pugscode.org/ | nopaste: sial.org/pbot/perl6 | ?eval [~] <m oo se> | We do Haskell, too | > reverse . show $ foldl1 (*) [1..4]
Set by diakopter on 9 July 2007.
diakopter @messages 03:00
lambdabot You don't have any new messages.
diakopter yes!
Aankhen`` glances sideways at diakopter. 03:02
agentzh heh 03:03
lambdabot agentzh: You have 1 new message. '/msg lambdabot @messages' to read it.
agentzh @messages
lambdabot fglock said 15h 15m 55s ago: how about call the accessor/key "match", and include a "filename" or "str" pointer
agentzh i've got one :)
Aankhen`` @moosages 03:05
lambdabot You don't have any new messages.
Aankhen`` Mwahaahahaha.
agentzh @tell fglock will do. but i prefer names leading with '_' :)
lambdabot Consider it noted.
Aankhen`` I prefer names with a leading *!
diakopter @tell Aankhen`` message this
lambdabot Consider it noted.
Aankhen`` $#*@)#*$@#*%$%&(^%&(#($@#
lambdabot Aankhen``: You have 1 new message. '/msg lambdabot @messages' to read it.
Aankhen`` There goes the neighbourhood. 03:06
Aankhen`` What the…? 03:09
diakopter op #perl6 03:14
lambdabot diakopter: You have 1 new message. '/msg lambdabot @messages' to read it.
diakopter @mooooooooooooosages 03:14
lambdabot Unknown command, try @list
diakopter @mooooosages
lambdabot Unknown command, try @list
Aankhen`` dalek = nutty.
diakopter @moooosages 03:15
lambdabot Unknown command, try @list
diakopter @mooosages
lambdabot Unknown command, try @list
Aankhen`` You trying to get yourself banned again? :-P
Aankhen`` Uh. That's a bit of a wide ban, isn't it? 03:15
Ban (*!*@feather.perl6.nl) affects: dalek devbot6 evalbot_r16941 Juerd pugs_svnbot specbot6 wolverian
diakopter WHOA
diakopter yikes 03:15
Aankhen`` You diakopter, you. 03:16
diakopter sry :/
Aankhen`` Don't make me come over there.
diakopter is bored waiting for mythbuntu to finish installing correctly for once
Aankhen`` is reading "WhyAttributeGrammarsMatter".
(Which involves trying to figure out what the heck an Attribute Grammar is.) 03:17
diakopter har. I'm sure someone has written "Attribute Grammar[ian]s Considered Harmful" 03:18
Aankhen`` Heh, no doubt. 03:18
I'd make a crack about "Computers Considered Harmful", but it wouldn't be very funny. :-\
mncharity @tell fglock sorry I missed you. I should be around early morning and late evening tomorrow (EST). 03:28
lambdabot Consider it noted.
diakopter eeenteresting 03:41
moritz moin 06:12
agentzh morning! 06:13
unobe morning :-)
moritz can I tell `cpan' somehow to install into a different directory? 06:14
agentzh no idea 06:15
Aankhen`` moritz: You can probably pass PREFIX=/foo or INSTALL_BASE=/foo in the Makefile.PL args.
moritz I just downoaded the tarball manually now ;) 06:16
Aankhen`` Heh.
Alias_ moritz: o conf init, and set PREFIX=/path when it asks for Makefilel.PL params and Build.PL params
pugs_svnbot r16957 | moritz++ | trunk: 06:18
: [irclog]
: * added POD to IrcLog.pm and IrcLog/WWW.pm
: * index.pl now displays all possible dates as a calendar
diff: dev.pugscode.org/changeset/16957
lambdabot Title: Changeset 16957 - Pugs - Trac
moritz Alias_: thanks
agentzh moritz++ 06:19
moritz Aankhen``: the index page might welcome some HTML/CSS tweaking - it's just an evil hack atm ;) 06:20
and I guess we need a big 'Today'-Link
pasteling "evalbot_r16956" at 193.200.132.135 pasted "Pugs build failure" (6 lines, 278B) at sial.org/pbot/26165
agentzh will need Aankhen``'s help with CSS/HTML soon as well ;)
Aankhen`` Sure thing, but give me a few moments. I need to sort out this thing with Vista switching keyboard layouts on me. 06:21
agentzh heh 06:22
agentzh won't install vista anyway.
Aankhen`` Hmm.
It seems to have something to do with the language bar.
Oh, I should probably install Komodo so I can tweak stuff. 06:23
devbot6 dev: Ticket #14 (enhancement closed): More information in pugs -V output. <dev.pugscode.org/ticket/14#comment:4> 06:26
pugs_svnbot r16958 | moritz++ | trunk:
: [irclog] added 'Today'-Link to index page
diff: dev.pugscode.org/changeset/16958
lambdabot Title: Changeset 16958 - Pugs - Trac
moritz index page is slow now :( - propably due to too much Date::Simple magic 06:34
I wonder if I should fix it, or cache it ;)
pasteling "evalbot_r16958" at 193.200.132.135 pasted "Pugs build failure" (6 lines, 278B) at sial.org/pbot/26168 07:02
pugs_svnbot r16959 | moritz++ | trunk:
: [irclog]
: * much more efficient index calendar creation (less date magic)
: * whitespace/tab fixes in cgi/index.pl
: * fixed vim modelin in cgi/index.pl
diff: dev.pugscode.org/changeset/16959
lambdabot Title: Changeset 16959 - Pugs - Trac
meppl good morning 08:43
moritz hi meppl ;)
masak mornin', meppl
meppl good morning moritz
lumi Morning, M people 08:44
masak ;)
meppl good morning masak and lunik
moritz any comments regarding the new irclog index page? 08:45
it certainly looks neat, but it wastes space as well
masak moritz: url? 08:50
moritz masak: moritz.faui2k3.org/irclog/ 08:51
lambdabot Title: IRC logs
moritz and as a comparison the old one: 209.85.135.104/search?q=cache:5uibC...ceweasel-a
lambdabot tinyurl.com/yp8qyt
masak definitely an improvement 08:52
I think it wastes space in a good way
if you're too worried about it, I'd suggest adding some CSS to remove the table cell borders 08:53
moritz I'll let Aankhen`` do that, I don't want to take that pleasure from him ;)
masak ok, me neither ;) 08:54
moritz unfortunately HTML::Calendar::Simple is not so CSS friendly
masak nothing is set in stone 08:55
moritz aye ;)
Alias_ Although building in CSS support from the start helps a LOT 08:56
<-- PPI::HTML
moritz Alias_: do you know by chance if PPI or Text::VimColor generate "better" markup for perl? 08:57
Alias_ PPI::HTML is by far the most accurate markup in existance 08:58
wolverian vim isn't particularly accurate.
Alias_ You might need to play and tweak the CSS a bit, and maybe subclass PPI::HTML for anything esoteric
But it's absolutely better than anything else
Just a bit of a CPU hog for big files (because of PPI) 08:59
moritz Alias_: ok I'll try it out
Alias_ I was using an Apache plugin for a while, but then google started trawling my repository and it killed the server somewhat
moritz vim has the advantag of knowing >300 languages/config files
Alias_ ya
moritz Alias_: use robots.txt ;) 09:00
Alias_ I do now
After google deleted a production database
Because I accidentally left a database management web app exposed, and it went down the tool.pl/DatabaseSetup?cmd=purge;db=whatever link
moritz :( 09:01
Alias_ ya, not fun
Big lesson for the future though
I'm going to rewrite my connection pool at some point so that it only hands out read-only connections to GET requests 09:02
moritz "never leeave a management toll without password protection"
Alias_ Oh it had password support
The config file got reset due to version control, and if the config doesn't list a password, it doesn't use it 09:03
Something multi-factor like that from memory
moritz ouch
trippeh Heh, any data modifying action should be a POST though, not GET... 09:13
This is one of the reasons
Alias_ yep
wolverian or PUT, but ...
(or DELETE, for that matter)
Alias_ Although this makes data-stored sessions interesting
wolverian wishes browsers weren't so crippled
Alias_ Because strictly speaking you couldn't nudge last_request dates on the session in a GET
(which could be bad, but I'm not sure) 09:14
Maybe I could rejigger it so that it had permissions to update the session table, but that's all
Instead of being a pure readonly connection 09:15
I suspect my connection-pooling needs are getting complex enough to write a standalone module :) 09:16
3 classes of connections (shared read-only, shared auto-commit, and transaction pool)
moritz Alias_: are you writing a http server in perl? 09:16
Alias_ Just regular webapps 09:17
But I have a fairly interesting ORM layer
Better than anyone else, in my opinion
Alias_ But unfortunately based on code-generation, so not CPAN'able 09:17
moritz self made?
Alias_ yeah
I talk to mst quite a bit, and he's gradually stealing some of the better ideas from it 09:18
moritz ORM-Compiler, YaY
Alias_ for DBIx::Class
moritz (idea stealing)++
masak Alias_: what's particularly good about it?
Alias_ You design the structure in entity-relational terms 09:19
And the codegen produces both a sane schema and sane code to talk to it
masak nice
Alias_ And it can handle 100% of all structures
Handles importing and exporting circular data references
masak so far it sounds like Hibernate to me
Alias_ It can do queries with infite arbitrary foreign key traversal 09:20
moritz "combine the crazyness of perl with the power of databases"
Alias_ $query = Class->select;
moritz or was it the other way round? ;)
masak Alias_: cool 09:21
Alias_ $query->add_condition( 'self.parents.friends.children.school.teachers.pets.name', '=', 'Spot' );
Alias_ It generates it's own schema 09:21
So the code can create it's own database structures 09:22
And it can handle superfluous tables and columns
wolverian "its" :)
moritz Alias_: how do you distinguish references from string literals?
Alias_ It knows what entity names are reserved words in various databases and renames around the problem
moritz: hmm?
It has a model of the entity-relationships 09:23
moritz Alias_: I guess 'self.parents. ...' is treated as a database column...
Alias_ not necesarily
It's not 1 to 1
moritz Alias_: and 'Spot' is a string constant, right?
Alias_ .friends may involve traversing a many to many rable
table
moritz ok, but it's not treated as a string constant
Alias_ or going backwards across a relationship, because it's the OTHER side of the relationship that has the column
moritz whereas 'Spot' is treated as a constant, right?
Alias_ the entity model knows the difference between data and relationships 09:24
The "=" condition class knows the data types it's allowed to build a condition for
So if you tried that on something that wasn't a string or a number, it would barf
moritz so if you have some external input, $dog_name, how can you be sure that it doesn't contain a string that describes a column and messes up your whole system?
Alias_ huh? 09:25
Because of sql placeholders...
The SQL generation handles placeholders
moritz so you add sql placeholders in add_condition and the like?
Alias_ As it builds the SQL it keeps track of them and aggregates them in the correct order as it merges the parts of the query
"=" is implemented by code
moritz I think we are talking about different levels... 09:26
Alias_ There's also things like "approx"
moritz if I do a $object->add_condition($foo => $bar)...
Alias_ you don't
moritz why not?
Alias_ conditions are (subject, condition, value, value, ... ) 09:27
like in rdf
Although obviously, nothing like rdf :)
A condition expression is three parts
moritz ok, add_condition($foo, '=', $bar)
how does it know how to interpret $foo and $bar?
Alias_ $foo is going to refer to some model element 09:28
So it's split by . and mapped to the entity model
moritz ok
and $bar?
Alias_ Depends on the condition
moritz let's assume '='
Alias_ sql_equal_to( $resolved_property, $bar ) 09:29
sql_equal_to knows it needs $bar to be of data type $resolved_property->data_type
moritz and how does that handle $bar?
Alias_ And it returns [ 'sql fragement including a ?', $bar ]
fragment 09:30
moritz ok, so $bar is _alway_ treated as constant?
Alias_ ah, I get you
moritz s/alway/always/
Alias_ you want column A = column B
moritz I want to know how it dispatches between string B and column B 09:31
Alias_ You'd need a new condition
add_condition( $foo, 'column-wise equal', $bar )
I don't currently have an implementation of that
moritz ok
then I'm happy for the moment ;)
Alias_ But it's probably possibly
resolve the properties of $bar like for $foo, make sure that they are in the same table 09:32
moritz ... or that you do a join
Alias_ I don't do joins
moritz ok
Alias_ It's all done with subselects
I mostly use postgres, which can do query rewriting across up to 7 subselects and turn it into The Right Thing
moritz is that efficient? 09:33
ok
Alias_ There's three choices
moritz ends up saying "ok" all the time
Alias_ 1. Implement it in wetware (my head)
moritz ?eval "ok" xx 5
evalbot_r16959 ("ok", "ok", "ok", "ok", "ok") 09:33
Alias_ 2. Implement query rewriting in Perl (also from my head)
3. Implement it inside the database (not in my head)
wolverian postgres++ # not crap 09:34
Alias_ Also, do you have any idea how complicates a join would be for self.parents.friends.children.scho....pets.name
which goes across at least one many to many
moritz no, I don't, and I don't want to have an idea ;)
let other people handle the complexity ;) 09:35
Alias_ The nice thing about subselects is that while it might not always be stunningly fast, you can reach 100% coverage for all relationship types
recursively
I do a couple of small tricks to limit the problem
if you had a self.parents.friends.children.scho....pets.name and self.parents.friends.school.pupils.name conditions, it would merge the two at the "school" level 09:36
So it merges conditions in a tree-like shape (in entity relationship terms)
Which seems to work
If you run explain on postgres for equivalent join and subselect queries, it actually implements it in terms about halfway between the two 09:37
in processes
table scans, index commands etc
Anyways, my code does some funky stuff when you say $object->store 09:38
Because the values might need to be stored in multiple SQL commands
So it pulls transactions internally sometimes from the connection manager
And hence, I need a fairly robust connection manager
moritz that doesn't sound very portable across databases 09:39
Alias_ codegen :)
I have implementations for MySQL, Postgres and Oracle
SQLite I can't build, because it doesn't have sequences
But for the most parts, it does work properly between the three
moritz I witdraw my last comment ;)
Alias_ The speed penalty is probably worse on some database 09:40
But really, I only added MySQL support because I felt I had to, not for the speed
moritz ;)
Alias_ And because I expected client-requests for it
It's already slower, since I have to use innodb, and emulate sequences 09:41
wolverian sqlite doesn't have sequences?! weird, I thought it did
Alias_ nope :(
He added auto_increment instead 09:42
wolverian oh
Alias_ You can't get a table id before insertion
Which is absolutely critical
wolverian not first class, then
Alias_ ya
And you can't emulate it
wolverian too bad.
thanks for telling me that :)
Alias_ Because sequence actions need to occur outside of transactions
And SQLite has a global lock
So if you are inside a transaction, and you try to pull a sequence value, it deadlocks 09:43
If I REALLY need it, I'll write an emulation implementation that does the sequence table in a seperate database 09:44
And according to audrey, there's a new thingy that lets you connect to two different sqlite files as one virtual database
If the locks are file-granular, that might work
I've never had the time for a proper proof of concept though
wolverian sequences for sqlite might be an interesting soc project 09:45
pugs_svnbot r16960 | fglock++ | trunk: 09:57
: [kp6] perl5x - fixed code-blocks
diff: dev.pugscode.org/changeset/16960
lambdabot Title: Changeset 16960 - Pugs - Trac
chris2 i need some quick help from a chinese. any here? 11:49
bloonix hallo lichtking 12:13
s/g/d/
lichtkind hi bloonix
moin
ja noch bin ich prinz :)
bloonix ein lustiger schreibfehler :)
wie geht es dir so? 12:14
sorry! how are you?
lichtkind nene schon gut deustch is mir lieber tschechisch grad am liebsten 12:15
ganz gut danke bin nur etwas verwirrt 12:16
bloonix ich dachte an die anderen hier :)
was ist denn los?
lichtkind bloonix: your right i had 5 tabs in my irc client an thought im in perlde, what are you doing here? infected by my article ? 12:19
bloonix: btw its ready now if you like to see it 12:20
moritz welcomes the german perl6 guys ;)
lichtkind hihi moin
bloonix I had a bit trouble on perlde so I though to avoid it. yes I read your perl 6 introduction on p-c.de and wanted to start a bit and where could I get a better start as here? :) 12:21
hi moritz 12:22
moritz bloonix: perl-6.de/tutorials/ if you want a german introduction
lambdabot Title: Perl 6 - Tutorials
bloonix moritz: already done ;) 12:23
moritz bloonix: then what more introduction do you want?
bloonix on p-c.de are a lot of shortcuts on one site. I like that
moritz something like a cheat sheat? 12:24
bloonix I need a lot of code examples
to know how can I write nice letters in Perl 6 ;)
moritz bloonix: then look into the pugs repository... 12:25
in examples/, or in src/perl6/Prelude*
or for regexes in src/perl6-Perl-6.0.0-STD.pm
lichtkind bloonix: do you mean the perl 6 tutorial on p-c.de wiki, there is also link to perl-6.de/tutorials/ i 12:27
lambdabot Title: Perl 6 - Tutorials
bloonix moritz: I should put a link on my home page to your introduction 12:28
moritz bloonix: do that please ;)
bloonix: and if you have suggestions what else to add, please let me know as well
bloonix lichtkind: I read the article if you like 12:31
lichtkind moritz: =<> goes also in perl6 like =$*IN ?
bloonix: cool but its now cold and gone, (third deadline :)) 12:32
moritz lichtkind: afaict it's like the p5 <@ARGV> 12:33
moritz lichtkind: but I'm not entirely sure about that, maybe it's =$*IN 12:34
lichtkind moritz: so maybe you could this clear up in perl-6.de/tutorials/anfaenger/variablen
lambdabot Title: Perl 6 - Variablen
moritz lichtkind: yes, I should 12:35
lichtkind moritz: it should be more entertaining :) exactly that whyt you disliked on my artikles :) 12:38
i mean the trivial style
lichtkind moritz: i dont grok fully your structure but its growing so it will be good :) 12:41
moritz lichtkind: I'm not so good in writing entertaining stuff ;) 12:41
lichtkind: what don't you grok? the different tutorials?
or the directory structure? 12:42
or how the menus are built?
lichtkind moritz: the idea of the structure of the whole site
also the design could support the content more
moritz I'm not aware of a very elaborate structure - just a normal tree structure like most other static sites 12:43
lichtkind: could you tell me more about the design, and how I could improve it? 12:44
lichtkind moritz: when i had more time, now im fiddeling on my own project and preparing my next foo article because im next month away in spain 12:45
moritz: this time a non perl 6 topic
moritz lichtkind: allright, no hurry ;)
lichtkind moritz: but as i promised at the bottom on the article in next foo will also a small perl6 in depth article maybe about operator for scalars 12:46
lichtkind moritz: last night i hacked Alien::Wx into my project and a bunch of problem fell over me :) 12:47
moritz lichtkind: I think there are more exicting topics in p6, if that's what you want
lichtkind moritz: i want a step by step in depth tutorial and thats something simple to start with, but think there are exciting new things in it like chained comparsion type forcing and so on 12:48
moritz lichtkind: aye
lichtkind good :) see you have to take a shower :) 12:50
and my new linux machine also seeks my attention
but a good perl 6 tutorial is pretty high up on my todo 12:51
pugs_svnbot r16961 | fglock++ | trunk: 13:04
: [kp6] fixed code-block compilation inside regex
diff: dev.pugscode.org/changeset/16961
lambdabot Title: Changeset 16961 - Pugs - Trac
pasteling "evalbot_r16960" at 193.200.132.135 pasted "Pugs build failure" (6 lines, 278B) at sial.org/pbot/26175 13:06
masak fglock: what's the progress status of kp6? 13:11
fglock masak: it's following the plan, roughly - svn.pugscode.org/pugs/v6/docs/kp6-roadmap.pod 13:12
lambdabot fglock: You have 2 new messages. '/msg lambdabot @messages' to read them.
fglock currently improving the regex engine
masak fglock: ok
fglock next phase will implement more library functions 13:13
moritz fglock: the kp6 prelude, so to say?
fglock yes, and it is missing some low level classes too 13:15
these are the things that it needs in order to run the Perl6-STD grammar 13:17
agentzh wants to hack on kp6. 13:19
fglock agentzh: there are some simple things to do 13:20
avar agentzh: I started fiddling with writing grammars in PCR, noticed some differences from p6
agentzh oh? 13:21
avar like when I wrote C<< token digit { <[0..9]> } token digits { $<num> := [ <digit>+ ] } >> in p6 I could do:
agentzh there are always simple things :)
avar "123" ~~ /<digits>/; say $/<num>;
but to get that in pcr I have to do: 13:22
C<< token digit { <[0..9]> } token digits { $<num> := [ <digit>+ ] { return { num => $$<num>> } } } >>
agentzh avar: really?
lichtkind fglock++
avar I.e. it doesn't automatically make named and numbered captures part of the result object..
agentzh avar: i think it can be easily fixed by adding tokens to Pugs::Grammar::Base 13:23
avar agentzh: I just wrote that up by hand but yes, roughly
What? Adding a token digit {} ?
diakopter Alias_: you still around?
agentzh avar: the match object is okay in PCR, no?
fglock avar: that should work
agentzh avar: yup
avar I'm talking about the general idea of result objects not a digit token in particular 13:24
agentzh avar: named captures: $match->{'name'}
avar: numbered captures: $match->[num]
avar I had a testcase where the grammar I was porting from pugs didn't work at all unless I did what I said above, manually made a hash of things that were already being captured as $<> := [] 13:25
I'll try to get it to you later-ish
agentzh avar: okay
fglock avar: [] captures a string, always (but let me check) 13:26
fglock if you use / $<x> := [ some captures here ] / then you lose the data structure, and only get the matched string 13:28
you need to use parentheses instead
avar I'll try that (later) 13:29
but whatever the issue was it was behaving differently in pugs and pcr 13:30
moritz wow, somebody found my homepage searching for "pugs" on result page number 47 (!) ;) 13:37
pugs_svnbot r16962 | fglock++ | trunk: 13:49
: [kp6] v6/docs/p6regex-on-p5regex.txt - progress report
diff: dev.pugscode.org/changeset/16962
lambdabot Title: Changeset 16962 - Pugs - Trac
pmurias hi 15:26
moritz hi pmurias ;)
masak hello pmurias 15:30
diakopter Juerd: ping 15:37
Juerd diakopter: pong
diakopter Juerd: may I ask the status of getting more ram for feather? (merely curious) or moving it to xen on a big(ger) box? 15:38
moritz Juerd: if money is still the problem, i might donate something (not enough for a bigger box, but probably for more RAM) 15:40
Juerd diakopter: Microgrant was approved; I'm waiting for payment info. 15:42
moritz cool
Juerd moritz: Money is no longer the problem, but any donation would still be welcome. Every dollar or euro can help to get bigger CPUs, or perhaps a spare hard drive :)
moritz Juerd: I'm willing to donate up to 150€ if you tell me what you do with it ;) 15:43
Juerd moritz: A bigger CPU in the new hardware. 15:44
moritz Juerd: write me an email with your IBAN and BIC code, then I can transfer the money to you
Juerd moritz: Which address? 15:45
IBAN: NL62ABNA0578740923 15:46
BIC: ABNANL2A
Note that I have to charge 19% VAT btw
If you have a european business, I can send you an invoice and it's 0% then
moritz Juerd: I do have
Juerd Please let me know the VAT number 15:47
moritz I don't have it yet, I only opened my buisiness the other day :(
Juerd Oh. I can send you the invoice later
moritz after the bank transfer? 15:48
Juerd moritz: In any order you wish :) 15:49
I don't know if your country allows payment without prior invoice.
moritz I hope so ;)
moritz Juerd: I've done my transfer now, you should receive the money during the next few days 15:50
Juerd moritz: On behalf of dozens of feather users, thank you :) 15:53
moritz this is so cool - in the EU transfers to other countries may not cost more than within the country, so it's free for me ;)
Juerd: you're welcome
Juerd Yes, it's free -- iff you say "shared costs"
If you say "I pay everything" or "recipient pays everything", then voila: expenses!
Which is a problem because many people pick "I'm paying" because they don't want the recipient to pay... 15:54
moritz Juerd: if we host that p6explain/decrypt/perldoc project on feather I'll want an account - right now I'm fine without
I didn't have a choice :( 15:55
Juerd moritz: If you want an account, please request it by email.
moritz regarding the paiment thing
Juerd That's good :)
moritz Juerd: Aye, I'm just moosing randomly...
Juerd Being forced into the free option is good.
moritz btw what location in the pugs repository would you suggest? 15:56
Juerd I haven't looked in pugs's repository for a few months, so I wouldn't know 15:57
lichtkind Juerd: hello 16:06
[particle] mortiz++ for volunteering currency 16:10
moritz [particle]: use your tab key if you want to karma me ;-) 16:11
[particle] d'oh!
moritz ;)
[particle] particle-- ;)
masak moritz: re location in the pugs repo: I'd say ext/ 16:23
though I'm by no means an expert either
agentzh ext/ ?
moritz masak: I thought about it, but ext/ is reservered for P6 modules...
agentzh is it a Perl 6 module?
moritz: right :) 16:24
moritz masak: and it'll mostly be a collection of data, with some p5 and p6 modules perhaps
masak hm
moritz I'll use misc/ for the moment
masak sounds good
moritz if it gets a lot of contributers, we can still move it to the pugs root or somewhere 16:25
Juerd lichtkind: Hi. 16:28
Limbic_Region takes bets on how long he will last before being disconnected 16:34
lichtkind Juerd good to see you, i also wasnt much around
Limbic_Region looks around for Juerd but doesn't see him 16:35
ahh, there he is as an op
moritz good disguise, such a doughnut ;) 16:37
pugs_svnbot r16963 | moritz++ | trunk: 16:50
: Added P6Explain, the first attempt to build a character based syntax database.
: Comments and contributions are welcome!
diff: dev.pugscode.org/changeset/16963
lambdabot Title: Changeset 16963 - Pugs - Trac
pasteling "evalbot_r16962" at 193.200.132.135 pasted "Pugs build failure" (6 lines, 278B) at sial.org/pbot/26180 16:51
pmurias mortiz: a possible name for the project might be p6hints 17:29
* moritz 17:30
[particle] perlhints or perlexplain or perlnurse *g* might be better 17:31
moritz pmurias: I like that name ;) 17:32
pmurias mortiz: good ;) 17:35
pugs_svnbot r16964 | diakopter++ | trunk: 18:06
: new svnindex style... let's see what it looks like.
diff: dev.pugscode.org/changeset/16964
lambdabot Title: Changeset 16964 - Pugs - Trac
pugs_svnbot r16965 | fglock++ | trunk: 19:05
: [kp6] perl5rx - the $/ under construction is available inside code blocks; but for now it is a copy
diff: dev.pugscode.org/changeset/16965
lambdabot Title: Changeset 16965 - Pugs - Trac
pasteling "evalbot_r16964" at 193.200.132.135 pasted "Pugs build failure" (6 lines, 278B) at sial.org/pbot/26182 19:09
mncharity fglock: ping? 19:33
fglock_ mncharity: pong 19:49
mncharity: I'd like you to sanity check p6rx-in-p5rx if possible 19:50
it's a new kp6 regex backend 19:51
mncharity nifty 19:55
how can I help?
fglock_ I'd like if you could take a look at the architecture and comment on it 19:56
fglock_ I'm afraid it can have scalability problems, or hit perl 5 regex bugs 19:57
WiNK79 Hi guys..can any one help me with adding perl-devel package onto perl5.8.8
i cannot find where I can get the Windows version of the package..? i tried using the cpan shell.. 19:58
but nothing was found..
mncharity fglock_: looking... 19:58
fglock_ WiNK79: I don't think there is a perl-devel for Windows - but you should ask at #perl
WiNK79 oh...ok....thanks...will do.. 19:59
fglock_ mncharity: there is a sample code here (with instructions for compiling/running) svn.pugscode.org/pugs/v6/v6-KindaPe...en-test.pl
lambdabot tinyurl.com/2hf4xk
mncharity ah :) I was about to ask "uhm, where?" 20:00
fglock_ hmm - please comment out the { return 123 } block - it makes the output very verbose by dumping the whole kp6 object hierarchy :P
the "docs" are here svn.pugscode.org/pugs/v6/docs/p6reg...5regex.txt 20:01
WiNK79 hi fglock: no response...:P 20:02
fglock: from #perl
fglock_ the original prototype is here (easier to understand) svn.pugscode.org/pugs/v6/v6-KindaPe...subrule.pl 20:02
lambdabot tinyurl.com/29lv29
fglock_ WiNK79: how about #perl at irc.perl.org 20:03
WiNK79 fglock: ok..will try that one..thanks
fglock_ WiNK79: or perlmonks.org
fglock_ mncharity: btw, the root directory for running kp6 is v6/v6-KindaPerl6/ 20:05
mncharity: the idea is that the match creates a linked-list of operations, which grows and shrink as the perl5 regex matches/backtracks; the list is converted to a Match object on demand 20:07
fglock_ the linear structure is much easier to manage than the match tree 20:07
mncharity ./lib5/KindaPerl6/Runtime/Perl5/Match.pm vs lib5regex/KindaPerl6/Runtime/Perl5/Match.pm ? 20:10
ah, neat idea 20:11
fglock_ lib5regex is the p6rx-on-p5rx Match class; lib5 is for the mp5 regex engine
mp6
I mean: lib5regex/ is the p6rx-on-p5rx runtime; lib5/ is the mp6 runtime 20:13
mncharity so it's just that one file in lib5regex/ I should be looking at?
oh, no docs, and output of the test script 20:14
fglock_ yes, that's the runtime; you can see the compiled script by piping into perltidy
perl kp6-perl5.pl --perl5rx < examples/token-test.pl | perltidy 20:15
mncharity re "after the match finishes, the operations are interpreted, and the result is a Match object", you will also have to do it during the matching, if the regex itself uses the match. 20:17
fglock_ the emitter is: svn.pugscode.org/pugs/v6/v6-KindaPe...l5Regex.pm
lambdabot tinyurl.com/2h6f5d
fglock_ yes, it does that
it is currently a copy, but I can make it rw by using references
or by making a diff 20:18
pugs_svnbot r16966 | diakopter++ | trunk: 20:26
: oops, didn't mean to commit that :/
diff: dev.pugscode.org/changeset/16966
lambdabot Title: Changeset 16966 - Pugs - Trac
mncharity looks around for fglock_ 20:40
Aankhen`` Idiot moment. Could two conforming Perl 6 implementations generate different ASTs for the same snippet of code, all other things being equal? 21:54
lumi Aankhen``: I don't think we have an AST spec 21:57
Also, I think you can optimize on the AST level
spinclad Aankhen``: sure
Aankhen`` Thanks.
I felt like the answer was obvious, but for some reason I had a mental block. 21:58
(Hence the "idiot moment". I tend to have a lot of those. :-P)
nothingmuch w 12 22:40
dduncan I would expect there to be an AST spec, at least insomuch as that the syntax for writing Perl 6 macros can take or return ASTs 22:50
avar the std parser is pretty much an AST spec 22:51
dduncan mind you, that could just be a role/interface to the ASTs, and not be how they are implemented internally
[particle1 perhaps once somebody is near the point of implementing perl6 macros, a standard will develop. 22:53
mncharity @tell fglock btw, I've comments, whenever is convenient. 23:57
lambdabot Consider it noted.

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!