Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2006-07-16

Perl 6 | Reference Documentation | Rakudo

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

All times shown according to UTC.

Time Nick Message
00:06 nothingmuch joined perl6
00:28 fglock joined perl6
00:37 svnbot6 r11557 | fglock++ | v6 - prelude: NaN, Inf are exported
01:09 ruoso joined perl6
01:14 scook0 joined perl6
01:16 markstos joined perl6
01:22 swaroop joined perl6
01:22 swaroop joined perl6
01:54 elmex_ joined perl6
02:03 unobe joined perl6
02:06 unobe joined perl6
02:07 Termy joined perl6
02:17 Termy joined perl6
02:19 Termy joined perl6
02:57 unobe joined perl6
03:00 unobe is now known as unobe_away
03:23 rashakil joined perl6
03:39 SubStack joined perl6
03:41 Termy joined perl6
03:57 merlyn v6 on the cpan depends on Sub::Multi, which isn't on the CPAN. ?
03:59 audreyt @seen clkao
03:59 lambdabot clkao is in #perl6. I last heard clkao speak 1 day, 3 hours, 26 minutes and 22 seconds ago.
04:01 audreyt merlyn: it has an underscore
04:01 merlyn oh, well that's no good.
04:01 merlyn can't install from the CPAN if it's not normal
04:01 audreyt I'll release.
04:01 merlyn the index from Sun, 16 Jul 2006 01:29:12 GMT doesn't contain Sub::Multi
04:02 Termy me voy... buenas :)
04:02 Termy left perl6
04:02 clkao hi
04:02 audreyt clkao: hey!
04:02 clkao zz
04:02 audreyt clkao: release 0.002?
04:03 clkao sure
04:03 audreyt ok. I bumped version for you
04:04 svnbot6 r11558 | audreyt++ | * Sub::Multi - version bump to 0.002
04:04 svnbot6 r11559 | Darren_Duncan++ | perl5/Pugs-Compiler-Perl6/ : in both v6.pm and README, restored the 'invalid header' example from v6-alpha to v6, so it is actually invalid
04:07 svnbot6 r11560 | audreyt++ | * perl version dep: 5.8.x
04:08 cmarcelo joined perl6
04:09 clkao audreyt: meanwhile, get yourself comaintainership. i need some sleep
04:10 audreyt okay.
04:11 dduncan joined perl6
04:12 dduncan about that last commit, why 5.8.0 exactly ... considering that it is close and makes the largest number of fixes or improvements, how about 5.8.1 instead?
04:12 dduncan then we know we're getting Unicode 4 for one thing
04:12 audreyt uhm, for Sub::Multi doesn't matter
04:13 audreyt sorry, it's just for Sub::Multi not pug
04:13 audreyt pugs
04:13 cmarcelo left perl6
04:14 dduncan still, I question why anything would list 5.8.0 as a minimum dependency ... if you're going above 5.6.x, I would start with 5.8.1
04:14 dduncan but whatever
04:16 audreyt I understand this sentiment
04:16 audreyt as 5.8.0, like 5.6.0, is very unstable
04:18 clkao the force_todo_junction in test.pm can be undefined. wouldn't that generate a warning? or was it intentional and i missed something?
04:19 audreyt it might generate a warning
04:19 audreyt but pugs's warning system is a bit borken
04:19 audreyt it was intentional
04:19 clkao well, in v6 it does generate a warning
04:19 clkao should that default to an empty string then?
04:19 audreyt yes.
04:19 audreyt it was a junction
04:20 audreyt pmichaud discussed with me and we made junction a nonmandatory req for Test.pm
04:20 audreyt so it was lost in refactoring
04:20 audreyt so empty string is fine
04:20 clkao *nod*
04:20 clkao how's jifty/moose?
04:21 audreyt obra shows it passes most of HM tests
04:21 audreyt some type constraint errors remain
04:21 audreyt though he's not here so we havn't synced what remains
04:22 audreyt meanwhile, doing ::Action::Param design now
04:22 audreyt also Jifty's autovivification and structural constraints and Class::Accessor compat
04:22 audreyt makes interesting Moose features
04:22 clkao mmmm
04:22 audreyt so also going to exercise my Moose commit bit
04:22 clkao would you like to do that for svk as well? ;)
04:22 audreyt I've long wanted structural constraints
04:22 audreyt for perl6
04:23 dduncan quick question that's a tangent to my last p6l post ...
04:23 audreyt and Moose would be a good place to experiment with them
04:23 dduncan does := work anywhere, such as on attributes or array elements?
04:23 dduncan eg $!foo := $!bar
04:23 dduncan or @foo[0] := @bar[0]
04:24 audreyt yes and yes.
04:24 dduncan excellent
04:24 dduncan probably the main places I would use it right now
04:24 audreyt though for compat arrays
04:24 audreyt maybe it'd be forbidden
04:24 dduncan I'm just thinking of normal arrays, such as declared with @foo or []
04:24 audreyt my int @x; @x[0] := ...
04:24 audreyt is probably insane
04:24 audreyt but normal arrays, sure.
04:25 dduncan eg, for easiest perl 5 implementation, the Set::Relation type will use an array internally to hold its elements
04:25 dduncan if I derive one from another like with a UNION, I would want any elements in common to just be stored once
04:26 dduncan hence assignement to array elements
04:26 spinclad joined perl6
04:26 dduncan on a related note, if I wonder if I can get that using @foo.push(), or whether I have to subscript only
04:27 dduncan still, a minor matter for now
04:29 audreyt .push takes rvalue
04:29 audreyt so it can't do implicit alias
04:47 clkao 499 tests pass tonight
04:49 merlyn 499 tests pass in the night...
04:51 audreyt clkao: have you written ingo about uploading?
04:51 audreyt clkao: also can you post a smoke.html somewhere (say on feather)?
04:51 clkao yes, it's on hiveminder
04:52 audreyt url?
04:52 clkao and it's going to feather/~clkao/
04:52 clkao in pugs group
04:52 audreyt okie
04:53 clkao nedd to slzzp
04:55 clkao currently only data_type/subtype causes infinite loop. it's about parsing "sub only_accepts_odds(Num::Odd $odd) { $odd + 1 }"
04:55 clkao i tried to add signature_term to accept : and then perl segfauled
04:55 clkao :/
04:55 clkao so we now parse about 12% of the tests, and pass 5%
04:58 mugwump hmm, can objects have a custom way of generating .id ?
04:58 audreyt yes,
04:58 audreyt method id
04:58 audreyt I think it should really be called object_id somehow
04:58 mugwump say two objects are the same .id, then they =:=, right?
04:58 audreyt ruby people learned that lesson painfully
04:58 audreyt no, they would be ===
04:58 mugwump heh
04:58 audreyt =:= is something else altogether
04:59 audreyt $x =:= $y means variable($x) === variable($y)
04:59 merlyn the flying colon!
04:59 merlyn colon biplane
05:00 mugwump hmm, I want the effect of setting the object ID to be that if somebody creates a new object and after BUILD it determines that the object ID is the same as an existing object, to return the existing one
05:00 mugwump BUILDALL even
05:01 mugwump ie, a custom rule that $x === $y  implies $x =:= $y
05:01 audreyt ok.
05:02 mugwump the application is, say, an object encapsulating a stream of bytes.  we make the object id the sha1 sum of the contents, and if somebody creates a new object with the same contents, it's the same object
05:03 audreyt I understand.
05:04 audreyt simply have "method id" overridden will do that.
05:05 audreyt hm. help wanted
05:05 audreyt http://search.cpan.org/dist/v6-alpha/HACKING looks bad
05:05 lambdabot Title: The CPAN Search Site - search.cpan.org
05:05 audreyt because http://search.cpan.org/src/FGLOCK/v6-alpha-0.009/HACKING is not POD
05:05 audreyt is someone up to reformat HACKING into proper POD?
05:06 audreyt README too, if possible
05:06 audreyt oh, nvm README
05:06 audreyt just perl5/Pugs-Compiler-Perl6/HACKING
05:06 mugwump you have to be careful in the Makefile.PL not to deliver extra .pod files like that :)
05:06 mugwump I do that with some of my modules, I have a Changes.pod
05:15 audreyt *wave* &
05:19 Dr_Pi joined perl6
05:20 Dr_Pi /msg NickServ Help
05:27 Dr_Pi /msg NickServ IDENTIFY ivanfrey iF6412
05:31 Dr_Pi joined perl6
05:33 Dr_Pi /msg compiler-per6 HELP!!!!
05:33 Dr_Pi left perl6
05:35 ivanfrey joined perl6
05:38 dduncan mugwump, regarding: mugwump: the application is, say, an object encapsulating a stream of bytes.  we make the object id the sha1 sum of the contents, and if somebody creates a new object with the same contents, it's the same object ...
05:38 dduncan is it possible for several streams to result in the same sha1 sum?
05:39 dduncan if so, then the returned object would be different than what was input
05:39 mugwump generally the assumption is that never happens
05:40 mugwump esp. in this case the sha1 sum hashes the length, too
05:41 swaroop audreyt, I've created a pod-ified version of HACKING, where do I submit the patch ?
05:41 dduncan in any event, what you described at BUILD time could more or less satisfy what I was wanting in my p6l post, or perhaps satisfy it better than that ... because duplicate objects are never created in the first place, never mind possibly eliminated if compared
05:41 mugwump swaroop: in the pugs tree
05:42 mugwump 17:06 < audreyt> just perl5/Pugs-Compiler-Perl6/HACKING
05:42 swaroop mugwump, yes, but should I ask someone to commit the patch for me or can I do it myself (it's not been reviewed...)
05:42 dduncan I thought of suggesting something like this myself, but I didn't know if it was normal for there to be a central index of all objects of a class
05:42 dduncan in the meta-something-er-other
05:42 mugwump swaroop: just commit, do you need a committer invitation?
05:42 swaroop mugwump, yes
05:45 kanru joined perl6
05:45 swaroop mugwump, ?
05:46 mugwump sorry, I was just checking http://rt.openfoundry.org/Foundry/Project/?Queue=270.  You'll need a project admin to send you the invitation
05:46 lambdabot Title: Pugs -- OSSF
05:47 mugwump there are a lot of admins :)
05:47 swaroop mugwump, ok :), but whom do I ask ?
05:48 mugwump Well normally asking here is the right place.  However this is a bad time of day for interactive questions.  Unless someone with an admin bit steps forward, then perhaps send a request to the perl6-compiler list.  If you're in a hurry drop it in a nopaste somewhere and I'll commit it for you.
05:49 swaroop mugwump, ok sure
05:49 swaroop mugwump, thx :)
05:49 dduncan curiously, that admin list seems to include me
05:50 mugwump well perhaps you'd like to take swaroop aside and run through the initiation rites with him then, dduncan ?
05:50 dduncan but given mugwump's involvement, I'm surprised it doesn't involve him
05:50 * mugwump mehs
05:50 dduncan I will have to learn how to do that though
05:51 dduncan never actually used admin privs before
05:52 dduncan I'll figure out how to login first, and then see if sending an invite is obvious
05:52 svnbot6 r11561 | snowstalker++ | Pod-ified according to Audrey's request (
05:52 svnbot6 r11561 | snowstalker++ | http://colabti.de/irclogger/irclogger_log/perl6?date=2006-07-16,Sun&amp;sel=72#l135 )
05:52 lambdabot Title: #perl6 2006-07-16,Sun
05:52 mugwump login password should be same as svn
05:52 swaroop I recollected that I had a previous login as a diff id, so committed :)
05:53 swaroop If you guys could send me another committer bit, i'd like to use the same user id across parrot and pugs :)
05:53 mugwump ok, dduncan : login, there is a "Project Admin" menu on the left.  Go to "Members", then there is a form
05:55 Southen_ joined perl6
05:56 unobe_away is now known as unobe
05:59 dduncan okay, I've finally logged in ...
06:02 dduncan okay, swaroop, what is your email address?
06:05 dduncan also, mugwump, check your email
06:05 dduncan I've tried to upgrade you to an admin
06:05 dduncan see if you get an email saying that
06:06 dduncan swaroop, I will need your email address in order to send you a committer invitation
06:08 Eidolos joined perl6
06:11 dduncan mugwump, if you find out swaroop's email address before I do, you can now add him yourself, afaik
06:11 dduncan oh wait ...
06:12 mugwump swaroop: what is your parrot account that you want activated?
06:12 mugwump thanks dduncan
06:12 dduncan so snowstalker is swaroop's committer id?
06:12 dduncan I didn't backlog for some reason, so I guess he's in already
06:13 dduncan I'm in the same situation though, different committer id from chat id
06:13 dduncan that's legacy for you
06:15 mugwump yeah.  I guess there's no way to rename an account.
06:15 unobe joined perl6
06:15 dduncan I wish there was ... would have done it a year ago then
06:18 dduncan oh, fyi ... I don't know if I've told you or just others, but in light of the availability of v6.pm, I do not intend to further develop my newest CPAN modules in Perl 5 code, but only the Perl 6 versions, because they should integrate into Perl 5 reasonably well any time now
06:19 mugwump great to hear.
06:19 dduncan and other perl 5 code should be able to use it / be used by it
06:19 dduncan its what I really want to do
06:19 dduncan and it gives more impetus to test Moose et al in action
06:19 dduncan indirectly
06:20 dduncan fyi, the moment v6.pm is able to fully run any of my perl 6 modules, they will go on CPAN independently like a normal perl 5 distro
06:21 dduncan and to assist that I will be keeping my language footprint requirements reasonably small
06:21 dduncan anyway, back to work ...
06:21 swaroop dduncan, sorry just came back, my address is [email@hidden.address]
06:22 dduncan thank you, but did you just say before that you already had a different committer id?
06:22 dduncan eg, are you snowstalker?
06:22 swaroop dduncan, yes, but I want to change the user id to just 'swaroop'
06:22 swaroop dduncan, yes
06:22 nothingmuch joined perl6
06:23 dduncan you know, it occurs to me that I could likewise change my id from Darren_Duncan to dduncan by just making another account ... if they accept the same email address for 2 of them
06:23 dduncan I resisted that thought a year ago, but maybe now it'll be different
06:24 dduncan meanwhile, sure, will try it for you
06:24 swaroop dduncan, thx :)
06:25 dduncan check your email
06:26 swaroop dduncan, got it
06:26 dduncan now, I don't know if it'll work for me exactly, since I want to use the same email for both ... on the other hand, I've already discovered that they handle the emails case sensitively, while the domain system doesn't, so I'll try a different case if I need to ...
06:26 unobe joined perl6
06:29 dduncan okay, so it is case-insensitive in some cases
06:30 dduncan I'll leave it alone for now
06:49 dduncan left perl6
06:50 wesle1 joined perl6
06:58 Aankhen`` joined perl6
07:03 nothingmuch morning
07:10 b00t joined perl6
07:16 shachaf joined perl6
07:56 iblechbot joined perl6
08:06 larsen joined perl6
08:14 kanru joined perl6
08:15 pmurias joined perl6
08:15 pmurias nothingmuch: morning
08:17 nothingmuch pmurias: morning
08:17 nothingmuch busy-ish
08:17 nothingmuch but i will still respond
08:23 shachaf joined perl6
08:36 pmurias clkao: if i wanted to write an new emitter for pX, what would be the (best) way to do it?
08:37 pmurias (i'm thinking of writing an AST emitter)
08:38 kanru joined perl6
08:39 xinming_ joined perl6
08:55 xerox_ is now known as xerox
08:58 nothingmuch pmurias: what AST?
09:05 revision17_ joined perl6
09:12 elmex joined perl6
09:15 miyagawa joined perl6
09:20 nothingmuch w 16
09:24 szbalint joined perl6
09:27 miyagawa joined perl6
09:49 edwardk joined perl6
10:03 gaal hello there
10:04 edwardk heya
10:04 gaal here's some advance notice :)
10:04 gaal http://perlmonks.org/?node_id=561539
10:04 lambdabot Title: OSDC::Israel - Call for Participation
10:04 gaal February, and everyone's invited :)
10:04 edwardk its a bit of a swim =)
10:05 * nothingmuch would *love* to see some haskell at OSDC
10:05 Southen_ is now known as Southen
10:05 gaal nothingmuch: word.
10:05 nothingmuch yo homie rezpect
10:06 * gaal goes off for lunch.
10:31 xinming joined perl6
10:45 pmurias nothingmuch: the new shiny p6notes.hs and notes/p6ast_draft AST
10:46 xinming joined perl6
10:46 nothingmuch pmurias++
10:47 nothingmuch but arguably that is more of an execution oriented AST
10:47 nothingmuch i would also like to see v6 refactored in such a way that an AST that has high source fidelity but is very much not a parse tree is the first stage
10:50 pmurias nothingmuch: what are the pros and cons of Jifty and Catalyst?
10:50 nothingmuch Jifty is a full stack
10:50 nothingmuch it requires you to bend to it a bit
10:50 nothingmuch but gives you much more as a result
10:50 nothingmuch catalyst is more of a toolkit type mix and match approach
10:50 nothingmuch use what works for you
10:50 nothingmuch but has lower expressiveness as a result
10:51 nothingmuch you can think of ruby on rails as a tricycle (can't remove training wheels)
10:51 nothingmuch and of jifty as an uber cool mountain bike
10:51 nothingmuch and of catalyst as a DIY kit
10:53 * pmurias dosn't like mountain bikes were much
10:53 chris2 joined perl6
10:55 wolverian I definitely like catalyst. haven't touched jifty yet, scared of the docs (or lack of) and mason :)
10:55 wolverian but, yes, catalyst is very diy.
10:56 pmurias s/were/very/
10:57 ruoso joined perl6
11:00 nothingmuch wolverian: mason is not a "strong" dep anymore
11:00 nothingmuch but this needs a little more work
11:00 nothingmuch that said
11:00 nothingmuch i've seen a lot of jifty during YAPC and obrathon
11:00 nothingmuch and I was *really* impressed
11:00 nothingmuch but you need lots of freedom to be able to decide to use Jifty
11:01 pmurias why?
11:07 pmurias eq. convincing your boss or does it forces you to do some things a specific way?
11:14 pmurias i suppose the latter
11:17 ludan joined perl6
11:24 svnbot6 r11562 | nothingmuch++ | update stub template
11:24 Aankh|Clone joined perl6
11:27 svnbot6 r11563 | nothingmuch++ | update other stubs as well
11:28 mauke_ joined perl6
11:43 mauke_ is now known as mauke
11:48 ludan is now known as panda
11:49 panda is now known as ludan
11:51 cognominal joined perl6
12:06 scw S05.pod line 637:
12:06 scw       / <after pattern> /     # was /(?<pattern)/
12:06 scw should be /(?<=pattern)/
12:07 scw (asked perlre to ensure that I didn't forget the syntax :p)
12:09 gaal nothingmuch: moose
12:09 gaal are you not jetlagged?
12:17 nothingmuch joined perl6
12:36 Aankh|Clone is now known as Aankhen``
12:57 buetow joined perl6
13:00 kanru joined perl6
13:04 weinig joined perl6
13:09 miyagawa joined perl6
13:15 penk joined perl6
13:34 Aankh|Clone joined perl6
13:41 SubStack joined perl6
13:41 SubStack_ joined perl6
13:42 SubStack joined perl6
13:47 buetow joined perl6
14:00 miyagawa joined perl6
14:03 pmurias scw: you are right
14:04 pmurias (?<pattern) is invalid
14:18 jkanter joined perl6
14:22 stclare left perl6
14:26 Aankhen`` joined perl6
14:45 Bit-Man_ joined perl6
14:45 Bit-Man_ is now known as Bit-Man
15:00 elmex joined perl6
15:12 anm joined perl6
15:17 frederico joined perl6
15:30 lollan joined perl6
16:00 lisheeghan joined perl6
16:10 iblechbot joined perl6
16:10 DHGE joined perl6
16:16 lisheeghan left perl6
16:19 froh-doh joined perl6
16:28 wesle1 joined perl6
16:33 Termy joined perl6
16:35 svnbot6 r11564 | audreyt++ | * bump Parrot requirement to 0.4.6 (not yet released),
16:35 svnbot6 r11564 | audreyt++ |   for vsoni++'s recent work on making imcc reentrant.
16:35 svnbot6 r11564 | audreyt++ |   This is needed because parrot currently lacks a sane way to
16:35 svnbot6 r11564 | audreyt++ |   distinguish 0.4.5 from 0.4.5-devel in C-preprocessor land.
16:38 svnbot6 r11565 | iblech++ | * Improved the error messages of the smokeserver.
16:38 svnbot6 r11565 | iblech++ | * Diagnosed the cause for the rejection of the v6.pm-smoke (http://xrl.us/o5jt):
16:38 svnbot6 r11565 | iblech++ |   "Not all of the required fields (pugs_version osname) exist."
16:38 lambdabot Title: Metamark
16:40 clkao iblech++
16:48 Aankhen`` joined perl6
17:05 svnbot6 r11566 | audreyt++ | * now that PARROT_REVISION support svk checkout (r13320 by
17:05 svnbot6 r11566 | audreyt++ |   yours truly), we can go back to support Parrot 0.4.5
17:05 svnbot6 r11566 | audreyt++ |   more reasonably with #define magick.
17:05 svnbot6 r11567 | audreyt++ | * Excise PIL-Run from Pugs -C/-B's list of helper directories
17:05 svnbot6 r11567 | audreyt++ |   now we've switched fully to v6-alpha.
17:09 * xerox quotes audrey's blog "iblech++'s multi-month PIL2JS journey on getting a few more tests passing each day"
17:09 xerox Man, this iblech!
17:29 chris2 joined perl6
17:34 nothingmuch ingy: ping!
17:36 clkao moose
17:36 xerox moo
17:45 DHGE left perl6
17:47 svnbot6 r11568 | nothingmuch++ | Mailing list summary for 2006-03 by kudra++
17:52 svnbot6 r11569 | clkao++ | Simple emit for %hash{}, passes statements/array_values_in_bool_context.t
17:52 svnbot6 r11570 | clkao++ | force_todo_test_junction is actually a string, not junction.
17:55 shachaf joined perl6
18:00 miyagawa joined perl6
18:01 lisheeghan joined perl6
18:14 clkao eval in perl5 sets $@ to empty string. but test.pm seems to be expecting $! to be undefined when eval works
18:14 merlyn in the old days, $! was never unset, only set
18:14 vsoni hi audreyt
18:15 merlyn I don't know when someone broke that, but it's broken now
18:15 merlyn oh wait.  it apparently still works right
18:16 merlyn eval shouldn't affect $! at all though, unless things inside the eval do syscall-related things
18:25 clkao i was saying perl6 test.pm ;)
18:32 Jedai joined perl6
18:40 svnbot6 r11571 | clkao++ | kill warnings about redefining my $i.
18:40 svnbot6 r11572 | clkao++ | When testing defined-ness of $!, it is testing the emptiness of $@ in perl5.
18:43 svnbot6 r11573 | nothingmuch++ | fix pod links
18:52 svnbot6 r11574 | nothingmuch++ | hack dumb links so that there aren't escaping issues
19:18 penk joined perl6
19:22 clkao ?eval my $g; for (0 .. 5) sub { $g = $g ~ $_; }; $g
19:22 evalbot_11550 is now known as evalbot_11574
19:22 evalbot_11574 \"012345"
19:22 clkao ?eval if(1) sub { 1 }
19:22 evalbot_11574 Error:  unexpected "s" expecting comment, operator or block
19:23 clkao so for can take a sub?
19:36 merlyn how come Module::Compile doesn't do the right thing when there's no .pmc file?
19:36 merlyn is that impossible?
19:36 merlyn shouldn't it just act like a source filter then, compiling each time?
19:36 merlyn or am I missing sometihng?
19:37 merlyn if the .pm file and the .pmc file act *differently*, that's clearly not the intent of what .pmc was meant to do
19:37 diotalevi So you found a bug.
19:37 merlyn I don't konw.
19:37 merlyn it's not a bug if it's working as designed.  I'm just trying to figure out if this is the design
19:38 merlyn in which case, I'd argue with the design. :)
19:38 diotalevi You're correct that they're  supposed to act the same.
19:39 diotalevi When it doesn't, it's a bug.
19:39 obra &
19:40 diotalevi Unless the intent changed in the last month.
19:40 merlyn well, I'm not convinced Module::Compile intends to fulfill my assertion about .pm and .pmc
19:40 merlyn not from a reading of the code
19:40 merlyn it looks instead as if ".pm is the source for .pmc" is more of the intention
19:41 merlyn and that the .pm is only accidentally in the @INC for convenience
19:41 merlyn but was never meant to be loaded
19:41 diotalevi No....
19:41 diotalevi Maybe for the M'C using thing you're looking at but not M'C in general.
19:41 merlyn OK.  I create Spew.pm, which uses Module::Compile -base.  I write Mission.pm, which says "use Spew".
19:42 merlyn if I say "perl -MMission", it should act the same whether I've said "perl -c Mission.pm" or not
19:42 diotalevi yeah.
19:42 merlyn but apparently, perl -c Mission.pm compiles a Mission.pmc, that has an entire different behavior than the original .pm
19:42 merlyn that's what I'm saying is broken
19:42 diotalevi d'oh.
19:43 merlyn I should be able to "use Mission" that works the same whether the .pmc is there or not
19:43 merlyn although a lot slower if the pmc doesn't exist
19:43 merlyn since it'll have to go through the source filtering
19:43 pdcawley joined perl6
19:43 diotalevi yeah. typing 1 handed. vry slow. :-/ chips in other.
19:44 merlyn so am I misunderstanding Module::Compile, or .pmc, or my own observation (and reading of code)?
19:45 merlyn in other words, I view .pm => .pmc as "unoptimized => optimized", not "source => object"
19:45 diotalevi Well I suppose you don't write a .pmc when you load a .pmc.
19:46 merlyn apparently, Module::Compile currently views .pm => .pmc as "source => object"
19:46 pmurias merlyn: if there isn't a .pmc doesn't it create one automaticly, so the behaviour is the same
19:46 diotalevi that's a designed in difference.
19:46 merlyn "create one automatically"?
19:46 merlyn this is not my observation
19:46 merlyn and what if it can't write it?
19:47 merlyn what if I just install .pm in my tree as root, and you come along as a user
19:47 merlyn it should still work
19:47 merlyn albeit more slowly. :)
19:47 pmurias it should
19:48 merlyn as in, it currently does, except for my demonstrated bug, or as in it ideally should?
19:48 diotalevi If you load a .pm then a .pmc should be written if possible and the filtered version should be loaded regardless. If you load a .pmc then you just get the filtered version.
19:48 merlyn diotalevi - OK, then I'm seeing a bug
19:48 diotalevi btw merlyn, I've noticed that you can't use left shift anywhere in your module.
19:49 diotalevi there's a currently failing test against that.
19:49 merlyn yeah, that silly precdence of operators.
19:49 merlyn or rather, operators with two meanings
19:49 diotalevi yeah. the one module i wanted to use M::C for had a few left shifts in it. They're not really optional either.
19:49 diotalevi Bummer for me.
19:50 diotalevi Well... I guess not 100% required. I mean I could go emulate left shift. That'd suck though.
19:50 merlyn ick
19:51 diotalevi /power-of-two.
19:51 merlyn heh
19:51 diotalevi er, *power-of-two.
19:51 diotalevi wrong direction.
19:51 merlyn eval 'sub left_shift { shift <' . '< shift }'
19:52 merlyn left_shift(10, 4)
19:52 diotalevi the hell?
19:52 merlyn what?
19:52 diotalevi your shifts even have undefined order of operation.
19:52 diotalevi in perl5 anyway.
19:52 merlyn it still works. :)
19:52 diotalevi yeah, yeah, whatever.
19:52 merlyn heh
19:53 merlyn anyway, so I'll report this bug in Module::Compile, and hope that it doesn't stop me from writing an article
19:58 merlyn Ahh - there's the problem.  I need to return text from pmc_compile.  Not just return. :)
20:00 merlyn return has to be not just an integer
20:00 diotalevi I gather you can name &pmc_compile as &FILTER...
20:00 merlyn otherwise the original text is used.  OK
20:00 merlyn not a completely intuitive interface, but I now see why I thought it wasn't working
20:01 merlyn now I just have to wire the rest of Spew into it
20:02 diotalevi also, I email Ingy about the left shift thing. Right now parsing is a simple regexp. I suggested PPI and apparently there was some discussion between him and adamk about it and decided not to use PPI.
20:06 stevan merlyn: FYI - I added "no Moose" support into the most recent version
20:06 stevan still can't do Moose::has yet, but that shouldnt be too hard
20:09 diotalevi stevan: can I tell do dependant destruction in moose yet?
20:09 diotalevi wait, bad time to ask the question.
20:09 diotalevi gotta go.
20:13 kudra joined perl6
20:14 kudra hello
20:17 unobe is now known as unobe_away
20:20 diotamobile joined perl6
20:21 stevan diotalevi: I am not sure what you mean by dependent destruction
20:22 Khisanth joined perl6
20:22 diotamobile joined perl6
20:25 larsen joined perl6
20:26 vel joined perl6
20:27 merlyn Yup.  definitely getting different behavior if the pmc isn't there
20:27 merlyn Now i just have to figure out how to describe the problem to submit a bug report
20:28 lisppaste3 joined perl6
20:28 merlyn I guess this is not ready for prime time.  I'll have to find a different topic now
20:29 merlyn darn it.  wanted to promote perl6 stuff more. :(
20:31 * kudra thinks it is bedtime.
20:31 kudra Goodbye
20:39 * merlyn files an rt bug
20:39 merlyn ... http://rt.cpan.org/Ticket/Display.html?id=20494
20:39 lambdabot Title: #20494: newly-created .pmc file is not loaded/executed when built
20:48 weinig is now known as weinig|bbl
21:12 penk joined perl6
22:11 gjvc joined perl6
22:25 clkao seen nothingmuch
22:25 jabbot clkao: nothingmuch was seen 4 hours 51 minutes 15 seconds ago
22:30 furrybeagle joined perl6
22:35 * nothingmuch shows
22:35 salamndr joined perl6
22:36 nothingmuch clkao: ?
22:39 clkao nothingmuch: where's your music
22:44 nothingmuch clkao: on my laptop
22:44 nothingmuch i can push it
22:44 nothingmuch or put it on my server
22:44 nothingmuch which do you prefer?
22:48 clkao from your server is good
22:52 nothingmuch what do you want? should i make a list of all the files?
22:53 nothingmuch ssh key please
22:55 fglock joined perl6
22:56 nothingmuch *yawn
22:56 nothingmuch *
22:56 nothingmuch i want to sleep soon
22:56 nothingmuch clkao: hurry up
23:03 bsb joined perl6
23:10 clkao fglock: around?
23:11 fglock clkao: hi
23:12 clkao fglock: hey, i was just trying to add 'loop' support.
23:13 clkao (decided to see why statement/* isn't passing
23:13 clkao fglock: turned out "loop { zz() }  while 1;" is parsed, expect a 'w' is eaten, if i remove the "<expression> while" rule, as it should require loop { } while.
23:13 fglock ok
23:14 clkao other things are 'elsif' (unless/if), given (given/do)
23:15 fglock 'loop <block>', right?
23:15 clkao no, "loop <block> while <expression>"
23:16 clkao fglock: http://scsys.co.uk:8002/2798
23:16 lambdabot Title: freenode_web paste from "clkao" at 82.35.85.110...
23:16 clkao my current diff, can you tell what's obviously wrong?
23:19 fglock clkao: it looks ok - is it committed?
23:19 clkao no
23:19 clkao it doesn't work
23:20 clkao i can't make the 1st part's perl6_expression optional
23:20 clkao adding ? doesn't work
23:20 clkao maybe i should just commit it and let you fix it ;)
23:21 fglock clkao: I think that's a bug in PCR '?'
23:22 clkao :((
23:22 fglock I think it only works after '()'
23:22 clkao fglock: and this causes anonsub to fail
23:23 clkao for unknown reason
23:23 fglock it's easier to define an optional-expression, and then use that
23:23 clkao bare_block that is.
23:23 fglock I'll apply the patch here and try to fix it
23:23 clkao ok.
23:29 leon joined perl6
23:31 leon Anyone here?
23:31 nothingmuch slightly
23:31 leon :-)
23:31 nothingmuch good night folks =)
23:32 elmex joined perl6
23:34 leon it's kinda silent for a channel with 130 people in it...
23:35 leon can anyone here answer a few questions about perl 6 development?
23:43 fglock clkao: it was missing a <ws> before '(while)'
23:43 clkao ah
23:44 clkao and the optional thingy in exp[123]
23:44 pingu joined perl6
23:44 clkao should we do maybe_empty_expression_no_block ?
23:44 clkao sounds very ugly ;)
23:47 cognominal joined perl6
23:48 fglock testing...
23:50 fglock clkao: ok!
23:50 svnbot6 r11575 | fglock++ | v6 - parser: 'loop(;;){block}' and 'look{block}while exp;' - applied clkao++ patch with a few changes
23:50 leon who should I report to for vandalized (spam) pugs wiki page?
23:51 fglock s/look/loop/
23:52 fglock leon: just say it here - people usually read the logs
23:53 leon is this channel always this silent?
23:54 leon i'm a bit surprised to see a channel with 133 users but only one message per minute
23:54 leon just wondering...
23:55 fglock it depends on which timezone most developers are - and most people just go sleep and don't logout :)
23:56 mauke_ joined perl6
23:56 leon that would explain a lot..
23:58 leon what is a better time to be here? a few hours earlier?
23:58 clkao fglock: are you going to commit that?
23:59 fglock clkao: it's r11575

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

Perl 6 | Reference Documentation | Rakudo