripting languages use dlopen because they have to (ie they don
10:32:21arnetheduckt know which shared library will be used at compile time)
10:33:10arnetheduckalso, you don't need -devel packages to link with libraries - I would think that nim libraries work with specific versions of a library, and should thus link with a specific version
10:34:10arnetheduckie instead of passing -lXXX, an alternative would be -l:libXXX.so.Y (a specific version).. -devel packages are merely there for developer convenience
10:36:19*rok quit (Quit: rok)
10:37:18*silven quit (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
10:50:05*Demon_Fox quit (Quit: Leaving)
11:01:18*Salewski joined #nim
11:03:02SalewskiJust a remark, dlopen can give namespace problem, see https://mail.gnome.org/archives/gtk-list/2015-March/msg00016.html # But a solution was available, see that thread.
11:04:51*elrood joined #nim
11:06:58*Salewski quit ()
11:21:11dom96Salewski: sounds like something you should report on Github
11:29:05*yglukhov quit (Remote host closed the connection)
11:40:58*yglukhov joined #nim
11:41:07*Matthias247 quit (Read error: Connection reset by peer)
11:45:03*yglukhov quit (Ping timeout: 240 seconds)
11:47:08*lubos_cz quit (Ping timeout: 276 seconds)
11:48:51federico3dom96: some packages put their modules under <pkgname>/ , some use src/ , some private/ - is there a preferred way documented somewhere?
11:49:25dom96whether you put it under src/ or not is up to you
11:50:41dom96The rest is documented here: https://github.com/nim-lang/nimble#creating-packages
11:51:05federico3tnx
11:57:21Araq_arnetheduck: interesting concept of "convenience" ;-)
12:02:28arnetheduckAraq_, well, stretching it a bit, but not too far - the unversioned symlink exists for no other purpose (ie it's not used at runtime)..
12:02:40arnetheduckalso, if nim wants to avoid devel, it must also not use header files, and that is a bigger issue than getting it to link correctly
12:05:06*lubos_cz joined #nim
12:05:34*yglukhov joined #nim
12:07:02arnetheduckanyway, mostly curious.. I I can imagine that for nontrivial libraries, dlopen + dlsym becomes slow
12:07:49Araq_I can imagine that it stays fast
12:08:10dom96it's all loaded into memory anyway is it not?
12:09:51*yglukhov quit (Ping timeout: 246 seconds)
12:10:02arnetheduckslow, as in, takes a while on startup.. ie, if loading gtk libs, your app takes a little longer to show up
12:10:33*enthus1ast joined #nim
12:13:36*rok joined #nim
12:15:15*nsf quit (Quit: WeeChat 1.4)
12:15:53dom96true. It would be interesting to see benchmarks for this.
12:20:01elroodpreferrably by someone who knows his way around RTLD_LAZY etc.
12:29:40*yglukhov joined #nim
12:33:53*br3w5 joined #nim
12:34:16*yglukhov quit (Ping timeout: 244 seconds)
12:40:07*endragor quit (Remote host closed the connection)
12:40:45*endragor joined #nim
12:45:12*br3w5 quit (Quit: br3w5)
12:53:56*yglukhov joined #nim
12:57:04enthus1asti'm playing with ed25519.nim but when i'm running the test.exe i receive
12:57:04enthus1astTraceback (most recent call last)
12:57:04enthus1asttest.nim(5) test
12:57:04enthus1asted25519.nim(64) createKeypair
12:57:04enthus1astSIGSEGV: Illegal storage access. (Attempt to read from nil?)
12:57:46enthus1asti fiddled around with the nim file a little but could not figure out whats the problem here
12:58:20*yglukhov quit (Ping timeout: 250 seconds)
12:59:12enthus1astniv, can you build it with the 0.13?
12:59:30*niv wakes up
12:59:41nivuhm.
12:59:49nivi did dev it against .13. not sure whats wrong there.
12:59:52nivis that on linux?
12:59:54nivosx?
12:59:59enthus1astwindows 7
13:00:16*Ven joined #nim
13:00:20nivhmm.
13:00:37nivi didnt test on windows, but i'll check
13:01:36enthus1astthis would be super nice :)
13:02:41nivcan you pastebin/gist your code that triggers the segv?
13:02:57enthus1astthats your test.nim
13:03:08enthus1asti just compiled it and run it
13:03:17nivfunny. i just tried compiling/running the test code in the repo on win10 and it worked fine
13:03:26enthus1astmhhhh
13:03:29nivi dont have win7 around unfortunately
13:03:39enthus1astsex
13:03:40enthus1astsec
13:03:42enthus1ast...
13:04:01nivits literally just a thin wrapper around https://github.com/orlp/ed25519/
13:17:47*rok quit (Quit: rok)
13:21:29enthus1astniv, you just build it with `nim c test.nim` right?
13:29:18*yglukhov joined #nim
13:29:26niventhus1ast: yes
13:29:41nivand then run the resulting test.exe
13:30:10*br3w5 joined #nim
13:33:15*nsf joined #nim
13:46:03*br3w5 quit (Quit: br3w5)
13:50:37*themagician quit (Ping timeout: 252 seconds)
13:52:46*Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
13:56:42*themagician joined #nim
13:57:07*tankfeeder joined #nim
14:17:00*PMunch joined #nim
14:55:33*enthus1ast1 joined #nim
14:55:45*enthus1ast quit (Ping timeout: 246 seconds)
14:59:45*Arrrr joined #nim
14:59:45*Arrrr quit (Changing host)
14:59:45*Arrrr joined #nim
15:01:45*yglukhov quit (Read error: Connection reset by peer)
15:02:08*yglukhov joined #nim
15:10:12*yglukhov quit (Remote host closed the connection)
15:22:04*yglukhov joined #nim
15:26:19*yglukhov quit (Ping timeout: 244 seconds)
15:38:58*PMunch quit (Ping timeout: 252 seconds)
16:09:39*br3w5 joined #nim
16:10:03*br3w5 quit (Client Quit)
16:10:08*Trustable joined #nim
16:20:14*br3w5 joined #nim
16:22:24*yglukhov joined #nim
16:35:57*Trustable quit (Remote host closed the connection)
16:36:03Araq_niv: maybe you got the calling conventions or linking wrong?
16:43:43Araq_One of C's great strengths is that its cost model is very predictable: it is easy to look at any small fragment of C code can instantly get an accurate idea of what machine operations will have to be performed to execute that code.
16:43:44Araq_"
16:43:59Araq_oh god, when will people stop with this nonsense.
16:44:15Araq_nobody can see cache misses in the C code.
16:44:36Araq_that's a factor of 100 in your "cost model" that's hidden from you.
16:44:54elroodare they easier to spot in any other language?
16:45:42Araq_no. there is no such thing as an easy cost model on a half modern CPU
16:46:55Araq_also auto-vectorization kills any further "I can see the asm behind the C" advantage
16:47:35Araq_a modern C compiler is millions of lines of code dealing with all sort of code transformations.
16:49:11Araq_it's absurd to claim "C is close to the metal" this day and age and only shows ignorance and hubris
16:50:07*ludocode quit (Read error: Connection reset by peer)
16:50:21*ludocode joined #nim
16:52:02elroodit's the closest-to-metal language we have bare pure assembly, the portable assembly analogy didn't emerge out of the blue, and while with today's cpus the second part of your cited statement for sure is absolute nonsense, the beginning of it still stands
16:53:57elroodwhether a lack of sophisticated abstractions is indeed a strength remains debatable, of course
16:55:39elroodregarding your comment for niv, would the code then work on win10?
16:56:20nivim not dealing with calling conventions at all, im just compiling the code directly with nim through the compile macro
16:57:04nivhttps://github.com/niv/ed25519.nim/blob/master/ed25519.nim#L1
16:57:13nivand then i just do .importc:.
17:03:34*libman joined #nim
17:08:56Araq_niv: which OS?
17:09:30Araq_elrood: in former times C was announced as a "high level language". now what?
17:09:40Araq_see the Unix ad youtube video for proof.
17:16:03*br3w5 quit (Quit: br3w5)
17:16:49*sarlalian quit (Ping timeout: 260 seconds)
17:18:50*sarlalian joined #nim
17:19:55elroodand those were times when algol and lisp were already around, so if one wanted to pick on a statement about c for its absence of sensibility, why not choose that one? ;)
17:20:19*enthus1ast1 quit (Ping timeout: 260 seconds)
17:23:33*enthus1ast joined #nim
17:31:26nivAraq_: well, works for me on osx, linux, and win10 ..
17:32:51tankfeederniv, i have ed25519 implementation on picolisp
17:33:04tankfeederhttps://bitbucket.org/mihailp/tankfeeder/src/f790d5322f8f7e9e1bd19d3d831b8223064450d9/crypto/ed25519.l?at=default&fileviewer=file-view-default
17:33:08tankfeederjust in case
17:33:51nivtankfeeder: err, okay? not seeing what this has to do with my crappy nim wrapper though
17:35:41tankfeederits native. its ok
17:36:16nivah, so you can compile picolisp to a C-linkable? thats cool
17:37:03tankfeederit works in both ways
17:37:25tankfeederauthor created his own high level assembler
17:37:32tankfeederyou must see it
17:38:03tankfeederhttps://bitbucket.org/mihailp/picolisp/src/74db5264f7aafb36532ccf465c0a9860523e257a/src64/flow.l?at=default&fileviewer=file-view-default
17:39:51*ludocode quit (Read error: Connection reset by peer)
17:40:08*ludocode joined #nim
17:43:03*enthus1ast quit (Ping timeout: 246 seconds)
18:03:34*yglukhov_ joined #nim
18:05:14*yglukhov quit (Ping timeout: 260 seconds)
18:05:19*yglukhov_ quit (Read error: Connection reset by peer)
18:08:52*yglukhov joined #nim
18:10:53*yglukhov quit (Read error: Connection reset by peer)
18:11:09*br3w5 joined #nim
18:19:12*yglukhov joined #nim
18:21:14*yglukhov quit (Read error: Connection reset by peer)
18:23:31*Jesin quit (Quit: Leaving)
18:27:49*yglukhov joined #nim
18:32:22*br3w5 quit (Quit: br3w5)
18:34:06*Jesin joined #nim
18:51:02Araq_niv: so your wrapper crashes on every OS?
18:52:01*tankfeeder quit (Quit: Leaving)
18:52:33nivAraq_: i cant repo the crash that enthus1ast on win7 reported. it works on all my machines (osx, lin, win10).
18:52:52Araq_ah ok, hmm
18:56:33cheatfateAraq_, just a simple question how to properly importc of #define to var or to const?
18:57:43Araq_read the logs lol
18:57:49Araq_from today.
18:58:09cheatfatethanks i got an answer from my lovely posix.nim
18:58:57*rok joined #nim
19:21:16*PMunch joined #nim
19:28:09dom96That reminds me. IcyFoxy, are you making a libsodium wrapper? because federico3 already made one.
19:33:25cheatfateIs anybody can run my gist https://gist.github.com/cheatfate/ff5dec454f00c62085f3b49787eb3263 on MacOS?
19:33:29cheatfateand give me results
19:36:12dom96cheatfate: 10240
19:36:12dom969223372036854775807
19:36:33cheatfatedom96, thanks
19:36:41dom96np
19:52:31*shodan45 joined #nim
20:09:44*Trustable joined #nim
20:10:00*Trustable quit (Remote host closed the connection)
20:12:04*Trustable joined #nim
20:12:30*Trustable quit (Remote host closed the connection)
20:17:14*ludocode quit (Read error: Connection reset by peer)
20:17:33*ludocode joined #nim
20:17:54*Trustable joined #nim
20:34:35*Arrrr quit (Quit: WeeChat 1.4)
20:47:37*shodan45 quit (Quit: Konversation terminated!)
20:51:19*endragor_ joined #nim
20:53:06reactormonkdom96, ![]() inserts an image - should I change your comment?
20:53:28dom96reactormonk: lol no. I didn't include it inline on purpose.
20:53:32reactormonkdom96, kk
20:53:45dom96because professionalism :)
20:54:51*endragor quit (Ping timeout: 246 seconds)
20:55:26*endragor_ quit (Ping timeout: 244 seconds)
20:55:51*libman quit (Remote host closed the connection)
21:01:23*fastrom1 quit (Quit: Leaving.)
21:14:00*yglukhov quit (Remote host closed the connection)
21:16:47*elrood quit (Quit: Leaving)
21:25:54*yglukhov joined #nim
21:29:52*Matthias247 joined #nim
21:30:03*yglukhov quit (Ping timeout: 244 seconds)
21:44:16*MyMind joined #nim
21:45:16*Mat4 joined #nim
21:46:08*Sembei quit (Ping timeout: 250 seconds)
21:50:27*yglukhov joined #nim
21:55:21*yglukhov quit (Ping timeout: 276 seconds)
22:08:29*Mat4 left #nim (#nim)
22:12:00*Matthias247 quit (Read error: Connection reset by peer)
22:26:49*yglukhov joined #nim
22:31:01*yglukhov quit (Ping timeout: 244 seconds)
22:32:34dom96I think it may be time to set up SSL for nim-lang.org and the forum with Let's Encrypt (even if it breaks Windows XP). Thoughts?
22:35:47Araq_dom96: hell yes.
22:37:33dom96I was going to set up Slack for us (by creating the IRC<->Slack bridge) until I realised that it's not just a simple checkbox...
22:37:39*dom96 yawns
22:37:42*rok quit (Quit: rok)
22:38:26*PMunch quit (Ping timeout: 276 seconds)
22:41:27cheatfatedom96, i think you can always buy ssl certificate if you want to support windows xp
22:42:15dom96cheatfate: I'm guessing that Win XP users are shown error for many sites as a result of Let's Encrypt anyway
22:42:20dom96so hopefully they're used to it?
22:42:56cheatfatedom96, i think only windows xp users with default not updated internet explorer...
22:43:13dom96hrm, that wouldn't be so bad
22:43:20*Demon_Fox joined #nim
22:43:23dom96are you sure that the SSL certs are not an OS thing though?
22:44:04cheatfatei think (but not sure) browsers install their own CA authorities to system
22:45:36cheatfatefirefox uses its own base of CA authorities... dont know about chrome
22:47:03*gokr quit (Ping timeout: 240 seconds)
22:48:46cheatfatebut looks like chrome on windows using os ca base...
22:58:43dom96This new Common Lisp site looks very nice http://lisp-lang.org/
23:00:11cheatfatedom96, it looks responsive :)
23:01:28tautologicoyou can access a slack team via IRC easily, but I guess the idea was integrating this here IRC channel with slack
23:03:01*yglukhov joined #nim
23:03:43dom96tautologico: yep, as far as I can see in order to do that I need to run a bot.
23:06:04cheatfatedom96, please, help asyncdispatch :) stop spawning new projects
23:06:33*Trustable quit (Remote host closed the connection)
23:06:37Xeplease don't use slack
23:06:44Xei will veto nim if it uses slack
23:06:57dom96cheatfate: too late now :P
23:07:11*yglukhov quit (Ping timeout: 244 seconds)
23:07:48dom96Xe: Why? It will be connected to this channel.
23:07:57Xedom96: please use https://matrix.org
23:08:17dom96The idea is to lower the barrier of entry to chat with us
23:08:21dom96nothing more
23:08:29Xeyes, but slack is harmful
23:09:39dom96Perhaps. But in what way can it harm us?
23:09:41cheatfatedom96, we have web client for irc, what can be easier? or installing custom client to speak would be better?
23:09:44*kingofoz quit (Ping timeout: 260 seconds)
23:09:54*kingofoz joined #nim
23:10:31dom96cheatfate: many people already have Slack installed
23:10:39dom96the idea is to get them to join us
23:11:41cheatfatedom96, i dont think messenger is a problem... if they want they join even if you will setup messaging by telnet
23:12:13dom96cheatfate: I can say for certain that if this chat required telnet that no one would join.
23:15:40cheatfatedom96, the best way where you can spare your time is to solve some bugs, imho
23:16:24cheatfatedom96, i'm waiting for your help for 2.5 month
23:21:53dom96I appreciate how you think I should be spending my time, but it is my time. Why are you waiting for my help on those issues? Why are they so important to you?
23:27:14*yglukhov joined #nim
23:31:39*yglukhov quit (Ping timeout: 246 seconds)
23:32:04cheatfatemaybe because there many work/projects based on asyncdispatch which has stack overflow inside like a `time bomb` so there no possibility to use this work in `production` stage...
23:33:06cheatfateand its important to me, just because i'm want to use asyncdispatch in my projects and i can't just because i can't put in production code with `stack overflow` inside
23:36:50dom96Alright. I will look into fixing it tomorrow.
23:37:35dom96Would now but I don't think I can produce a good solution at this time :)
23:39:50cheatfatedom96, thank you
23:40:42dom96cheatfate: can you remind me what the reason was that we can no longer use callSoon with a queue?
23:41:53cheatfatedom96, we can use callSoon but we need to have another place for dispatching queued callbacks...
23:43:41cheatfatedom96, https://github.com/nim-lang/Nim/issues/4160 this shows why we can't dispatch callbacks in poll()
23:45:52cheatfatetasyncdiscard.nim test is a main problem
23:46:15dom96does it? It just shows your reproduction
23:46:40cheatfatemy #4160 is just improved version of tasyncdiscard.nim
23:51:04dom96yeah, but why can't we dispatch callbacks in poll()?
23:51:37cheatfatebut tasyncdiscard.nim and #4160 don't using poll at all
23:55:05dom96oh
23:57:55*brson joined #nim
23:57:55*brson quit (Client Quit)
23:58:01dom96Only thing I can think of right now is to change asyncdispatch so that it doesn't run async procs immediately