|
weekly Perl 6 status meetings with phase transitions: Tue 19:00 UTC | IR clogs at irclog.perlgeek.de/phasers/today Set by moderator on 14 June 2011. |
|||
|
14:14
colomon left
18:12
masak joined
|
|||
| masak | :) | 18:12 | |
| pmichaud | my report for 2011-06-21: | 18:15 | |
| What I did: | |||
| nom stuff: | |||
| * generic nqp::op namespace | |||
| * lazy lists, including infinite lists and iterators | |||
| → created opcodes for fast creation of lists and iterators | |||
| → binder uses same functions for creating slurpies | |||
| * PAST::Want nodes so that int/num constants to pir:: and nqp:: opcodes aren't boxed when they don't need to be | |||
| * pir::const:: for access to pasm symbolic constants | |||
| * Reorganized the locations of operators in source tree | |||
| * new Range type, super-fast for generating numeric ranges | |||
| * gather/take | |||
| * Enum/Pair | |||
| * List.map, for() loops | |||
| * hash methods (.pairs, .kv, .keys, .values, { }, etc.) | |||
| * Refactored to use the new &say and .gist spec as given in perl6/specs rev 669e0dd | |||
| * preliminary list slices (still need Positional role) | |||
| * return, return-rw, take, take-rw, next, last, redo, die, fail (cheat) | |||
| * Cleaned up Stringy/Str | |||
| * term:<time> | |||
| * Nil | |||
| * Added DUMP() which lets us see a more detailed dump of a datastructure | |||
| nqp stuff: | |||
| * added nqp::opcodes | 18:16 | ||
| * added PAST::Want node type for selecting PAST based on context | |||
| * updated some error messages | |||
| master stuff: | |||
| * enabled register sharing via PAST::Stmt | |||
| * draft release announcement | |||
| What I plan to do: | |||
| * convert CORE.setting to use nqp:: instead of pir:: (where possible) | |||
| * performance improvements to list handling | |||
| * more CORE.setting methods and functions | |||
| * implement LoL | |||
| * Work on S07 redraft | |||
| EOR | |||
|
18:23
kboga joined
|
|||
| masak | pmichaud++ | 18:26 | |
| moritz | pmichaud++ indeed | 18:29 | |
| My report: implemented more stuff in nom | 18:30 | ||
| the usual back-and-forth spec discussions | |||
| rest: same as last week | |||
| EOR | |||
|
18:40
colomon joined
|
|||
| Util | pmichaud++ | 18:48 | |
| Pre-report: | |||
| # New RC Perl 6 solutions: | |||
| * rosettacode.org/wiki/Levenshtein_distance | |||
| * rosettacode.org/wiki/Partial_functi...pplication | |||
| * rosettacode.org/wiki/Character_matching | |||
| * rosettacode.org/wiki/Sort_disjoint_sublist (Added 1-line solution) | |||
| # Plan to do: | |||
| * Finish YAPC::NA::2011 talk | |||
| * $WORK | |||
| EOR | |||
| sorear | pre-report: | 18:51 | |
| * finished removing Niecza's grammar's own symbol table, so now the compiler only has *one* | 18:52 | ||
| * made many small improvements to "declared but not used" checking, in particular it works with top-level definitions and operators now | |||
| * added is tighter/looser/equiv | 18:53 | ||
| * changed 'subset', 'enum', and 'constant' to default to 'our' | |||
|
18:54
colomon left
|
|||
| sorear | * fixed class A::B { ... }; class A::B { }. about 3 times, infact :( | 18:54 | |
|
18:54
mberends joined
|
|||
| sorear | * added checks for duplicate definitions of attributes, methods, and our symbols, with source locations | 18:54 | |
|
18:55
colomon joined
|
|||
| sorear | * started trying to add MY:: et al, which has turned into a major refactor of package handling (but it's a huge simplification) | 18:55 | |
| EOR | 18:57 | ||
|
18:57
colomon left
|
|||
| mberends | sorear++ | 18:58 | |
| masak | pre-report: arranged an NPW with jnthn++. gave two (reprised) talks during it. patched the nom branch for the first time. edublogged, but missed yesterday. will make up for it today. EOR | 18:59 | |
| moritz | sorear++ masak++ | 19:02 | |
| masak | o/ | ||
| moritz | it's #phasers time - any other reports? | ||
| Util | o/ | ||
| mberends | o/ but nothing to report | ||
| sorear | o/ | ||
| TimToady is mostly working(ish) on a 5.14 Camel... | |||
| sorear is still backlogging #perl6 and #phasers | |||
| sorear is now done | 19:03 | ||
| mberends | TimToady: as you're the original p6'er, is 5.14 weirdly nostalgic for you? | 19:04 | |
| masak | I'd like to know what the probability is of nom merging before release. :) | ||
| pmichaud | How many times do you look at something in 5.14 and say "oops"? | ||
| moritz | masak: 0% | 19:05 | |
| pmichaud | masak: I guarantee that nom will merge before a release. We just don't know which one yet. :) | ||
| PerlJam | I just want to know if the 5.14 book is full of references to Perl 6 ;) | ||
| mberends | TimToady: (well, you were the original p5'er too, etc etc :) | ||
| TimToady | mberends: it's kind of a drag, really | ||
| pmichaud | nom won't make it into the June release, for certain. I'm not sure where we'll be for July. | ||
| TimToady | PerlJam: we're trying to limit P6 refs to things that are actually going to be considered informative to a P5 programmer :) | 19:06 | |
| pmichaud | We still have a ways go go, since we don't have Test.pm running yet (which jnthn++ and/or I expect to have working by this weekend) | ||
| PerlJam | So no "smart matching works like this in 5.14, but in Perl 6 it's much better..." ? :) | ||
| TimToady | PerlJam: that can fall in the category of "this works differently because..." | ||
| anyway, not trying to hijack the meeting | 19:07 | ||
| masak | pmichaud: thanks for being epsilon informative :P | ||
| pmichaud | masak: glad to help! :) | ||
| TimToady | pmichaud: it's usually worse than "oops" :) | 19:08 | |
| sorear | niecza-v7 release on Monday. I'm starting to get kinda annoyed by all the missing features in the v6 which I'm currently using for the build :) | ||
| I've also been thinking about how to write a Perl 6 interpreter without requiring a bootstrap at some level | 19:10 | ||
| tadzik | not late! | ||
| sorear | tadzik: ? | 19:11 | |
| moritz | tadzik: you're within c.t. :-) | ||
| sorear | the main interesting issue is "augment slang" (equivalently, is parsed macros) in user code, which requires the user Perl 6 code to be parsed using the user grammar engine | 19:12 | |
| tadzik | I had some report, but I eated it :) | ||
| sorear | but it seems that a pure C Perl6 could use its own grammar engine to parse, without a setting | ||
| also figuring into this thought is TimToady's comment that the setting doesn't need to be a Perl 6 module | 19:13 | ||
| TimToady | it just has to look like an outer scope by the time you get there | 19:14 | |
| moritz | sorear: it just has to try hard enough to look like one :-) | ||
| pmichaud | Rakudo long had a non-p6 setting, fwiw. :) | ||
| sorear | anyway this set of ideas is making me think I can build a Perl6 that will compete with "time perl -E 'say 2 + 2'" | 19:15 | |
| moritz | which worked fine, as long as nobody tried to introspect it | ||
| sorear | and I want to try, although it seems kinda silly to start a new implementation before niecza is done :D | 19:16 | |
| mberends | we tend to think that done === dead, so long live niecza! | 19:17 | |
| PerlJam | another implementation of Perl 6 can only improve the chances that Perl 6 will get used ;) | 19:18 | |
| masak | sorear: it's a simple cost/benefit calculation. will starting the new implementation bring more boons than focusing on niecza? | 19:19 | |
| TimToady | and can you trick someone else into taking it over? :) | 19:20 | |
| sorear | The v6 announcement made it to reddit, where the overwhelming response was "the name sucks". I'm sort of tempted to agree... | ||
| PerlJam | sorear: rebranding is always an option | ||
| moritz proposes "the compiler formerly known as niecza" as new name :-) | 19:21 | ||
| sorear | masak: not now | ||
| tadzik | akaniecza? :) | ||
| masak | sorear: if the schedule is "not now", I'm interested in contributing. :) | 19:22 | |
| sorear | the New Design wants a low-level language with good coroutine support, extremely fast loading, dynamic code generation, and an efficient 6model-integrated garbage collector | ||
| masak | (I have tuits, but not now.) | ||
| sorear | "in other words, it wants M0" | ||
| PerlJam | sorear: Just call it "hakuna matata" | 19:23 | |
| mberends | TimToady: niecza without sorear++ is unthinkable today, like pugs without au++. I hope we'll all do whatever we can to increase its bus number to avoid a similar fate. | 19:24 | |
| jnthn | oh hai | 19:25 | |
| Sorry I'm late | 19:26 | ||
| mberends | o/ jnthn | ||
| masak | jnthn! \\o/ | ||
| jnthn is in Gothenburg :) | |||
| sorear | o/ jnthn | ||
| Is there any difference whatsoever between $a and $::a? | |||
| moritz | any more #phasers topics for tonight? | 19:29 | |
| sorear | hmm | ||
| don't think so | |||
| moritz | tadzik: do you have a short gsoc report for us? | 19:30 | |
| jnthn | To report, breifly: did stuff in nom. Will now focus on native types, roles, etc. | ||
| tadzik | moritz: not really. Me report is: had two exams, these are not the results I'm looking for | 19:31 | |
| sorear | I want to understand leading ::, roles, native types better | 19:33 | |
| but I think those are not #phasers topics | |||
| jnthn | sorear: Not sure what the leading :: is about there | 19:34 | |
| Just 'cus std parses it doesn't mean it means anything ;) | |||
| masak | isn't leading :: for being able to interpolate strings into variable names? | ||
| sorear | jnthn: S02 talks about leading :: in vagueish terms | 19:35 | |
| jnthn | ah | ||
| jnthn will look a little later | |||
| sorear | AIUI, leading :: means nothing in sigilled names, and in unsigilled names it means "treat this syntactically as a type name, even if no such type was declared" | ||
| jnthn | In a signature it's a type capture | 19:37 | |
| I know about that case | |||
| masak | :) | ||
| jnthn | it's the sigil'd one that looked weird to me :) | ||
| maybe "ignore it" is fine though :) | |||
|
20:16
masak left
21:02
colomon joined
22:00
tylercurtis left,
tylercurtis joined
22:20
[particle]1 joined
22:22
[particle] left
|
|||
| colomon | sorear: where does big_plus (for example) get used? My greps aren't turning it up anywhere.... (I've written a preliminary bif_pow, still needs work.) | 23:05 | |
| afk # don't know why I'm being called... | 23:06 | ||
| sorear | colomon: NieczaPassSimplifier | 23:07 | |
| though that was probably a mistake in retrospect | |||
| colomon | .man or .pm6? | 23:08 | |
| *.nam | |||
| .pm6, I see. trying now. (But apt to be called away in a moment.) | 23:12 | ||
| sorear | sorry I'm so unavailable | 23:22 | |
| it goes like this | |||
| 1. CORE.setting contains a definition of &infix:<+>. | |||
| 2. Direct calls (not via higher-order-functions) to that are rewritten by the Simplifier into a Op::Builtin(name => 'plus') node. | 23:23 | ||
|
23:23
kboga left,
kboga joined
|
|||
| sorear | 3. Op::Builtin generates a low-level code node of type bif_plus. | 23:23 | |
| 4. Low-level nodes of unknown type default to calling methods on Builtins. | |||
|
23:48
mberends left
|
|||
| colomon | ah. I've definitely done something wrong with that, then. alas, I need to do some $work atm instead of straightening it out. | 23:58 | |