🦋 Welcome to Raku! raku.org/ | evalbot usage: 'p6: say 3;' or /msg camelia p6: ... | irclog: colabti.org/irclogger/irclogger_log/raku
Set by ChanServ on 14 October 2019.
00:30 Tirifto left 00:34 rindolf left 00:44 pecastro left
tony-o .tell guifa sql::slang might fit your definition of "integrated into the language" 01:34
tellable6 tony-o, I'll pass your message to guifa2
01:42 wamba joined 02:41 jmchael left 03:06 Manifest0 left, Manifest0 joined
xinming SmokeMachine: when we use infix<in>, the RHS is an empty array, This will cause Red to generate sql which is like xxxx IN ( ) Which results an error, What is the sensible behavior in this case do you think please? 03:37
Or we should raise error too just like postgres does? 03:38
03:39 xelxebar left
xinming I'll write a patch which raise error just like postgres does. 03:39
03:39 xelxebar joined 04:14 __jrjsmrtn__ joined 04:15 _jrjsmrtn left 04:42 rindolf joined 04:59 thdrmz left 05:00 thdrmz joined 05:04 thdrmz1 joined 05:05 thdrmz left, thdrmz1 is now known as thdrmz 05:11 thdrmz left, thdrmz1 joined 05:13 thdrmz1 is now known as thdrmz 05:53 zln_ joined 05:54 zln left 06:02 vivelakira joined 06:05 vivelakira left 06:23 wamba left 06:36 webstrand left 06:44 brtastic joined 07:08 parabolize left
xinming What is the best way to make die on warning? 07:26
07:40 brtastic left 07:41 brtastic joined 07:58 cog joined 08:02 cog__ left 08:06 brtastic left, wamba joined 08:21 ranpoom joined 08:33 brtastic joined 09:02 Doc_Holliwood joined 09:04 wamba left 09:15 Altai-man joined 09:28 wamba joined 09:43 leah2 left
CIAvash xinming: colabti.org/irclogger/irclogger_lo...10-29#l150 ? 09:52
09:57 leah2 joined 10:18 sjm_uk joined
Geth doc: 6343e91df0 | (JJ Merelo)++ | doc/Language/variables.pod6
Align and move left
10:31
doc: 842a581618 | (JJ Merelo)++ | doc/Language/variables.pod6
Avoid wrapping
linkable6 Link: docs.raku.org/language/variables
10:41 wamba left 10:43 webstrand joined 10:46 Sgeo left 11:06 pecastro joined 11:16 sjm_uk left 11:53 sjm_uk joined 11:55 aborazmeh joined, aborazmeh left, aborazmeh joined 11:57 sena_kun joined 11:58 Altai-man left 12:02 gugod left 12:05 aborazmeh left 12:29 wamba joined 13:16 jmchael joined 13:31 sjm_uk left
gfldex lolibloggedalittle: gfldex.wordpress.com/2020/12/27/as...ctability/ 13:40
I found myself trying to avoid naming 2020. Good riddance. 13:41
14:08 pmurias joined
pmurias timotimo: that mode would keep you from getting distracted by lesser stuff while having important irc conversations? 14:10
14:17 zln_ left, xelxebar left, zln joined, xelxebar_ joined 14:19 tejr left 14:20 tejr joined 14:21 asymptotically left, gugod joined 14:24 asymptotically joined
xinming CIAvash: ??? 14:25
14:31 aborazmeh joined, aborazmeh left, aborazmeh joined 14:32 wamba left
timotimo pmurias: probably! 14:32
14:36 aborazmeh left 14:38 aborazmeh joined, aborazmeh left, aborazmeh joined 14:39 parabolize joined 14:43 notagoodidea joined 14:47 aborazmeh left 14:49 aborazmeh joined, aborazmeh left, aborazmeh joined 15:05 wamba joined
notagoodidea Hi all! Does Raku expose the GMP binding used in Rakudo/MoarVM? 15:06
pmurias Raku has built in big integers 15:15
15:17 orinthe left
notagoodidea Yes I know, but I looking at the functions for Int/Numeric in the doc, I don't see some usually found in GMP binding. 15:17
So if the GMP in full binded to provide BigInt in Raku, can we access those binding? For example, in Julia GMP is also used to provide BigInt built-in and offers to expose GMP binding via the standard lib (Base). 15:19
15:28 orinthe joined
timotimo we don't use the gnu multiprecision library in moarvm 15:31
someone is working on a patch to get a fork of gmp in 15:32
notagoodidea Ah that why it was mixed up in my head.
that's* 15:33
timotimo at the moment we have libtommath
gmp is not great to use inside of something like moarvm, it doesn't really offer error handling, instead seems to prefer sending a signal
15:34 aborazmeh left
timotimo i guess that's cool if you have only a single thread? 15:34
notagoodidea Yep probably. Never really tried to leverage GMP in multi-threaded environment. I mostly played with the Julia binding when messing with the collatz conjecture for fun. 15:35
Geth doc: uzluisf++ created pull request #3752:
Document several metamethods
15:42
notagoodidea I am just a beginner at Raku in the "What if" & "How I tranlsate what I know" phase so weird questions pops up in my mind. 15:44
lizmat notagoodidea: weird questions are ok in my book, as they show out-of-the-box thinking 15:49
MasterDuke the moarvm gmp branch builds on linux and passes all tests (and is much faster at bigint math operations), but needs someone who knows (and has access to) windows to get it building there 15:51
notagoodidea Does it leverage multi dispatch to provide a full binding to GMP API or can we expect to call directly GMP binding if it is merged? 15:52
15:56 Altai-man joined
notagoodidea lizmat: Thanks. otb thinking I don't know but we all have a background and concept that we try to hold to when learning something before being able to let it go and embrace a new way to approach topics. 15:58
15:58 sena_kun left 15:59 domidumont left, domidumont joined
lizmat multi dispatch only exists at the HLL level 15:59
notagoodidea Ah, I have been bitten by a "Numeric overflow", is it an implementation choice (as big int are provided) ? 16:00
lizmat the GMP API binding is ultimately exposed as NQP ops in Rakudo
Seeing numeric overflow errors are rare in Raku, afaik :-)
because bigints are the default 16:01
notagoodidea Yeah, I just tried 10**100**100.
timotimo well, how big is that number?
m: say (10 ** 5) ** 6; say 10 ** (5 ** 6); say 10 ** 5 ** 6 16:02
camelia 1000000000000000000000000000000
1000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000…
16:02 cpan-raku left
timotimo er, yeah 16:02
MasterDuke notagoodidea: there are no nqp or rakudo changes, it's all under the hood
timotimo m: say ((10 ** 5) ** 6).chars; say (10 ** (5 ** 6)).chars; say (10 ** 5 ** 6).chars
camelia 31
15626
15626
16:03 cpan-raku joined, cpan-raku left, cpan-raku joined
timotimo m: say 100 ** 100 16:03
camelia 100000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
timotimo m: say (100 ** 100).chars
camelia 201
MasterDuke gmp does allow bigger numbers than libtommath
timotimo how does that grow 16:04
10 ** (100 ** 100) is just a number as long as (1 with 201 zeroes) digits? 16:05
notagoodidea timotimo: Add one unt to the second exponent, it add an order of digits in the number. 10*100*1 = 101 digits, 10*100*2 = 1001 digits.
s/unt/unit
timotimo m: say (100 ** 100).log
camelia 460.51701859880916
timotimo does that mean the number of digits fits into a 460 bit number? 16:06
aukkras m: say (100 ** 100).log/2.log 16:07
camelia 664.3856189774725
aukkras 665 bit number
timotimo so you were expecting rakudo to do what, exactly?
IBM Physical server Power System E980: 64 TB 16:08
how many of those would you need to rent at the same time to store the number?
MasterDuke m: say (2**10000000000).chars 16:09
camelia Numeric overflow
in block <unit> at <tmp> line 1
timotimo i mean i guess you don't have to store it in ram, you could store it on a few .. kilometers? of magnetic tape
MasterDuke ^^^ succeeds with gmp
timotimo i think i'm getting the calculations wrong 16:10
honestly, it's possible that there's an optimization for 2 ** x for very high x 16:11
since that can be satisfied with a single page of memory plus lazily allocating zeroed pages :P
16:12 gordonfish left
timotimo 10 ** x on the other hand, not as easily, unless i guess you also implement BCD? 16:13
m: say (10 ** 100).base(2)
camelia 10010010010011010110100100101100101001100001101111100111010110000101100100111100001001100010011001110000010111111001110001010110011100100000010001110001000010001101001111100101010101011001001000011000010001010100000101110100011110001000000000000000000… 16:14
timotimo i'm too lazy to look up if there's a closed form for digit index to digit for 10 ** x in log2
notagoodidea timotimo: If you want to get the number of digits in an number : ($n).log10 + 1 and floor it if it is decimal. But not sure it was your question/observation. 16:15
timotimo ok, how many bits do you need to store the result of 10 ** (100 ** 100)? 16:16
16:16 gordonfish joined 16:17 zln left
raku-bridge <frost> m:sub f {return 1,'a'}; my ($a,$b); $a,$b = f; say "$a $b"; 16:21
evalable6 WARNINGS for /tmp/vq7KT6CDKb:
1 a
Useless use …
frost, Full output: gist.github.com/0abc9b0cc661f264f5...d2bd464339
SmokeMachine xinming: I thought I’d fixed that... but if it’s an empty array, it should be changed to something that’s always false, like `1 = 0`
raku-bridge <frost> m:sub f {return 1,'a'}; my ($a,$b); ($a,$b) = f; say "$a $b"; 16:22
evalable6 1 a
MasterDuke m: sub f {return 1,'a'}; my ($a,$b); $a,$b := f; say "$a $b";
camelia WARNINGS for <tmp>:
1 a
Useless use of $a in sink context (line 1)
Use of uninitialized value element of type Any in string context.
Methods .^name, .raku, .gist, or .say can be used to stringify it to something meaningful.
in block <unit…
aukkras m: (100 ** 100 ) * (10.log / 2.log) 16:23
camelia WARNINGS for <tmp>:
Useless use of "*" in expression "(100 ** 100 ) * (10.log / 2.log)" in sink context (line 1)
16:23 cetjs2 joined
cetjs2 hello 16:23
aukkras m: say (100 ** 100) * (10.log / 2.log); 16:24
camelia 3.3219280948873626e+200
cetjs2 What is Web Frameworks for raku? what is popular qnd laravel-like?
raku-bridge <frost> Cro? 16:25
cetjs2 what is cro? 16:26
lizmat cro.services
SmokeMachine cetjs2: I think it’s cro
16:28 Tirifto joined
SmokeMachine cetjs2: are you going to use db? 16:29
Geth doc: 5e1260bcbb | (Luis F. Uceta)++ (committed by Juan Julián Merelo Guervós) | doc/Type/Metamodel/PrivateMethodContainer.pod6
Document several metamethods
linkable6 Link: docs.raku.org/type/Metamodel::Priv...dContainer
notagoodidea Ah. Would take the products of the log2, no? But trying with 10**10**3 give Inf whith log2.
m: (10**10**3).log2
camelia ( no output )
notagoodidea m: (10**10**3).log2.say
camelia Inf
SmokeMachine cetjs2: you can also take a look on Red if you want a ORM... 16:30
cetjs2 SmokeMachine, I think using sqlite3... who is red orm? this like hibernate or elegant orm? 16:33
SmokeMachine cetjs2: it’s more Raku ish... and it works with SQLite too... memory and file... 16:35
modules.raku.org/dist/Red:cpan:FCO
cetjs2: ^^
cetjs2: fco.github.io/Red/ 16:40
16:40 guifa2 joined
guifa2 o/ 16:40
tellable6 2020-12-27T01:34:07Z #raku <tony-o> guifa sql::slang might fit your definition of "integrated into the language"
guifa2 tony-o absolutely! Dunno how I forgot about that one
SmokeMachine guifa2: I would also say Red is “integrated into the language”... 16:41
guifa2 SmokeMachine: that was already in my list of modules to show off ;-) 16:42
SmokeMachine guifa2: :)
notagoodidea m: say 1 + (100**100) * 10.log10 16:43
camelia 1e+200
notagoodidea m: say (1 + (100**100) * 10.log10).base(2).chars 16:44
camelia 665
notagoodidea timotimo: 665 bits for 10**100**100, it seems low no? 16:45
timotimo yeah
aha 16:46
wolframalpha to the rescue
guifa2 Hmm, what might the best way to split a string when you want to capture a combining grapheme?
combining character* 16:47
timotimo the resulting number has 100 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 001 decimal digits
guifa2 ah, probably $foo.Uni.split 16:48
notagoodidea m: say (1 + (100**100) * 10.log10).chars
camelia 6
notagoodidea timotimo: So the 1 + $exposant * base.log10 is correct. 16:49
But I converted the number of digits in bits, that's why is so low -_- 16:50
timotimo right
so, moar is perhaps justified in calling that an overflow
notagoodidea m: say (1 + (100**100) * 10.log2).floor 16:51
camelia 332192809488736258156007321626732964933652037630384773984828053001819499610649784180077247442368308425414021275635919348520235492085887256739957544934325558774604993429445235567958526664961728068976640
notagoodidea Hahaha yep, it seems that would a correct estimate of the number of bits needed. 16:52
timotimo: So the next question si what is the limit? Is it a hard limit from MoarVM or libtommath or machine-dependant? 16:53
timotimo may very well be a hard limit in moar itself 16:56
notagoodidea But what surprise me too, is that the taking the log from (10**10**3) raises Inf. Does Numeric does not drop to FatRat? 17:00
m: say (10**10**3).log2.^name 17:01
camelia Num
notagoodidea ok, I am lost in the type hierarchy of Raku now. :) 17:02
lizmat TimToady decided it would fallback to Num, as the performance of FatRat degrades pretty quickly and pretty bad
if you want FatRats, you need to start with FatRats 17:03
notagoodidea Ok and Num is bounded by 64 bits float if I understand correctly from the doc? 17:04
lizmat yup, afaik :-)
notagoodidea lizmat: Um. even with FatRat, log2 returns Inf. `my $num = FatRat.new(10**10**3, 1); $num.log2` => Inf and $num.^name => FatRat. 17:08
lizmat hmmm... perhaps .log2 always returns a num because the underlying math lib doesn't know about Rats ?
guifa2 the problem is the log function; I think, which isn't designed to work with Rats (because of the iunerling
^^
notagoodidea Because log2 is supplied by the Numeric role and so its bounded by the Numeric restriction?
lizmat no, it's a limitation of the underlying math lib, I'd say 17:09
aukkras m: say (10**10**3) 17:11
camelia 10000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000…
notagoodidea hum ok. I am not sure to have a clear model between class and role (Num vs Numeric in this case) yet. 17:12
guifa2 Numeric is the numbers equivalent of Stringy 17:13
Numeric just describes something that can function as a number-like thing
notagoodidea lizmat: Ok, probably. Julia seems to handle it without trouble (it is GMP backed). 17:14
guifa2: And Num describe what number is? 17:15
lizmat notagoodidea: not sure what Raku isn't handling properly wrt to Julia ?
guifa2 Num is just a specific representation (that is backed internally by a floating point number)
notagoodidea lizmat: Throwing the Inf for the log2(10**10**3), sorry.
So not Raku per se but the undrelying libs for big int. 17:16
lizmat I seem to recall that's a deficiency of libtommath
notagoodidea guifa2: thanks, it is a bit clearer :)
17:16 ranpoom left
lizmat that gmp doesn't have, and one of the reasons for trying to migrate to it 17:16
guifa2 Theoretically, you could make a Hyperreal class, define it as Numeric, and it would play well with math functions. No idea how you'd encode it internally though hahaha 17:17
But the idea is that all number operations shoudl return something that's Numeric. So Int / Int doesn't get you Int, but a Rat. But most of the time, it's not important 17:18
(the idea being that people who do care about the difference know how to deal with it normally, and those who don't, don't have to worry about it)
guifa2 . o O ( unfortunately, while most folks knokw to use Numeric to accept all numbers, they're not as good about using Dateish or Stringy, although with the new coercion mechanics — vrurg++ — that might not be a huge issue anymore) 17:24
notagoodidea guifa2: Yep, the underlying logic is clear now, at least until nex time it will bite me :D 17:25
18:15 jmerelo joined
jmerelo .tell El_Che very interesting talk by Carmen Andoh... 18:17
tellable6 jmerelo, I'll pass your message to El_Che
guifa2 waves at jmerelo 18:21
did you see my FOSDEM proposal?
jmerelo guifa2: yep I did. Already graded it; however, last time I checked only grades were by me, so I'm waiting for someone else to check them out before we send out notices. 18:22
tellable6 2020-12-23T19:47:14Z #raku <guifa2> jmerelo when you get a chance, I'm finishing up my FOSDEM proposal, have a few Qs
guifa2 ^^lol @ tellable 18:23
18:23 domidumont left 18:27 mowcat joined, wamba left, jmerelo left 18:28 jmerelo joined
jmerelo Not sure my last message got in 18:28
Lights went out for a few seconds
guifa2: I see it got through. Still have some questions? 18:34
18:35 cetjs2 left
guifa2 jmerelo: nope, just checking that it got in on time ^_^ 18:40
18:41 pmurias left 18:48 aluaces left 18:50 aluaces joined 19:01 notagoodidea left 19:20 thundergnat joined
El_Che jmerelo: the goconf talk? Yes, it was surprisingly interesting and applicable to Raku 19:23
tellable6 2020-12-27T18:17:43Z #raku <jmerelo> El_Che very interesting talk by Carmen Andoh...
19:26 b2gills left 19:35 notagoodidea joined 19:40 Sgeo joined 19:43 oboe joined
oboe merry festivus all! can someone help me with stopping automatic type coercion? 19:44
i have a simple sub which takes an integer, but if instead the command line looks like --option rather than --option=number, then a boolean is created, which is auto coerce to true and then 1, which I was not anticipating! 19:46
El_Che are you typing the variables?
oboe i can do subset Integer of Int where (IntStr) and then sub MAIN(Integer :$option = 0) {...} but surely there is an easier way? 19:47
19:48 jmerelo left
oboe i had assumed that sub MAIN(Int :$option) would suffice! 19:49
moritz wait what? which part isn't enough? 19:50
if you want 0 as the default, you need to make that explicit
timotimo Bool is Int in raku
oboe sub MAIN(Int :$option) is letting in booleans!
moritz sub MAIN(Int :$option = 0) { }
oboe I only want integers!
moritz m: say Bool ~~ Int 19:51
camelia True
MasterDuke btw, why wouldn't you just do `sub MAIN(IntStr :$option = 0) {...}`? why the subset?
oboe to get rid of booleans. The integer command-line argument is shortening strings and the boolean option is shortening them by 1 which I wasn't expecting to pass type checking 19:52
as I wasn't expecting booleans to be let through 19:53
timotimo m: @*ARGS = <--foo True>; multi sub MAIN(Int :$foo where none(Bool)) { say "yay" }
camelia Usage:
<tmp> [--foo=<Int where { ... }>]
timotimo m: @*ARGS = <--foo True>; multi sub MAIN(Int :$foo) { say "yay" }
camelia Usage:
<tmp> [--foo=<Int>]
timotimo ok that's not how 19:54
m: @*ARGS = <--foo=True>; multi sub MAIN(Int :$foo) { say "yay" }
camelia yay
19:54 rindolf left
timotimo m: @*ARGS = <--foo=True>; multi sub MAIN(Int :$foo where none(Bool)) { say "yay" } 19:54
camelia Usage:
<tmp> [--foo=<Int where { ... }>]
timotimo here it rejects the boolean
oboe Thanks timotimo!
That reads way better
timotimo you'll want to change the outut from the usage and hel though, since "where { ... }" looks weird 19:55
19:55 b2gills joined 19:57 sena_kun joined
oboe Yes, I had found that range checking and auto usage generation are less that awesome 19:58
19:59 Altai-man left
oboe timotimo: your solution seems to still let booleans in! 20:02
sub MAIN( Int :$trim-trailing where { none (Bool) } = 0) { say 'trim-trailing: ' ~ $trim-trailing.^name ~ ': ' ~ $trim-trailing.Numeric ~ ': ' ~ $trim-trailing; }
evalable6 trim-trailing: Int: 0: 0
oboe if I run it with: raku test.raku --trim-trailing 20:04
i still get:trim-trailing: Bool: 1: True
timotimo dang
20:07 grep0r left
MasterDuke m: @*ARGS = <trim-trailing>; sub MAIN( Int :$trim-trailing where { .^mro[0] !=== Bool } = 0) { say "trim-trailing: " ~ $trim-trailing.^name ~ ": " ~ $trim-trailing.Numeric ~ ": " ~ $trim-trailing; } 20:09
camelia Usage:
<tmp> [--trim-trailing=<Int where { ... }>]
MasterDuke m: @*ARGS = <--trim-trailing>; sub MAIN( Int :$trim-trailing where { .^mro[0] !=== Bool } = 0) { say "trim-trailing: " ~ $trim-trailing.^name ~ ": " ~ $trim-trailing.Numeric ~ ": " ~ $trim-trailing; }
camelia Usage:
<tmp> [--trim-trailing=<Int where { ... }>]
MasterDuke m: @*ARGS = <--trim-trailing=3>; sub MAIN( Int :$trim-trailing where { .^mro[0] !=== Bool } = 0) { say "trim-trailing: " ~ $trim-trailing.^name ~ ": " ~ $trim-trailing.Numeric ~ ": " ~ $trim-trailing; } 20:10
camelia trim-trailing: IntStr: 3: 3
MasterDuke oboe: does ^^^ work for what you want?
oboe MasterDuke: checking... 20:11
That works MasterDuke. It still seems an aweful mouthful for just specify Int only. 20:12
subset Integer of Int where (IntStr) 20:13
is also obtuse
MasterDuke true. but that's kind of how inheritance works, right? Bool is an Int. You're asking for an Int and you're getting one, you just happen to not like its flavor 20:14
have you looked at GetOpt::Long? the builtin CLI argument handling has some limitations (i don't remember why), but if you're willing to use a module it can do more 20:16
oboe Thanks MasterDuke. I really like the builtin stuff. It seems so close to perfect! 20:17
I last used GetOpt::Long with perl5.
I take it mro is for method resolution order?
MasterDuke i believe the raku version is pretty similar, but i personally haven't looked at the details in a long time
gfldex m: @*ARGS = <--trim-trailing=3>; sub MAIN( Int() :$trim-trailing ) { dd $trim-trailing } 20:18
camelia IntStr.new(3, "3")
MasterDuke yea. the ^ signifies it's a meta-method, i.e., operating on the objects metaclass
gfldex oboe: you didn't asked Rakudo for coercion. You just asked for a type check.
20:19 wamba joined
oboe yep, i was asking for type check and seeing what i thought was auto type coercion under the radar 20:20
which was subverting the type checking 20:22
I was expecting the following to fail type checking: 20:25
m: sub MAIN( Int() :$trim-trailing = 0) { say dd $trim-trailing; } 20:26
camelia 0
Nil
oboe m: @*ARGS = <--trim-trailing>; sub MAIN( Int() :$trim-trailing = 0) { say dd $trim-trailing; }
camelia Bool::True
Nil
20:28 aluaces left 20:29 aluaces joined
oboe Aside: What is the offline way to access docs? perldoc seems fubar. 20:31
I'm reduced to: raku --doc /usr/local/Cellar/rakudo-star/2020.10/share/perl6/site/doc/*.pod6 20:32
MasterDuke rakudoc is the successor to p6doc, but i'm not quite sure its status 20:34
oboe There is an equivalence between sum and [+]. Is there another for [*]?
MasterDuke nope
oboe okay. no rakudoc. is that a zef install. i'm with rakudo-star. 20:35
MasterDuke believe so 20:36
20:38 leont joined 20:51 brtastic left 20:55 defaultxr left 21:13 oboe left 21:18 Doc_Holliwood left 21:22 sena_kun left 21:23 _hxearth\` joined, _hxearth\` left 21:29 wamba left 21:45 wamba joined 21:48 wamba left 21:54 jmchael left
guifa2 what's the best way to approach compiling modules when they have enough files that it occasionally dings the open file limit? 21:55
timotimo get the lazy devs to hunt down the leaking file descriptors 21:56
guifa2 timotimo: I'm the developer :-)
I was just using zef install . to test installing 21:57
22:04 sxmx left 22:05 jmchael joined 22:06 sxmx joined 22:08 guifa2 left 22:40 jferrero left
El_Che if only the dev will hit it, raise your ulimit? 22:59
23:01 wamba joined 23:05 xelxebar_ left, xelxebar joined 23:11 wamba left 23:18 rypervenche left 23:41 rypervenche joined 23:48 leont left 23:57 jmchael left