Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2007-01-14

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:02 Psyche^ joined perl6
00:05 andara left perl6
00:06 prism joined perl6
00:18 Psyche^ is now known as Patterner
00:25 nekokak joined perl6
00:44 mako132_ joined perl6
00:47 jferrero joined perl6
00:54 lisppaste3 joined perl6
00:58 yaxu joined perl6
00:59 Psyche^ joined perl6
01:13 Limbic_Region joined perl6
01:16 Psyche^ is now known as Patterner
01:29 yaxu left perl6
01:35 Schwern joined perl6
01:47 buubot joined perl6
01:48 mako132_ joined perl6
02:24 mako132_ joined perl6
02:33 takesako__ joined perl6
02:34 buubot joined perl6
02:46 buubot joined perl6
02:56 neonse left perl6
03:04 Schwern joined perl6
03:04 cmarcelo joined perl6
03:06 trym joined perl6
03:06 gnuvince joined perl6
03:11 ayrnieu joined perl6
03:50 pbuetow joined perl6
04:00 ProperNoun joined perl6
04:11 b00t joined perl6
04:14 nipra joined perl6
04:14 svnbot6 r15070 | cmarcelo++ | * MO: More stuff for autoboxing (in MO/si.hs), now accepting
04:14 svnbot6 r15070 | cmarcelo++ | monadic functions in the mkBoxClass constructor too (via new
04:14 svnbot6 r15070 | cmarcelo++ | syntax sugar).
04:14 avar ?eval 1
04:15 evalbot_r15069 1
04:49 spx2 joined perl6
05:04 b00t joined perl6
05:29 svnbot6 r15071 | Darren_Duncan++ | ext/Rosetta/ : minor doc edits, and copyright date updates ... nothing worth mentioning in the ChangeLog
05:30 Psyche^ joined perl6
05:39 pbuetow joined perl6
05:45 Psyche^ is now known as Patterner
06:14 BooK_ joined perl6
06:30 leed joined perl6
06:38 luqui joined perl6
06:45 rhizo joined perl6
07:07 trym joined perl6
07:12 nipra_ joined perl6
08:02 dduncan left perl6
08:07 nipra_ is now known as nipra
08:33 Hgud is now known as greenslap
08:56 elmex joined perl6
08:58 iblechbot joined perl6
09:15 miyagawa joined perl6
09:18 trym joined perl6
09:20 AlexLibman joined perl6
09:32 andara joined perl6
09:48 daddye1 joined perl6
09:48 daddye1 hello
09:49 daddye1 left perl6
10:30 andara joined perl6
10:41 sunnavy joined perl6
10:42 spx2 joined perl6
10:50 sunnavy joined perl6
11:14 sunnavy joined perl6
11:49 sunnavy joined perl6
11:52 GabrielVieira joined perl6
11:59 elmex joined perl6
12:25 iblechbot joined perl6
12:26 GabrielVieira2 joined perl6
12:31 larsen_ joined perl6
12:33 andara joined perl6
12:36 chris2 joined perl6
12:37 CardinalNumber joined perl6
12:52 dmq joined perl6
12:59 ozo joined perl6
13:10 neonse joined perl6
13:22 buetow joined perl6
13:27 buetow joined perl6
13:31 ProperNoun joined perl6
13:38 trym joined perl6
13:43 trym_ joined perl6
14:00 bernhard joined perl6
14:13 Limbic_Region joined perl6
14:39 nipra joined perl6
15:01 nipra_ joined perl6
15:23 jan joined perl6
15:27 spx2 joined perl6
15:47 jan joined perl6
15:53 xinming_ joined perl6
16:00 nipra_ is now known as nipra
16:18 spx2 joined perl6
16:19 spx2 joined perl6
16:27 spx2 joined perl6
16:51 spx2 joined perl6
16:52 spx2 joined perl6
16:53 spx2 joined perl6
16:54 spx2 joined perl6
17:01 spx2 joined perl6
17:02 andara joined perl6
17:04 spx2 joined perl6
17:05 spx2 joined perl6
17:11 glasser joined perl6
17:15 spoops joined perl6
17:17 marmic joined perl6
17:26 luqui joined perl6
17:43 cmarcelo joined perl6
17:45 zekvent joined perl6
17:52 weinig|bbl is now known as weinig
17:54 ktne joined perl6
18:10 ludan joined perl6
18:14 ktne hello
18:14 ktne is there any IDE supporting perl6 for win32?
18:17 cmarcelo ktne: Not exactly what you're asking, but there are Perl 6 syntax files for Emacs and VIM.
18:18 ktne hmm, i was thinking more about something like integrated debugger and such
18:18 ktne and not emacs :)
18:24 weinig is now known as weinig|food
18:29 Vex joined perl6
18:34 svnbot6 r15072 | kudra++ | The summary
18:52 rdice joined perl6
18:56 luqui joined perl6
18:56 luqui_ joined perl6
19:04 ozo joined perl6
19:07 [particle] joined perl6
19:10 weinig|food is now known as weinig
19:33 elmex joined perl6
19:40 araujo joined perl6
19:51 ProperNoun joined perl6
20:06 mncharity joined perl6
20:12 mncharity is now known as putter
20:14 Limbic_Region joined perl6
20:30 justatheory joined perl6
20:34 putter anyone around?
20:34 * putter looks for insight into "the path to Christmas"
20:35 * allbery_b is around but not useful for that question
20:36 allbery_b (in any sense...)
20:37 * Limbic_Region is far too around
20:38 putter allbery_b: hi.  are you a perl person?  haskell?  p6? ;)
20:38 * Limbic_Region shudders to think of how much weight he has gained since getting married
20:38 putter hi LR
20:38 Limbic_Region salutations putter
20:38 allbery_b I've done a little bit of stuff with p6 and am learning haskell.  wouldn't have a clue about roadmaps or etc.
20:39 putter I'm trying to figure out what the state of the (p6) universe is...
20:39 allbery_b ...and have been reticent abotu touching the tree because I don't really know what's up with mp6/kp6/etc. (i.e. is the haskell tree now a lame duck?)
20:40 Limbic_Region putter - from what perspective, spec or implementation or some other vantage point all together?
20:40 putter allbery_b: (potential developer survey:)  do you have a commit bit?  if you had say an hour to do some coding, any idea where you would look for tasks to do?
20:41 * Limbic_Region knows the answer to that
20:41 * Limbic_Region keeps silent until asked though
20:41 putter LR: big picture project management
20:41 Limbic_Region putter - I think this xmas is the xmas
20:42 Limbic_Region the spec is converging more rapidly
20:42 putter allbery_b: re "...and[...]duck?", very useful observation
20:42 Limbic_Region putter - but I think you should likely be talking to someone from any(@Larry)
20:43 * allbery_b honestly doesn't know
20:43 putter sounds like the "don't worry about breaking the tree" property doesn't exist anymore?
20:43 allbery_b mp6 and kp6 are steps somewhere in making perl6 self-booting.  is the haskell version part of that chain, or just aprimordial botstrap that will be discarded?  I've never known
20:44 lumi I think the tree is Audrey-insured, less Audrey means less tree breaking
20:44 allbery_b or are haskell-pugs and mp6 and/or kp6 parallel tracks?
20:44 putter lumi: makes sense
20:44 Limbic_Region allbery_b no one is putting all their eggs in one basket
20:45 lumi So afaict mostly there's just fglock moving [mk]p6 along
20:45 lumi atm
20:45 * Limbic_Region AFK &
20:45 putter tnx LR
20:46 putter lumi: hmm...
20:46 lumi One thing I thought about touching is the 1..10:by(5) parsing, but I don't know enough to do it
20:46 putter allbery_b: any other "things you've wondered about"?
20:47 allbery_b not really; been concentrating on becoming more proficient in haskell
20:47 putter k
20:47 putter lumi: it's haskell side?
20:47 lumi Yeah, the haskell parser
20:48 putter any idea the state of Moose?
20:48 allbery_b I did poke at the parser a bit (misparsing of file tests) a couple months ago, swing and a miss, decided I needed to understand what was going on better first --- hence the focus on haskell
20:49 putter ah...
20:49 allbery_b (specifically:  "- f" is parsed as "-f" currently, because the guts of the parser ignore whitespace)
20:50 lumi I don't know about Moose, I think stephan, nothingmuch and gaal are prime vict^Winterviewees for that
20:50 putter :)
20:50 lumi allbery_b: Where does this happen? test? I can have a look?
20:51 allbery_b there's a test for it at the bottom of the file tests, yes.  what happens is that when a '-' is seen it goes back to check for file test ops, but reenteringt he parser discards leading whitespace.  I wasn't able to figure out how to avoid it
20:52 * allbery_b no longer recalls the chain of calls he was playing with that handled filetest parsing
20:54 putter one possibility might be to move it to p6 side.  sub prefix:<-f>(mumble){Pugs::Internals::filetest_f(mumble)} or
20:54 putter some such.
20:55 putter checking to see if (1) the p6 version works, and (2) it still works if placed in the Prelude.
20:55 allbery_b Pugs/Parser/Operator.hs
20:56 allbery_b the actual problem is that the parser beng used is following haskell rules, so e.g. "-f" can't be parsed as a single symbol --- Haskell doesn't permit punctuation and word characters to be combined in a symbol.  (This is also the root cause of Haskell's odd parsing wrt negative numbers.)
20:56 Aankhen`` joined perl6
20:57 allbery_b so the low level lexical analysis sees e.g. "-f" and returns the "-", another token must be read to get the "f" and whitespace will be discarded automatically when that is done
20:58 * allbery_b could try switching the low level stuff to use different rules but is terrified of what that might break...
20:59 allbery_b since it can'
20:59 allbery_b t be done on a case-by-case basis easily.  although if I've picked up enough haskell and parsec guts it might be possible to come up with a special case
21:11 * putter wonders if it would help to have one day a week which is _the_ day for pugs development.  some time when one can count of everyone being available.
21:12 andara joined perl6
21:13 allbery_b right, Pugs/Lexer.hs, definition of "symbol" needs a special case for first character (only!) being '-', followed by a letter, followed by a non-word character of some kind.
21:13 allbery_b I might actually be up to that now...
21:13 lumi Oh, I see
21:13 allbery_b note that the existing special casing works anywhere inside a "token" as far as I can tell
21:13 lumi Are you doing that?
21:14 allbery_b not yet, I've just been retracing my tracks
21:14 avar When I do \x00..\x1f in rule {} will that always mean that range in ascii?
21:14 lumi It does say "horrible, horrible kluge"
21:14 lumi But I don't understand the bit about prec levels
21:16 weinig is now known as weinig|bbl
21:16 gaal uh, what is this -f whereof you speak? I don't think it means the same thing as it did in p5
21:16 lumi gaal:  No?
21:16 gaal anyway: the convention is ruleX is a token parser (ie, takes following whitespace), whereas verbatimX does not.
21:17 gaal there was a thread on this subject on p6 with a lot of contention, but it ended with a ruling
21:17 gaal unfortunately, i remmeber neither the actual ruling nor the title of the thread :)
21:18 lumi makeOp1 in Pugs/Parser/Operator.hs for those following at home
21:18 lumi It parses the "-" as a token, and then tries to parse which filetest it is if any
21:18 lumi So it will have eaten the whitespace already
21:19 cmarcelo putter: did you heard of parrot the (montly) Parrot Bug Day? maybe something like that for the rest of p6 (pugs, mp6, etc) would be what you want...
21:19 lumi symbol has neither "ruleX" nor "verbatimX" form
21:19 gaal yeah, that's a parsec-originated combinator
21:21 lumi But it can't be right there.. -X should be a token
21:21 allbery_b that's far too late
21:21 allbery_b the damage is done when name <- symbol returns the "-"
21:21 gaal I still think -X doesn't work like it did in p5
21:21 lumi Exactly
21:22 allbery_b symbol needs to recognize the filetest ops itself,m which will also fix the (broken) "horribe, horrible kluge"
21:22 gaal partly because Larry wanted to avoid the special moose
21:23 gaal zz&
21:23 lumi I think it isn't symbol's job
21:23 * allbery_b is hacking Pugs/Lexer.hs:symbol to parse '-' + letter + nonletter
21:23 lumi Night gaal
21:23 allbery_b but if you dont' do it in symbol then whitespace will be able to intervene
21:23 gaal I think you should find the spec :) re-z&
21:24 lumi I think special-casing before running deciding what to run symbol on would be better
21:24 allbery_b the filetest ops break the assumption that a punctuation character follwoed by a word character makes two tokens
21:24 putter cmarcelo:  i had not heard.  interesting idea.  the motivation is that as people (eg, audreyt) have less time available, rather than scattering through calendar time, and things bogging down, perhaps it could be concentrated, providing normal, albeit 1/7 speed, progress.
21:24 allbery_b and symbol is already filled with special cases of that variety...
21:25 lumi allbery_b: My reading of symbol is, it will parse any string you give it, and then check whether there's a boundary
21:25 allbery_b yeh
21:26 lumi Asking symbol to parse "-" and having it return "-f" instead might break things
21:26 allbery_b which is where it breaks as '-' followed by a word character is a break by its rules
21:26 allbery_b so that has to be suppressed
21:26 lumi allbery_b: Because makeOp1 asks it to parse "-"
21:26 allbery_b hm, it shouldn't be
21:26 lumi But makeOp1 knows it's doing something iffy, because it checks in lookAheadLiterals for the "-" case
21:27 allbery_b no, makeOp1 passes it e.g. "-f".  symbol decides that that doesn't make sense because it's punct followed by word and returns "-"
21:27 * allbery_b went through that last time he did this
21:27 allbery_b hence the horrible hack
21:27 lumi Are you positive? That's not what it looks like to me
21:28 allbery_b I banged my head against it for close to a week last time before I worked it out.
21:28 putter or some other pattern instead of once a week.  say particular hours on MWFSS?
21:28 allbery_b the "name" parameter to makeOp1 is the full operator, e.g. "-f"
21:28 lumi The hack specifically asks whether the parameter name, which is passed to symbol, is *exactly* "-"
21:28 allbery_b and that's what is passed to symbol
21:28 allbery_b and symbol parses that, then decides it makes no sense and returns "-"
21:29 allbery_b which si then trapped and the parser is reentered to get the rest of the operator name, but than eats whitespace
21:29 allbery_b because symbol checks for punct followed by word and says "wait, that can't be a single token"
21:29 allbery_b and returns the punct
21:30 luqui joined perl6
21:30 lumi name is a parameter for makeOp1; it's passed to symbol, which tries to parse that string, making sure there's a word boundary, and then passes it to, er
21:30 lumi It doesn't pass it (sorry) it returns it but it's ignored
21:31 lumi symbol first of all parses all of the string, and then compares the next character to the string's last character
21:31 lumi It doesn't look inside the string at all other than that
21:33 * allbery_b sighs
21:34 lumi allbery_b: Where in symbol do you see it looking inside s?
21:34 allbery_b so, I'm back to being thoroughly confused as to how file test ops are parsed at all
21:34 ruoso joined perl6
21:34 lumi I'm confused too why makeOp1 gets "-" as the name
21:34 cmarcelo putter: what's MWFSS (mon wed fri sa su)?
21:34 allbery_b or perhaps I should say "why does it bother to define them as operators since the claim is that it never *parses* them as operators, but only a shacks appended to "-"?
21:35 putter cmarcelo: yes
21:35 lumi I think just moving the "Are we in '-'?" check to outside the call to symbol might make things better
21:36 allbery_b oh, *that's* what the comment means
21:37 allbery_b unary '-' has hgher precedence than filetest ops, I think, so it wins
21:37 lumi What does it mean?
21:37 lumi Oh..?
21:38 allbery_b basically when it's parsing it recognizes '-' before e.g. '-f'
21:38 luqui you could do a negative lookahead... that's ugly though
21:39 allbery_b but once again that leads back to symbol, because it eats trailing whitespace immediately
21:40 cmarcelo putter: I think starting simpler (with a day or weekend) like Parrot maybe a good way to measure if this "arrangement" will work.. IMHO, as important as this is try to keep STATUS/TASKS files (or their substitutes) more "alive"..
21:40 allbery_b so you can't check for it after it's parsed '-'
21:40 luqui allbery_b, well, you could "try" before parsing the -
21:40 putter cmarcelo: agreed and agreed
21:40 luqui btw, what parser are we talking about?
21:40 lumi That wouldn't uneat the whitespace
21:41 allbery_b luqui: no, we'd need a custom case to parse with char '-' instead of with symbol, because with symbol we can never learn tif there is followeing whitepsace or not
21:41 lumi Okay, I have some kinda something
21:42 lumi Where do we nopaste?
21:42 luqui allbery_b, oh, I see the problem.  you're having trouble with automated parser generation
21:42 allbery_b the parser in question is Pugs/Parser/Operator.hs, where filetest ops currently don't parse quite right ("- f"is parsed as th filetest op "-f")
21:42 luqui i mean, given a list of operators, parse them
21:43 luqui this is inside a OPP parser, not a top-down, right?
21:43 luqui (well, the OPP is inside the top-down)
21:44 larsen_ joined perl6
21:44 allbery_b erm, tis is not debugging a parser written in pugs; this is debugging pugs' own parser
21:44 luqui hmm, that's weird; doesn't look like it should do that
21:44 lisppaste3 lumi pasted "Some kind of fix?" at http://paste.lisp.org/display/35292
21:45 * luqui understands
21:45 lumi Does this look vaguely correct?
21:45 allbery_b not sure.  I was angling toward ``char '-''' though which amoutns t the same thing (except possibly safer, haven't checked to see if string eats whitespace like symbol does)
21:46 lumi It doesn't
21:47 lumi It's Parsec's string
21:47 allbery_b oh wait, that's not going to work anyway
21:47 lumi What isn't?
21:47 cmarcelo lumi: what the guard C<< | "-" <- name = >> means? (right in the beginning of lookAheadLiterals defition)
21:47 lumi cmarcelo: It's a pattern guard, and it was already there
21:48 allbery_b the problem is that "-f " or "-f(" should be a filetest op, but "-foo" should be the value of "foo" negated
21:48 CardinalNumber joined perl6
21:48 allbery_b although the existing code might deal with that properly
21:49 lumi allbery_b: You're right, my fix is wrong
21:49 lumi m
21:49 * luqui can understand why it might misparse "-foo"
21:49 luqui but not why it would misparse "- f"
21:49 lumi Possibly not? parseFileTestOp checks the following
21:50 lumi luqui: It tries to parse "-", kills any whitespace, then comes back to us
21:50 luqui oh, I can see how it might do that
21:50 lumi Where we try to see if it's a filetest
21:50 * allbery_b is trying it anyway just to see what breaks, but seems to recall something like that causing passive parse breakage
21:50 lumi allbery_b: No, I think my fix *is* right
21:51 allbery_b er, *massive* parse breakage
21:51 lumi Haha
21:51 lumi I was wondering about that passive breakage :)
21:52 lumi Okay, I've got to go
21:52 lumi zz&
21:55 * allbery_b sighs as pugs decides it must rebuild the known universe --- this gonna be slow
21:55 jferrero joined perl6
21:58 Aankhen`` joined perl6
21:58 cmarcelo considering C<name> is a String in makeOp1, anyone knows what does << "-" <- name >> means?
22:01 larsen__ joined perl6
22:02 allbery_b ``| val <- expr'' is a pattern match guard
22:03 allbery_b the val is actually a pattern, if the pattern matches the guard succeeds
22:03 luqui cmarcelo,  it means if name == "-"
22:03 allbery_b except it's faster because it uses a pattern match
22:03 luqui i.e.  | "-" <- name = foo   -- returns foo if name == "-"
22:03 allbery_b (and you can use other pattern matching tricks but the code in question doesn't)
22:04 allbery_b *sigh* right, I'd svn up'd.  this is going to take forever
22:04 allbery_b (g4 macmini...)
22:05 cmarcelo Thanks, I see know. I didn't knew this syntax, my head was focused on "<-" as monadic sugar only..
22:06 luqui allbery_b, aye, I had a g4 two years ago at the hackathon
22:06 luqui it was unreasonable to do pugs development on it
22:06 luqui I ended up sshing to audreyt's machine
22:07 allbery_b pattern match guards are an extension, IIRC, so you might not have runinto them
22:07 luqui yeah, not part of haskell 98
22:07 allbery_b (in fact I googled and got SPJ's paper intoducing the extension)
22:09 cmarcelo Yep, it's an extension, for GHC it's described in User Guide, section 7.3.2.
22:10 allbery_b it's actually related to <- syntax in do, btw, since that is also a pattern match
22:10 allbery_b (which invokes "fail" if the pattern doesn't match)
22:12 allbery_b it's just that in most cases it's used in the trivial case x <- expr which will never fail
22:14 cmarcelo They are related but in Pattern Guards both side have the same type, while in list comprehensions (and do-notation), that's not the case..
22:15 cmarcelo That was confusing to see at first time.
22:15 luqui cmarcelo, well, it's almost the case.  pattern guards you have a <- a; do notation and list comprehensions you have a <- m a
22:15 luqui they still have to match; there's just an "m" in front sometimes :-)
22:16 cmarcelo :)
22:22 cmarcelo the problem with today's code is that it parses "- f " as being -f operator and it should instead parse as being "f" negated? or there are more problems?
22:23 allbery_b that's the problem, there's already a parse test for it
22:23 allbery_b I tried to attack it a couple months ago, and decided I needed to learn more haskell first
22:23 allbery_b so I've been poking at it again (nd re-confusing myself over the slightly bizarre interactions in lexing and parsing)
22:25 cmarcelo where in t/ is the test for this?
22:29 allbery_b syntax/parsing/filetest.t
22:29 allbery_b (sorry, machine being hammered flat by pugs compile :)
22:32 cmarcelo tks.
22:39 Dana77 joined perl6
22:39 Dana77 left perl6
22:41 cmarcelo allbery_b: why would the propose fix misparse "-foo", or it's just speculation?
22:41 cmarcelo *proposed
22:41 allbery_b just speculation
22:42 allbery_b I added some tests to the local t/syntax/parsing/filetest.t to check for it, when the build finishes (someday...)
22:43 cmarcelo ok. (reading here I didn't get why it would fail)
22:44 trym__ joined perl6
22:46 allbery_b basically, for the same reasn "- f" does:  because something might recognize an identifier first (although it arguably "shouldn't")
22:47 * allbery_b is just being paranoid because the parser keeps confusing him every time he tries to untangle it
22:49 cmarcelo (I tried here and I'm getting the << Unexpected " f" >>)
22:52 cmarcelo looks like, if it's not a file operator, we need to consume the whitespace as the rest of code was expecting..
22:54 putter I'm trying to understand p6 project bottlenecks.  Ie, a partial answer to "why isn't p6 done yet?".
22:55 larsen_ joined perl6
22:55 putter Anyone have thoughts?
22:55 cmarcelo allbery_b: actually I missed a "whiteSpace" call in the original fix.. re-testing
22:56 putter Actually, a better motivating questions is "why isn't p6 progressing more rapidly?".
22:56 allbery_b "because audreyt is the motivator and she's being attacked by real life(tm)"?
22:57 allbery_b well, not really
22:57 putter one thing mentioned has been lack of transparency and out of date task lists, leading to people not knowing what to do and drifting off.
22:57 allbery_b there's a fair amount of work going on, but it's all in mpp6 and kp6 which are self-bootstrapping perl6 compilers
22:58 * allbery_b suspects a certain part of it is just that haskell's a pretty formidable bar to participation
22:58 putter allbery_b: re audreyt, it does definitely look like total commits tracks audreyt's commits, with the largest excursion being the p5/js backends, which eventually again blocked on availability of audreyt.
23:00 putter audreyt availability == "free time" given past lack of success in finding a way to give her $3k/mo to work full time.
23:01 cmarcelo allbery_b: the fix worked here for the cases mentioned..
23:01 putter Is the above widely know do you think?  p6 progress rate ~~ audreyt time available == audreyt's "free time" because we can't find measly funding?
23:02 allbery_b well, I'd also claim it's an indication that we need an @Audrey
23:02 putter lol :)
23:03 cmarcelo note that, while haskell maybe a barrier for some participators ("bar" means that, right?), it captured people from Haskell/Functional world to participate...
23:03 allbery_b close enough.  (think pole vaulting)
23:04 allbery_b actually the haskell world credits pugs with bringing more people into haskell
23:04 putter cmarcelo, allbery_b: were you already aware of the "can't find $3k/mo, so no FT audreyt" concept?
23:05 * putter wonders whether it needs to be more widely known.  which might fix it.
23:05 trym___ joined perl6
23:05 * allbery_b was not specifically aware but is not especially surprised
23:05 allbery_b gnome started moving much faster when there were people paying developers
23:06 cmarcelo putter: I was aware of the correlation, not the specifics. Did she "self-funded" her participation last year?
23:06 putter cmarcelo?
23:06 putter yes
23:07 putter I'm not sure how much % time that represented.
23:07 putter It also seems to perhaps over-extend her a bit?  maybe.
23:08 putter what would an @audreyt look like.
23:08 putter I don't have a clear idea of how @Larry works.
23:08 cmarcelo putter: sorry, i didn't get (understand) your two last messages..
23:08 cmarcelo three
23:08 nelhage joined perl6
23:09 putter Which was going to be another question.  Regards what are bottlenecks in getting the spec written.
23:09 putter cmarcelo: you know the @Larry concept/
23:09 putter err, concept?
23:10 cmarcelo putter: I don't know much, but looks like the spec was getting lots of feedback/corrections/etc from pugs effort itself ("let's try use and see what happens")..
23:10 allbery_b basically we need more than one person who understands the code well enough to not only work on it, but also help othrs work on it
23:10 jisom joined perl6
23:10 cmarcelo putter: no.
23:11 allbery_b one person is a serious bottleneck not only when that one person doesn't have time, but if that quake a coupel weeks ago had been farther north then pugs might be in serious trouble right now
23:11 * putter googles...
23:11 cmarcelo putter: (sorry, i'll google it..)
23:11 allbery_b larry was "executive director" for perl <= 5.  for perl6 the job has been spread out, hence @Larry
23:12 putter not something google is good at :/
23:12 allbery_b it's mentioned somewhere in the perl6 web pages IIRC
23:12 cmarcelo re @Larry: I think I get the idea..
23:13 putter allbery_b: are all(@Larry) equivalent?  or are they specialized?
23:13 lisppaste3 joined perl6
23:13 allbery_b http://dev.perl.org/perl6/people.html is a start
23:13 lambdabot Title: Who's Who in Perl 6, parrot, & Pugs - perl6
23:15 putter so it sounds like "specialized, maybe"
23:16 allbery_b looks like it
23:16 allbery_b I'm not sure what part of that set is actually "@Larry", might just be Larry & Damian
23:16 cmarcelo I agree with you that more people with "project global understanding" would help. In a sense, STATUS/TASK works as a Stub for these people presence... (makes sense?)
23:16 allbery_b (larry's probably gonna come back and thwack me for being a clueless yutz :> )
23:17 putter makes sense
23:17 allbery_b complicating things is that ther's perl6, then there's pugs and parrot
23:18 putter I'm not actually sure whether *anyone* has "project global understanding" at this point, except maybe audreyt.
23:18 putter :/
23:18 allbery_b in terms of pugs, that's likely correct
23:18 allbery_b hence my @Audrey comment
23:18 cmarcelo A question: is someone helping fglock with mp6/etc?
23:19 putter good question
23:19 allbery_b no idea, really; I know there's activity but don't know who's been involved (as I said, I've mostly been concentrating on haskell skills)
23:20 putter re what is determining the pace at which the spec matures.  one factor is pugs serves as a forcing factor.  anything else?
23:21 putter If Larry had a FT p6 job, would the spec mature N times faster?
23:21 allbery_b pugs; parrot; people noticing things in the spec and asking questions, leading to discussion on the p6 list (quite a lot of activity in the smart matching spec recently due to that)
23:22 putter So speed pugs by N, and spec'ing speeds by N?
23:23 allbery_b or speed parrot by N --- seems to me from the weeklies that there's a fair amount of activity over there currently
23:23 cmarcelo putter: when we have pugs 6.28 (are you aware of this version numbering milestones for pugs?), probably the specs on Objects et al will get more discussion...
23:24 putter aware.  though I've never been too clear on how much we are using them, or how useful they are...
23:26 putter the last? release was said to be the one just prior to overhauling the AST.  did the ast overhaul actually happen?  or did audreyt run low on tuits?
23:27 cmarcelo I asked about mp6 because with an updated STATUS/TASKS, and more people knowing what's going on it would be easy to point someone what to do for helping, even not knowing details.. Today we just point the person to fglock (for example) or to the code, which isn't enough sometimes...
23:27 putter indeed
23:28 putter hmm, there is a little utility app.  something which grovels over the irc log and summaries when the best time to catch someone here is.
23:29 putter has nothingmuch or anyone been working on the p6 compiler?
23:30 * putter contemplates adding an "Availability" subsection to the Implementations section of http://en.wikipedia.org/wiki/Perl_6
23:31 cmarcelo (what's "tuits"?)
23:31 putter "get to it"s
23:31 putter "tuit" - a unit of available effort
23:31 allbery_b 'round tuits' in the circles I normally hang out in
23:32 putter "get a round to it"
23:32 allbery_b yep
23:33 Aankhen`` joined perl6
23:33 putter erps, that should have been  "get around to it" ->  "get a round toit" -> "get a round tuit"
23:33 putter sigh
23:34 cmarcelo AST overhaul is related to MO, which even has a bootstrap plan (docs/notes/mo-boo*), which I'm trying to help getting done, but the lack of global understanding is "holding" me a bit.. (we translated nothingmuch's MO to Haskell at YAPC::SA and imported the code to pugs tree)
23:34 putter there's another whole angle of people not appreciating how significant p6 is computer science wise.  but that doesn't seem to be the bottleneck just now.
23:34 perlbot joined perl6
23:35 * putter looks at mo...
23:35 cmarcelo hehe, thanks for the "tuits" explanation :)
23:37 cmarcelo right now I'm "playing" with autoboxing stuff (at src/MO/si.hs -- but watchout because the code is a little messy)
23:38 penk joined perl6
23:39 putter are gaal and iblech around much lately?
23:40 putter actually, a better question,
23:40 allbery_b gaal was around earlier
23:40 allbery_b dunno about activity
23:40 cmarcelo putter: since looks like you're collecting suggestions => lots of people seem to miss pugs.blogs.com posts (i don't know how this affects outsiders, since there are weekly reports of the mailing lists).
23:41 cmarcelo They called people's attetion for some aspects, "capturing" people interested in this aspects to contribute..
23:42 putter good suggestion.  I think there was an effort to get more people than just audreyt to post, but I'm not sure it gelled.
23:42 cmarcelo allbery_b: did your compilation finished?
23:42 allbery_b yes, I'm running full test suite now
23:43 cmarcelo Mark Stoberg(?) posted very nice things there for a while..
23:43 allbery_b although I'm wondering what missing whiteSpace you noticed earlier, suggesting I'll trip over the same issue
23:43 * putter notices p6 has two wikis at the moment.  one spammed.
23:44 bsb joined perl6
23:44 putter so how does the MO stuff fit into the project...?   MO permits X which permits Y which....?
23:44 cmarcelo allbery_b: I think was my mistake. The whiteSpace is only needed inside the C<< "-" <- name >> case.. (which the fix have)
23:46 allbery_b ahhh, first breakage found
23:46 allbery_b pointy blocks
23:47 cmarcelo putter: as I understand, MO gives a framework to implement OO/meta-OO stuff. Which means getting pugs 6.28 release. But the integration between MO code and Pugs isn't that clear for me (even with the notes/mo-boo* file) :P
23:47 jisom joined perl6
23:48 rhizo joined perl6
23:48 cmarcelo allbery_b: which test? (for example)
23:49 allbery_b t/blocks/statement_control.t parse error at line 7 is the first one I noticed
23:53 sonorous joined perl6
23:59 * allbery_b observes that looking fr an operator called "->" in haskell source is not especially rewarding...

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo