[01:47] *** ilbot3 joined [06:49] <[Tux]> test 21.750 [06:49] <[Tux]> test-t 12.120 [06:49] <[Tux]> csv-parser 25.602 [06:50] <[Tux]> going the right direction again :) [07:23] *** RabidGravy joined [08:00] *** sortiz_ joined [08:47] nqp: 01fd41b | (Pawel Murias)++ | src/vm/js/ (2 files): [08:47] nqp: [js] Start working on compiling things that are not complete QAST::CompUnits. [08:47] nqp: [08:47] nqp: Pass the first 3 tests of t/qast/01-qast.t [08:47] nqp: review: https://github.com/perl6/nqp/commit/01fd41b8b6 [11:41] *** vendethiel joined [12:20] nqp: eea75dc | (Pawel Murias)++ | src/vm/js/ (3 files): [12:20] nqp: [js] Compile BVals outside of CompUnits. [12:20] nqp: [12:20] nqp: Pass more of the qast test. [12:20] nqp: review: https://github.com/perl6/nqp/commit/eea75dcebe [12:20] nqp: 250beb3 | (Pawel Murias)++ | src/vm/js/nqp-runtime/reprs.js: [12:20] nqp: [js] Make the REPR put methods on STable in setup_STable rather then an assortment of hacky places. [12:20] nqp: review: https://github.com/perl6/nqp/commit/250beb389a [12:55] *** lizmat_ joined [13:08] *** skids joined [13:19] *** d4l3k_ joined [13:22] rakudo/nom: 268e9c6 | lizmat++ | tools/build/makeMAGIC_INC_DEC.pl6: [13:22] rakudo/nom: More WIP on magical increment/decrement [13:22] rakudo/nom: review: https://github.com/rakudo/rakudo/commit/268e9c673e [15:50] nqp: d98d15d | (Pawel Murias)++ | src/vm/js/Operations.nqp: [15:50] nqp: [js] Implement nqp::stmts. [15:50] nqp: review: https://github.com/perl6/nqp/commit/d98d15d3ac [15:50] nqp: 9397ca1 | (Pawel Murias)++ | src/vm/js/Compiler.nqp: [15:50] nqp: [js] Ignore only known things in the pre_serialization list of things. [15:50] nqp: review: https://github.com/perl6/nqp/commit/9397ca11f0 [15:50] nqp: d32f936 | (Pawel Murias)++ | src/vm/js/Compiler.nqp: [15:50] nqp: [js] Set the context to null when compiling pre/post deserialization code. [15:50] nqp: review: https://github.com/perl6/nqp/commit/d32f936345 [16:38] *** Ven joined [16:53] nqp: 48fd5f9 | (Pawel Murias)++ | src/vm/js/Compiler.nqp: [16:53] nqp: [js] Support .resultchild on QAST::Stmt and QAST::Stmts. Implement having multiple children for QAST::Stmt. [16:53] nqp: review: https://github.com/perl6/nqp/commit/48fd5f91d7 [16:57] *** Ven joined [17:09] *** Ven joined [17:16] *** Ven joined [17:23] *** Ven joined [17:33] *** Ven joined [17:46] *** Ven joined [17:57] *** FROGGS joined [18:22] *** perlpilot joined [18:23] *** Ven joined [18:38] *** Ven joined [19:04] *** Ven joined [19:12] rakudo/nom: 5bd1e96 | lizmat++ | src/core/IO/ (2 files): [19:12] rakudo/nom: Implement $*DEFAULT-READ-ELEMS for reading [19:12] rakudo/nom: [19:12] rakudo/nom: Inspired by https://www.mail-archive.com/perl6-language@perl.org/msg35150.html [19:12] rakudo/nom: [19:12] rakudo/nom: Takes its default from the DEFAULT_READ_ELEMS environment variable, [19:12] rakudo/nom: or assumes 65536 if that cannot be found either. [19:12] rakudo/nom: [19:12] rakudo/nom: Since this number will both be used for string based reading (so the [19:12] rakudo/nom: number reflects the number of chars read, which can be more than 1 byte [19:12] rakudo/nom: per char), as well as binary reading, it felt right to call it ELEMS [19:12] rakudo/nom: rather than BYTES or CHARS. [19:14] hmmm.. looks like dalek got kicked? the commit is https://github.com/rakudo/rakudo/commit/5bd1e9655d5bd660f660 [19:16] hm, now how do we get this into moar's ops that do buffered reading? [19:18] lizmat: can you reproduce .IO.open.split("\n") being even slower than .IO.lines ? [19:18] * timotimo AFK [19:19] (if you need an example file, just run anything with --profile=heap and make sure it stops after 3 or 4 heap snapshots have been taken) [19:19] timotimo: maybe use "\x0a" instead of \n ? [19:19] in my case, split was 2x as fast [19:26] oh, huh. [20:10] lizmat: I think I'd rather the read size be a property of a handle [20:11] lizmat: Also it perhaps wants to be in bytes if we're going to use it as the buffer size we pass "all the way down" [20:11] well, it was something *i* could to easily to get the discussion going :-) [20:11] having it at VM level on a handle would be excellent :-) [20:12] Aye, it's not in the language until it's in a released test suite, so no panic. :-) [20:12] indeed :-) [20:14] The mailing list discussion kinda suggests a default set in PROCESS and so overridable by a dynvar is also valuable though :) [20:15] But yeah, I think it'll wanna be bytes, since those are what you actually request from the operating system. [20:21] jnthn: but readcharsfh needs a number of chars [20:21] but maybe that should default to whatever we have for bytes [20:22] there's a difference between how many bytes to request from the OS to read into our internal buffers and what we want to have returned right-now [20:22] What's readcharsfh used for though? Not slurp nor .lines nor .get... [20:22] I guess for things that have their own buffer? [20:23] jnthn: yeah, things like .words .split .comb [20:23] Also, d'oh, I somehow thought today was Tuesday, not Wednesday, 'cus of the bank holiday... [20:24] my attempts at making p6sink cleverer about unless/if with only two children is breaking the restricted setting :< [20:24] Guess I'll see about mostly stealing tomorrow for Perl 6 things then :-) [20:24] .oO( If I'm not smart enough to know what day it is, maybe it's a good thing I didn't hack on Perl 6 stuff :P ) [20:24] um. no. it's breaking anything that runs [20:25] timotimo: Congrats, or something :P [20:27] *** Skarsnik joined [20:29] the things run to completion, apparently, and then i get Cannot look for method 'sink' on a null object [20:29] and --ll-exception doesn't get me more detail on that >_> [21:33] rakudo/nom: ae3a791 | lizmat++ | tools/build/makeMAGIC_INC_DEC.pl6: [21:33] rakudo/nom: More WIP on magic decrement [21:33] rakudo/nom: [21:33] rakudo/nom: Turns out we need a hash on boundary condition anyway. [21:33] rakudo/nom: review: https://github.com/rakudo/rakudo/commit/ae3a79129c [21:35] *** AlexDaniel joined [21:50] *** pouledodue joined [22:14] *** pouledodue left [22:21] *** skids joined