SVK users: search.cpan.org/dist/SVN-Mirror/ | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org | www.treehugger.com/files/th_images/paradigm.jpg
Set by audreyt on 17 August 2006.
00:02 rgs_ joined 00:05 justatheory joined
diakopter seen Alias_ 00:07
Alias_ yeah
diakopter i saw your use.perl.org post about cpan 00:09
i haven't watched the video you linked yet
but I had some ideas to suggest
Alias_ shoot 00:10
Although doing it in the comment would help, so others can see
instead/as well as here
diakopter true. 00:11
here, then there.
java has the Java Class Library 00:12
MSIL has the Base Class Library
Alias_ I actually think that's a fundamentally flawed method
diakopter explain? 00:13
Alias_ How many times have you had to port all your programs because Perl released a new version and all the modules changed
The most common answer is "never"
or "not since Perl 5" 00:14
I spend time with Java programmers, and half the conversation seems to be about porting from J2EE X to J2EE Y and so on
Also, it requires that you plan ahead on the major APIs, with committees and such 00:15
diakopter how is that an artifact of perl not having an officially blessed set of modules
Alias_ And then hope you pick right
Perhaps I'm reading your comment wrong 00:16
Your suggestion was a raw statement of fact, without any assertion or suggestion
diakopter hee; I didn't make a suggestion actually
Alias_ Try restarting with your actual suggestion
zgh the lack of central design means that many different people produce numerous solutions, then the best one can rise to teh top rather than being prefected ahead oftime
that's the great thing about CPAN
Alias_ Plus. Java has money behind it 00:17
diakopter right, but that relies on random choice
Alias_ random?
zgh natural selection is not random
Alias_ It relies on non-random choice actually
diakopter no one fully researches all the possible implementation choices before choosing one
and all their pros and cons
zgh no but they get suggestion on #perl
Alias_ They will if there's only 2 or 3 00:18
diakopter if they did so, they would have been able to write their own in shorter time
their own amalgamation, perhaps.
Alias_ For most subjects, CPAN has between 1 and 3 choices
zgh not EVERYONE needs to do that, jst like not EVERYONE needs to hack source of OSS to be successful
er s/of/for/
Alias_ It's a few very specific areas that attract modules like honey
Config files, OO abstractions, templates, MVC web frameworks
Roman numerals are something of an unusual exception to that 00:19
Unless there's something sexy about the field of roman numberals I don't know about
diakopter earlier, sorry for prefacing my suggestion with some lead-in
zgh in the words of audreyt, "CPAN is the language, Perl is the syntax."
Alias_ The other effect we see is that someone uploads shit, but then nobody else would have upload code that does the math to solve optimimal naval prop shaft design 00:20
optimal
And then as people use it, they send feedback and patches and such
And what was originally shit, becomes not bad, and then decent, and then good
diakopter yeah.
Alias_ unless it has a terrible API, which you can't really recover from 00:21
So the modules@ list often asks people to go talk to the existing authors
zgh Alias_: can you post the URL to your referenced blog post?
Alias_ I did it today in fact
diakopter the problem is, the patch submission and bug reporting and rating systems for many modules are absolutely terrible.
oops, scratch that "the problem is,"
Alias_ the patch submission and bug reporting and rating systems are identical for all modules
00:21 mdiep joined
Alias_ Thanks the gods 00:21
diakopter I mean, many of the maintainers don't reply to patches or bug reports, and probably most users don't submit them 00:22
Alias_ use.perl.org/~Alias/journal/30724
lambdabot Title: Journal of Alias (5735)
zgh Alias_: thx
Alias_ diakopter: If the maintainers don't reply, then you just apply to take over maintainership
And we (the modules@ pause admins) give you co-maint
zgh diakopter: this is an old "problem" that free/open source handles very well 00:23
not everyone needs to participate at that level for it to work
Alias_ zgh: And CPAN better than most
diakopter: Any modules you'd like to take over?
diakopter zgh: that doesn't mean there isn't room for improvement, even in cpan
zgh Alias_: agreed
diakopter: too true, there always is
but I tink CPAN is the most successful project of its type 00:24
if you can call it "a project"
Alias_ CPAN has almost no major problems
And it's quite unique in that regard
Not at large scales anyway
It has the usual supply of bugs and smaller issues that need fixing
zgh CPAN really is much more than a library archive at this point. It's just important to perl as the language itself.
diakopter that statement could be myopic, several-many years from now
zgh diakopter: which statement? 00:25
Alias_ diakopter: Good thing I keep up with the times :)
diakopter "CPAN has almost no major problems"
what are the exceptions (the major problems)
zgh diakopter: well what are you worried about specifically?
Alias_ CPAN Testers is a problem, it's dead largely 00:26
And there's some rippling waves of QA issues as a result
diakopter any others that you're willing to mention here :)
Alias_ ponders
zgh read the post. 00:27
Worying about too many choices misses the point.
there's already python
Alias_ I have a few lingering concerns about metadata
diakopter I personally am not worried about too many choices, too much. 00:28
zgh developers should take responsibility for researching what meets their needs.
python has a BDFL that likes to pick the "official" implementation of everything 00:29
i hear he tapped django as the official web framework
Alias_ yeah
diakopter Alias_: more on the metadata concerns?
zgh sucks for devs of different frameworks that might work better in some cases
00:29 nekokak joined
Alias_ diakopter: I am concerned the metadata is being specified by one subgroup without sufficient consultation 00:29
00:30 xerox_ joined
Alias_ I have a few general installation toolchain interoperability concerns as well 00:30
But those are being fixed, so from the large viewpoint it's solved 00:31
But everything else looks ok for the next doubling at least
i.e. To 20,000 modules
After 20,000 modules I think we're going to maybe outgrow RT 00:32
diakopter i thought best practical was reimplementing RT in jifty?
Alias_ But we have about a 2 to 2.5 year lead time on that problem
That would be nice
I think more that the design might start hitting limits 00:33
Hard to explain without deep understanding of the nature of the growth
And it's at least 50% gut instinct
CPAN Testers didn't reach a single critical point either, but 10,000 about where it's scaling limit was 00:34
diakopter who was it that was spec'ing that massive scale testing system
Alias_ me, actually 00:35
diakopter ah
Alias_ But as for the rest, transport is fine, the client situation is mostly ok, governence is ok, search and reputation is ok
Windows is looking much better
zgh glad to see people asking the hard questions about CPAN
Alias_ downstream could be better, but it's hard to know how to fix that
I'm sort of hoping metadata makes downstream better... 00:36
diakopter who was it that was mentioning cpanp and its future
Alias_ kane
Who wrote it
(probably)
diakopter here's a hypothetical for you. feel free to shoot it down, but i'll keep going with it as long as I can still answer most of the hard questions 00:37
hmm; more difficult to verbalize that I thought. 00:39
anyway, sorry to be so far offtopic 00:41
i'll continue when I gather my thoughts
zgh CPAN is more topical than ever
Alias_ CPAN is always topical 00:42
And yeah, things coming to a crunch for perl 6 too
oh, one major problem, although less important than QA, is UI design 00:45
Which hasn't been high enough priority on anyone's list to target
Our design is fragmented and largely sucks
I doubt we can unify things, but a general refresh would be nice 00:46
We don't control the user entry points particularly well atm
diakopter here's a open question that I won't be offended if no one answers seriously
in what markets does perl have the largest room for growth/incursion 00:47
s/the largest/any/
Alias_ Light Business Desktop 00:48
Prebuilt web apps 00:49
diakopter your choices for markets to discuss could include such spaces as company size, development/deployment platform, decision-maker role, etc.
personally, I think trying to attract *more* developers directly is a total lost cause 00:50
Alias_ I don't think we've done that for ages
diakopter most people who call themselves developers don't (didn't?) have much choice in what language/environment they use 00:51
alright, I'm just trying to think about all the options
00:51 hikozaemon joined
diakopter sorry to think extraneously 00:51
heh
Alias_ I think the biggest one is that we're on the edge of being a VB replacement
Something to write all those incidental business desktop applications in 00:52
diakopter technically, perhaps. try selling that to more than 3% of system architects
Alias_ "selling"?
We aren't in the business of selling anything
Just like with the biotech stuff, one day you wake up and everything's written in Perl 00:53
And CGI
diakopter yeah, during times of explosive segment growth.
Alias_ certainly
We adapt fast
diakopter I think it would be folly for the perl community to fail to market/position its product and its selling points appropriately 00:54
Alias_ We don't market of position anything
or
I'm only commenting on areas in which I see the potential for explosive growth
diakopter which areas are those
Alias_ The desktop one is a case of a number of factors all starting to come together at the same time 00:55
diakopter yes; there's a large saddle point here with the release/demise of Vista
potentially.
Alias_ It's nothing to do with that, although I just tested Vanilla 7 on Vista and it works fine
diakopter which factors are you talking about then wrt the desktop 00:56
Alias_ It's that a few particular technical areas required for Windows are starting to move from kinda solved, to competantly solved
diakopter oh, you're talking about perl on windows.
Alias_ Firstly, Perl + Windows is going through a revitalisation as we finally escape from ActiveState
That was a sticking point for a long time 00:57
Secondly, PAR is starting to mature
diakopter i thought you were referring to perl as a development platform for desktop apps on top of linux
Alias_ Thirdly, Wx and SQLite are becoming stable
Perl as a development platform on the desktop, regardless of operating system
diakopter ok 00:58
Alias_ As in, you can install desktop applications from CPAN that work on Windows, and Mac, and Linux
We're not there yet, but we're at about early beta of that concept
zgh Alias_: you got me really excited for a second.... I want desktops apps directly from CPAN! 00:59
Alias_ Throw in PAR to make packing and isntallation cleaner, and SQLite as a sane bundled database, add the talk-to-anything power of CPAN, and you have something interesting
diakopter most markets are made. it's very rare that market leaders emerge in markets such as biotech and cgi scripts without intensive planning/marketing.
Alias_ zgh: App::GUI::Notepad!
zgh: It works right now on all three
zgh scrambles to try it
diakopter my point is that perl might grow for a while, but without careful positioning, its growth rate will be limited. 01:00
Alias_ That's fine
diakopter in other words, with some effort, growth could be improved.
Alias_ I'm perfectly happy with Perl occupying it's natural market share, and not overstretching itself
diakopter right, b/c you have a job in it (as do I) 01:01
Alias_ I own a company actually that does consulting
diakopter same
Alias_ So if everything was Perl I'd make a lot more than I do now
diakopter same effect, I mean
Alias_ But then I'm not in a hurry to do that
You don't guide and position Perl very much
It's more like gardenning 01:02
01:02 kanru joined, frederico joined
diakopter it ought to be more like biowarfare 01:02
Alias_ You just sort of encourage it here, and prune it there, and it grows out in every direction regardless
diakopter that kind of growth is nice, but better facilitation/management of growth might be better for the long term. 01:03
Alias_ And what is biowarfare like? Something incredibly expensive and dangerous that everybody does, but nobody uses?
diakopter or so I'm proposing
such as?
Alias_ How would you propose managing growth?
That didn't involve scaring people off by telling them what to do
diakopter i'm not ready to make very concrete suggestions yet; i'm still in the talking out loud phase :) 01:04
Alias_ ah
One can manage growth, generally by doing nothing
diakopter i have some ideas, but their plausibility levels wax and wane
Alias_ When Perl::Critic started despite the fact I had a fledgling equivalent, I just shut the hell up and got out the way
diakopter err; feasibility.
Alias_ And Perl::Critic has gone far further by me not "managing" it and scaring off the owners 01:05
They have the control they want, I help when it's needed, and everybody wins
Making people work for you when you don't pay them is hard 01:06
Managing people that work for you that you don't pay is harder
Generally, the only way I know how to do it is by making them work for you without even realising it
i.e. Where in normal business you can have Minions, in Perl you need primarily Unsuspecting Minions 01:07
There's very little in the way of active business-style management
01:07 crythias joined
diakopter agents on a need-to-know basis 01:07
Alias_ Except occasionally to stamp out bad ideas before they spread
Like I said, it's a bit like gardenning 01:08
diakopter pugs is a bit like biowarfare.
Alias_ Perl on the Desktop will happen... we're currently at Notepad sophistication, but that's improving on it's own just fine
01:10 mjk joined
diakopter brb; my lap-heater is about out of juice 01:10
anyway. 01:14
mjk how to from ....> return to pugs>
can help me? 01:15
diakopter mjk: what do you need
oh
01:17 markstos joined
mjk i'm learning pugs in command line, but when i input a error command(for example, print jfklsdj ";), pugs displays ....> 01:18
i can't return pugs> status
and can't input other perl statements
how to return pugs> 01:19
diakopter Ctrl-Z will leave pugs
at least in my terminal
01:19 Alias_ left
mjk thanks , but no method to return pugs> 01:19
diakopter well, in the case of your example, 01:20
";
zgh diakopter: Ctrl-D in unix will leave pugs 01:24
diakopter: not that you were asking that :-)
diakopter zgh: care to continue re: cpan? 01:25
mjk i see
thanks
zgh CPAN is the language, Perl is the syntax! 01:26
zgh cheers.
svnbot6 r12592 | fglock++ | v6 - fixed the 'rule' declaration; 01:50
r12592 | fglock++ | - fixed a bug in the rule "metasyntax" parser;
r12592 | fglock++ | - fixed $/, $()
r12592 | fglock++ | - implemented syntax:<categories> in the parser 01:51
r12592 | fglock++ | - new class Pugs::Grammar::P6Term - implements the 'term' syntax category in pure p6
01:52 ferreira joined, avarab joined
markstos Here's a perhaps simple pugs bug: pugs -I seems to work like perl5, but is not listed in "pugs -h". It seems like a useful addition there. 01:55
Where I do I look to debug when there is a failure in the prelude? 'No compatible subroutine found: "&CGI::Application::import" at <prelude> line 63' 02:03
02:06 xdg joined 02:07 rashakil__ joined 02:08 markstos is now known as stosbma
stosbma markstos: try removing "use_ok()" from your test 02:09
02:09 stosbma is now known as markstos
markstos That did it! Thanks, stosbma! 02:09
Pugs Test.pm has some interesting output. Notice only two test scripts actually passed, but the ultimate conclusion is that things were "166.67%" OK! (before I fixed some tests, things were 300% Okay...) 02:47
Failed 11/13 test scripts, 15.38% okay. -6/9 subtests failed, 166.67% okay.
03:08 mako132_ joined
svnbot6 r12593 | clkao++ | v6 - Fix if/elsif/else. 03:45
03:53 mako132_ joined
svnbot6 r12594 | audreyt++ | * smartlinks.pl: Implement markstos++'s suggestion 03:54
r12594 | audreyt++ | to hide the filesystem-specific realpath in the
r12594 | audreyt++ | filename it displays.
r12595 | audreyt++ | * Pugs.Help: markstos++ suggests -Ipath should be documented 03:57
r12595 | audreyt++ | as part of "pugs -h".
04:00 kanru joined 04:07 frederico joined 04:13 cooljoke joined 04:16 lambdabot joined
xinming hmm, I'm really sorry for what I've done, I tried 'svk push', And It seems It pushed some old files to the mirror. 04:28
I hope anyone who can revert this for me, as now, I don't know what to do... :-/
hmm, It's likely some files from fglock. 04:29
Khisanth grab old version and recommit? 04:32
PerlJam stop using svk? 04:33
xinming PerlJam: ??? 04:34
PerlJam just jumping to the more radical "solution" :)
xinming Khisanth: I ever tried with my local svn server, But It seems, if you revert it, svk won't show the difference.
PerlJam: I wrote the test, and when I try to push it back, It seems there are many g while I push, 04:35
and luckily, when I saw many empty merge, I pressed C-c
:-P
Khisanth I was thinking something like svn cat -roldrev foo > foo; svn ci foo
xinming but my real content isn't pushed in.
I think I should check-out using svn and commit. 04:36
audreyt xinming: no, there's no old files 04:37
svnbot6 r12598 | audreyt++ | * More third-party/ cleanup:
r12598 | audreyt++ | System.FilePath is moved into third-party/
r12598 | audreyt++ | and upgraded to use Neil Mitchell's maintained branch.
r12598 | audreyt++ | Pugs.Rule.* is no more, as we are no longer forking Parsec.
audreyt the merges were just svnprops
svnbot6 r12598 | audreyt++ | (Four more cleanups to go before we complete the
r12598 | audreyt++ | third-party/ move as suggested by allison++.)
obra audreyt: my modinstall mirror is out of date. if schwern sent you an autoinstall patch, could you apply? 04:38
audreyt xinming: try "svk push -l" maybe?
obra: I don't deal in patches... he is a modinstall committer
obra Thanks. I wasn't an admin, so I couldn't make that happen 04:39
audreyt now you are, too
obra woot
xinming audreyt: when I'm in rev 12595, I push it, there are many empty merge to the svn server. :-/ I really don't know if the test is commited.
audreyt xinming: it wasn't... can you try "svk push -l"? 04:40
xinming audreyt: Ok.
audreyt just woke up... need to go 1)find food 2)vote for haskell-prime features and 3)return ti IRC 04:41
PerlJam xinming: wait, you stopped the merge process when you saw an empty merge? Just let it continue. All "svk push -l" is going to do is not show you the empty merges :-)
svnbot6 r12599 | yiyihu++ | r12615@Laptop: xinming | 2006-08-23 12:20:52 +0800
r12599 | yiyihu++ | Added a test for quote operators.
r12599 | yiyihu++ | Just basic test suite, with single adverb. (q:c q:h q:b etc...)
r12598 | audreyt++ | * More third-party/ cleanup:
r12598 | audreyt++ | System.FilePath is moved into third-party/
r12598 | audreyt++ | and upgraded to use Neil Mitchell's maintained branch.
r12598 | audreyt++ | Pugs.Rule.* is no more, as we are no longer forking Parsec.
r12598 | audreyt++ | (Four more cleanups to go before we complete the
r12598 | audreyt++ | third-party/ move as suggested by allison++.)
audreyt PerlJam: peace of mind is kinda important :)
xinming PerlJam: thanks. 04:42
PerlJam audreyt: yeah, it did freak me out the first time I tried to "svk push" some smallish thing and got a ton of empty merges out of it.
(even though I "svk pull" all the time)
audreyt it's an UI issue, indeed 04:43
the current output lvel of svk is something I'd expect --verbose to do
it's one of the noisest SCMs around :)
xinming but the problem is, svk push is already commited something empty to the server.
PerlJam xinming: not a problem. 04:44
04:44 yhsuan joined
audreyt xinming: yes, but that's fine 04:44
04:51 yhsuan joined 05:21 zgh joined 05:29 zgh joined 05:33 visq joined 05:34 lollan joined 05:40 zgh joined
nothingmuch gaal: it looks like a deriviative of your utf8 patch for DBD::mysql has been applied: search.cpan.org/src/CAPTTOFU/DBD-my...ChangeLog, rt.cpan.org/Public/Bug/Display.html?id=17829 05:52
05:54 zgh joined 06:02 zgh joined 06:06 lambdabot joined 06:09 zgh joined 06:14 zgh joined 06:15 agentzh joined 06:20 lambdabot joined
gaal nothingmuch: ah, excellent. 06:28
nothingmuch gaal++ 06:29
06:29 zgh joined
beppu Has anyone used perl5 DBI from within pugs successfully? 06:45
06:47 zgh joined 07:04 zgh joined
svnbot6 r12600 | agentz++ | [docs/feather/index.html] 07:04
r12600 | agentz++ | - added some more FAQs to feather's homepage at the request of Juerd.
agentzh Juerd: will add more stuff to the page some time later. :) 07:05
07:12 kanru joined
svnbot6 r12601 | agentz++ | [docs/feather/index.html] 07:16
r12601 | agentz++ | - changed the style sheet to search.cpan.org's.
07:19 iblechbot joined
agentzh where should i put the index.html in the Pugs repos for ~agentzh/syn/? 07:35
i think i'd better make it editable by others. :)
how about util/syn_index.html?
07:38 zgh joined
svnbot6 r12602 | agentz++ | [docs/feather/index.html] 07:40
r12602 | agentz++ | - added notes for the pscp utility.
Juerd agentzh: Thanks 07:43
svnbot6 r12603 | audreyt++ | * fix filepath cabal file
agentzh Juerd: :)
Juerd Hm, my checkout refuses to upgrade 07:44
2;0 juerd@feather:~/feather$ svk up 07:45
path /home/juerd/feather is not a checkout path.
It certainly was...
I'll do this later then 07:46
wolverian did you remove your ~/.svk ? :) 07:50
svnbot6 r12604 | audreyt++ | * oops, forgot to add System.Path. 07:52
Juerd wolverian: Never on purpose... 07:53
07:58 kane-xs joined 08:11 kanru joined 08:38 dvlt joined 08:39 dvlt left
svnbot6 r12605 | agentz++ | [docs/Perl6/Spec/Functions.pod] 08:39
r12605 | agentz++ | - adjusted the title format to conform with other synopses.
08:42 plisk joined 08:44 kanru joined
wolverian should zip still return a list of lists? maybe a list of captures instead? 08:47
audreyt how is that possible? 08:48
wolverian I don't know :)
I suppose that means I need to read on captures before assuming things.. 08:49
audreyt yeah :) 08:50
svnbot6 r12606 | bsmith++ | Add directory and --recurse support to prove6, just like prove. 08:51
agentzh feather.perl6.nl/~agentzh/syn/ 08:59
lambdabot Title: Index of /~agentzh/syn
svnbot6 r12607 | agentz++ | [util/smartlinks.pl]
r12607 | agentz++ | - replaced the "TITLE" dumb word at the first line in
r12607 | agentz++ | the HTML template with "Sxx".
agentzh we now have an index page.
:)
wolverian yay :)
09:04 ruoso joined
agentzh supper & 09:06
09:06 agentzh left 09:12 xerox_ is now known as xerox 09:31 elmex joined 09:32 plisk joined 09:38 b00t joined 09:46 plisk_ joined
meppl guten morgen 10:00
Patterner aloha 10:01
10:01 wilx` joined 10:02 buetow joined 10:03 buetow joined
meppl good morning patterner 10:12
10:19 JarJarBi1ks joined 10:21 pbuetow joined 10:22 visq joined, kane-xs_ joined, alinbsp joined 10:35 rgs_ is now known as rgs 11:02 aholanda joined 11:04 agentzh joined
agentzh Juerd: oh, the feather homepage is not synced every 5 minutes like what it says. sigh. 11:06
agentzh tries to find something to do. 11:07
Juerd agentzh: Indeed. My svk checkout is no longer acknowledged as an svk checkout, by svk. 11:16
Also, people have added things in that directory - as root.
agentzh Juerd: how about using ~audreyt/pugs?
Juerd So I have to investigate if I can move those things, or change the permissions, before I re-checkout.
agentzh her working copy is constantly updated by evalbot.
Juerd Haven't had time yet.
I'll look into this later 11:17
agentzh okay. :)
11:17 jferrero joined
svnbot6 r12608 | audreyt++ | * p6doc: Make it slightly more useful by allowing "p6doc s03" 11:24
r12608 | audreyt++ | as synonym to "p6doc Spec::Operator". Still need a lot of
r12608 | audreyt++ | usability work before this can be released to CPAN to replace
r12608 | audreyt++ | Perl6::Bible.
audreyt my laptop was damaged after all; now whenever it's moved or tilted too much, it reboots (and reboots, and reboots) 11:26
svnbot6 r12609 | agentz++ | [util/smartlinks.pl]
r12609 | agentz++ | - removed <hr /> from the resulting HTML. so there's no "oddly big dividers" reported by wolverian++
audreyt very annoying, psychologically :/
plan to turn it in for repair tomorrow or the day after
11:26 chris2 joined
audreyt but meanwhile I'm setting up workspace in my old laptop, so will afk for a while... 11:27
agentzh audreyt: simply throw that away. :)
audreyt I'd do that if it's clearly broken
but it appeared fine until I moved it...
agentzh hehe
audreyt but yeah.
agentzh audreyt: is there anyone working on webpugs? 11:28
audreyt no... would be lovely if you can look into it
agentzh audreyt: will have a try later. :)
audreyt another worthwhile task is make p6doc releasablle
ooh. agentzh++
agentzh :) 11:29
11:35 ludan joined 11:37 agentzh joined 11:42 avarab is now known as avar
agentzh wonders if some kind person has fixed Pod::Html installed on feather since the HTML outputs there are sexier than his own. 11:42
svnbot6 r12610 | agentz++ | [util/smartlinks.pl] 12:02
r12610 | agentz++ | - added left margin to the code snippets. this is
r12610 | agentz++ | indeed a long-overdue feature. :)
r12611 | agentz++ | [docs/feather] 12:14
r12611 | agentz++ | - added the ``syn'' subdir and syn/index.html which is
r12611 | agentz++ | the index page for feather.perl6.nl/~agentzh/syn
lambdabot Title: Index of /~agentzh/syn
12:14 plisk joined
agentzh not sure if i could access feather.perl.nl/syn/ once Juerd fixed his svk problem. :) 12:16
*perl6
12:18 bpphillips joined
Juerd agentzh: If you want, I can give you root access. 12:19
You appear to have more time available than I do 12:20
agentzh Juerd: that'll be cool. :)
Juerd agentzh: You may now execute commands as root by prefixing "sudo" 12:21
agentzh i have plenty of time recently since i'm in my summer vocation. ;-)
Juerd agentzh: Use it wisely.
agentzh Juerd: thanks! 12:22
I will.
12:22 Aankhen`` joined 12:24 yhsuan_ joined 12:30 yhsuan__ joined
svnbot6 r12612 | agentz++ | [docs/feather/syn/index.html] 12:40
r12612 | agentz++ | - use absolute urls to make the page relocatable
r12611 | agentz++ | [docs/feather]
r12611 | agentz++ | - added the ``syn'' subdir and syn/index.html which is
r12611 | agentz++ | the index page for feather.perl6.nl/~agentzh/syn
lambdabot Title: Index of /~agentzh/syn
agentzh Juerd: feather document root is at /var/www, aka /home/juerd/feather, right? 12:49
12:50 Limbic_Region joined
agentzh feather.perl6.nl/syn/ 13:01
lambdabot Title: Synopses on Feather
agentzh now the url to the synopses on feather is much shorter. :)
Juerd++ sudo++
Juerd: i'll recover the resync magic of the feather site in my crontab if you don't mind. :) 13:02
13:04 vel joined
agentzh please comment on the new feather home: feather.perl6.nl/ 13:07
lambdabot Title: feather.perl6.nl
13:10 frederico joined
agentzh oh, it's really lovely: feather.perl6.nl/syn/S02.html 13:12
lambdabot Title: S02
agentzh feels very happy.
gaal agentzh++, please see /msg 13:13
agentzh gaal: haven't seen your /msg. weird. 13:14
integral agentzh++ # totally excellent idea
agentzh inegral: thanks!
*integral 13:15
gaal: maybe try again? 13:16
integral hmm, the double quotes in those synopses are displaying as "``" and "''" rather than “”s
13:18 fglock joined
agentzh integral: looking 13:34
integral: that's difficult to fix. because ``'' are generated by Pod::Html. :/ 13:36
*is
integral: err, sorry, i think they're controled by css. :) 13:38
oh, no, no, no. apparently i need more coffee. i was looking at the dev.perl.org's pages...ah, yes, `` and '' are produced directly by Pod::Html. 13:40
13:41 SamB joined
agentzh integral: i think that's probably a *feature* rather than a bug since i'm already familiar with ActivePerl's HTMLs. 13:42
;-)
13:42 buetow joined
integral iirc search.cpan uses somehting like Pod::Simple::HTML which is supposedly "better" 13:47
13:51 cjeris joined 13:52 nothingmuch joined
agentzh integral: I've looked at Pod::Simple::HTML, but didn't find how to generate an index. 13:54
i really like indexes in the pages. 13:55
Pod::Simple::HTML lacks documentation. 13:56
okay, i've finally found the "index" interface in Pod::Simple::HTML's source code. oh well... 13:57
does anyone know how to obtain the source code of search.cpan.org? 13:59
svnbot6 r12613 | fglock++ | v6 - update some files 14:20
r12614 | diakopter++ | fixing whitespace on my entry.
agentzh integral: indeed you're correct, Pod::Simple::HTML is *very* nice.
it will be perfect if its author write some more doc for their baby. :) 14:21
one cannot figure out how to use it by himself without looking into the source code. 14:22
svnbot6 r12615 | fglock++ | v6 - minor fix in 'Term:<x>' grammar 14:32
14:35 marmic joined
gaal anyone remember how to limit output length of 'x' in perl -d? I know about o dumpDepth=4 but ISTR there's also a line length cap 14:41
14:46 justatheory joined 14:49 mdiep joined 14:50 yhsuan___ joined, yhsuan___ is now known as yhsuan
svnbot6 r12616 | agentz++ | [util/smartlinks.pl] 14:53
r12616 | agentz++ | - switched from the cumbersome Pod::Html to the cute
r12616 | agentz++ | Pod::Simple::HTML. Thanks to integral++ for
r12616 | agentz++ | suggesting this improvement.
r12616 | agentz++ | a lot of HTML-related issues are now gone immediately,
r12616 | agentz++ | and both the HTML code and the Perl 5 code look much
r12616 | agentz++ | cleaner now. :)
integral agentz++ 14:54
agentzh integral: thank you!
[particle] happily updates his working copy 15:00
i've been digging into smartlinks.pl to see what parrot needs to do to use this
agentzh wow... 15:01
particle: is there anything that i can help?
15:02 xinming_ joined
[particle] believe me, i'll let you know :) 15:02
agentzh great
[particle] we're starting to use smartlinks in the pge tests
they'll be part of the unique-id 15:03
C<< # L<S05/Bracket rationalization/capturing group> >> will appear before a group of tests related to that 15:04
agentzh yes, that's exactly what we're doing within the Pugs test suite.
[particle] the link will apply to all following tests, until another smartlink appears
agentzh yup, smartlinks.pl uses the same rule. 15:05
[particle] i want to write this in PIR
agentzh is amazed.
[particle] which means, first i need to write a perldoc parser/emitter 15:06
which, using pge/tge, isn't hard.
agentzh quite right. i did that in smartlinks.pl
[particle] yes, i saw that.
agentzh there's a pod parser/emitter as well as a smartlink parser.
[particle] are smartlinks standard perldoc (pod/kwid) format?
agentzh particle: i don't think so. 15:07
[particle] that's a shame. it should be valid.
agentzh just the simplest form of smartlinks is standard pod.
L<S04/section name>
PerlJam Convince whoever is writing the new POD spec that they're needed (I think that's Damian)
[particle] ingy/damian 15:08
agentzh where the keyword part is omitted.
good thought!
PerlJam ingy is easier to get a hold of on IRC. :)
agentzh true
[particle] are keywords space separated individual words, or are they phrases searched as one unit? 15:09
agentzh for example?
[particle] like in my example, L<S05/Bracket rationalization/capturing group> 15:10
is "capturing group" searched
agentzh currently, L<../../foo bar baz> results in the following regex: /foo.*?bar.*?baz.*?/
[particle] or is "capturing" searched, and "group"
ok
agentzh to preserve the space, you need to quote them using " or '. 15:11
[particle] okay, perfect
agentzh L<../../"foo bar" baz/>
sorry, i mean L<../../"foo bar" baz>
[particle] have you thought about spec coverage reports based on smartlinks?
plus todo/skip status, of course 15:12
agentzh in terms of how many paragraphs in the synopses have test snippets?
[particle] seeing a document highlighted all in green would make me happy :)
yes
agentzh hmm 15:13
gaal wants to see test snippet highlighted in green while you want to see synopsis text colored in green.
that's fun. :)
[particle] -Ofun++ 15:14
agentzh at last, we shall see the whole documents are in green...
[particle] green is the new orange
agentzh hehe
that's crazy.
i'll think about that later. of course. :) 15:15
[particle] of course. just let it rattle around in your brain for a while
agentzh lol
please let me see your PIR version of smartlinks.pl or something like that when you finished. 15:16
that's damn cool.
[particle] undoubtedly! but it'll take a while, as there are modules you use that haven't been written in pir yet 15:17
agentzh yes!
there's no PIR version of Pod::Simple::HTML, for example.
[particle] indeed.
agentzh particle: why not use the smartlinks.pl directly? 15:18
[particle] however, once i've got an ast, writing an html emitter isn't that hard
well, i could of course steal it...
but where's the fun?
in fact, i probably will steal it first, replacing bits until it's all pir 15:19
agentzh ah, the fun part...yeah...
[particle] btw your process_paragraph sub could be cleaned up a bit, i think
agentzh oh? 15:20
[particle] i just need to remember the correct p5regex syntax
agentzh use another module to do the dirty job?
[particle] well, you can do s/[LCFIB]<<(.*?)>>/$1/go;
and then have two lines 15:21
agentzh ah, yes. thank you.
i've been silly. :)
[particle] but i'm trying to think of the way to match same numbers of < >s
agentzh fixing
PerlJam What's with the /o ? It won't do anything useful for that RE
[particle] true, it won't
agentzh just add that by habbit. 15:22
[particle] /o is a pre-5.6ism iirc
PerlJam it's a bad habit IMHO
agentzh okay, will change that too. :)
PerlJam /o *only* ever did something useful if there's a variable in the pattern.
agentzh thank you
PerlJam: true.
[particle] pj: you remember the incantation to match the same number of > as < ? 15:23
PerlJam (I have a bit of a prejudice against /o because once it took me quite a while to track down a bug due to /o deep inside a module I was using)
[particle] so this extends to L<<<...>>> etc?
PerlJam [particle]: not right off. 15:24
15:24 kanru joined
[particle] k 15:24
can you have multiple syn-dirs? 15:26
agentzh yes, perlpod allows L<<<< ... >>>>.
[particle] hrmm, no. $syn_dir is scalar
agentzh particle: what do you mean?
[particle] i mean, i need to deal with perl6 synopses (S05 in particular) and parrot PDDs 15:27
agentzh that won't be difficult to implement.
[particle] no, should be straightforward 15:28
agentzh heh, yes.
15:39 kanru joined 15:43 hexmode joined
integral [particle]: m/(<+)(.*?)(>+)(??{ length($1) == length($3) ? "" : "(?!)"})/ 15:44
or: m/(<+)(.*?)((??{ my $z = $1; $z =~ y/</>/; $z }))/ 15:45
svnbot6 r12617 | agentz++ | [util/smartlinks.pl] 15:48
r12617 | agentz++ | - cleaned up sub process_paragraph according to
r12617 | agentz++ | the suggestion from particle++
r12617 | agentz++ | - stripped the /o regex modifiers per particle++ and
r12617 | agentz++ | PerlJam++
r12617 | agentz++ | [util/t/smartlinks.t]
r12617 | agentz++ | - fixed a bug in the test.
[particle] integral++ # the second one looks right 15:50
integral the first should too! :-P
actually, a (?!>) at the end and a (?<!<) at the beginning might also be handy 15:51
[particle] the first makes it a non-capture if it doesn't match length?
integral makes it fail, yes
[particle] backtrack and continue or fail the entire regex? 15:52
integral backtrack and continue
[particle] s/regex/match/
ah, then that'll work too :)
integral so that's why you need some anchoring at the front and back
15:56 JarJarBi1ks is now known as JarJarBinks 16:00 penk joined 16:01 kanru joined 16:02 agentzh joined 16:05 rlb3 joined
agentzh integral: surprisingly your second regex causes my activeperl 5.8.7 to crash. :/ 16:05
integral excellent! I guess it's the y/// being nested :-/
TimToady agentzh: if you want to make the ordering of the keywords independent, then you can put each of them inside (?=^.*keyword) and turn them all into lookaheads.
agentzh TimToady: actually i pretty like the ordering of the keywords is significant. :P 16:06
TimToady (I do this when searching my radical dictionary...)
it's unexpected if you call them "keywords" though. 16:07
agentzh hmm, right.
thank you, and i'll switch to the new semantics soon. :)
TimToady or differentiate ordered keywords syntactically 16:08
a b c|d e f
agentzh stares at the line.
TimToady or a.b.c d.e.f 16:09
or something
agentzh *nod*
TimToady the latter makes unordered the default, but the . is unobtrusive
and really stands for .*?
and manages to match a literal dot 16:10
if they happen by accident
agentzh yeah, that's much more intuitive.
TimToady course "a b c" can still be literal 16:11
16:11 kanru joined
TimToady or use .. for "followed by" 16:11
or ... is almost English
agentzh L<xx/xx/"i don't".."know".."if I really like it">
TimToady for 'something left out here' 16:12
agentzh L<xx/xx/'but it'...seems...'good now'>
hmm, '...' is just a bit long.
while '.' is too easy to be confused by perl's string concatenation operator ".". 16:13
so '..' is okay, i think.
indeed they're not truly "keywords" in the common sense. 16:14
[particle] how about something using perl6 syntax
TimToady so maybe it would be better to not call them keywords and keep a b c ordered
agentzh furthermore, we can have "true" keywords by using spaces as the word separator.
TimToady and use .. for omissions and | for alternation 16:15
[particle] ordered words, search anchors,
(instead of keywords)
agentzh yes, we need a better name for it.
TimToady keyphrases?
agentzh sounds good. :) 16:16
[particle] i lik it
huffman++
agentzh hehe
TimToady I think space should still imply at least \s+ so that you don't have to worry about reflowing paragraphs
agentzh i just want to keep the keyphrases' syntax as simple as possible. 16:17
16:17 iblechbot joined
TimToady maybe even .*? still to ignore X<<...>>-ish things 16:17
agentzh TimToady: X<<...>> currently is stripped by smartlinks.pl.
TimToady I'm fine with current .*? meaning 16:18
agentzh me too.
i think the problem is just the name "keyword"
TimToady *nod* and the non-capability of starting over at the front for independent words
agentzh so keyphrases may fix all the problems. :)
TimToady maybe just allow additional /phrase chunks and AND them. 16:19
(using the (?=^.*?...) trick 16:20
agentzh oh, what's that?
[particle] are the keyphrases bordered by \b ? 16:21
TimToady using lookahead to start over. Or you can just do separate searches.
agentzh particle: currently not.
TimToady maybe space should imply .+? rather than .*? 16:22
agentzh *nod*
fixing
fixed 16:23
16:23 chaoslawful joined
agentzh i'm worried about the use of \b since the keyphrases are not necessarily \w+ 16:24
TimToady only put the \b if the front or back is \w? 16:25
agentzh good point!
agentzh is happy.
16:28 rashakil joined, Psyche^ joined
TimToady wanders off... & 16:28
agentzh finds TimToady's suggestions very interesting. 16:39
16:39 Psyche^ is now known as Patterner
agentzh TimToady: btw, i really like the (?=^.*?) trick in general. :) 16:41
16:41 xinming_ is now known as xinming 16:43 jferrero joined
svnbot6 r12618 | agentz++ | [util/smartlinks.pl] 16:45
r12618 | agentz++ | - replaced .*? with .+? when concatenating
r12618 | agentz++ | keyphrases. TimToady++
r12619 | agentz++ | [util/t/smartlinks.t] 16:51
r12619 | agentz++ | - updated the unit tests to reflect the .*? ==> .+?
r12619 | agentz++ | change.
obra audreyt: can you release a new Module::Install? 16:52
16:56 chaos joined 17:02 kanru joined
audreyt kane-xs_: M::AutoInstall now checks PERL5_CPANPLUS_IS_RUNNING. 17:06
svnbot6 r12620 | agentz++ | [util/smartlinks.pl]
r12620 | agentz++ | - added \b to the resulting regexes of each keyphrase
r12620 | agentz++ | if it starts or ends with /\w/.
r12620 | agentz++ | [util/t/smartlinks.t]
r12620 | agentz++ | - updated the unit tests to reflect this change.
kane-xs audreyt++
this is a good first step.. i'll aim for something even more indicative in the future (ie, telling you what file is currently being run or somesuch)
svnbot6 r12621 | agentz++ | [t/README] 17:09
r12621 | agentz++ | - s/keyword/keyphrase/g
r12622 | fglock++ | * v6 - fixed $/<x>() and $/{'x'}()
r12623 | agentz++ | [util/t/smartlinks.t] 17:12
r12623 | agentz++ | - s/keyword/keyphrase/g
agentzh Juerd: is there anyway to run sudo via cron? 17:13
audreyt why would you want to sudo 17:14
?
obra: M::I 0.64 is up. 17:17
obra tx 17:19
agentzh audreyt: because i have to write to feather's document root.
copy files to /var/www...
audreyt don't do that... let your dir be ln'ed into it
TimToady symlink out instead? 17:20
apache will generally let you follow symbolic links by default, I think
agentzh ah, yes!
audreyt and if not, ask Juerd to turn that option on in httpdconf
TimToady or in the .htaccess if you want to limit the capability to a single directory. 17:21
(can be so limited in the httpd.conf too)
agentzh thanks!
audreyt np :) *goes back to attempted sleep* 17:22
agentzh it's 1:20 AM or so at audreyt's place. :) 17:23
TimToady 2 or 3 for you?
agentzh i'm in the same timezone as her. 17:24
TimToady ah so you're only slightly insane
agentzh :)
usually i go off at 9:00 PM.
TimToady I thought your connection usually shut down at night?
agentzh tonight is an exception.
because Juerd has assigned root access to me. :)
TimToady: right. 17:25
TimToady planning to turn feather into a Chinese spambot while Juerd isn't looking, are you? :) 17:26
agentzh TimToady: certainly not. :) 17:28
ayrnieu or a Chinese gold-farmer? 17:29
agentzh is still unfamiliar with simlinks.
TimToady cd /fake/location; ln -s /where/it/really/is/truename fakename 17:30
ayrnieu agentzh - what bothers you about them?
agentzh TimToady: thanks :)
ayrnieu: i'm setting up feather's document root. 17:31
TimToady if fakename is the same as truename you can just put . instead of fakename
PerlJam agentzh: for smartlinky goodness?
ayrnieu ( also, ln -s /where/nothing/actually/is fakename ; and ln -s ../../might/be/the/same fakename )
agentzh PerlJam: sure but more.
TimToady: trying now...
17:33 prefiks joined
agentzh Unix's symbolic links are very charming. :) 17:35
saddly there're no such things on wondows. :(
gaal there were hard links in FAT 17:36
agentzh didn't know that.
gaal AKA cross-linked clusters :) 17:37
it was considered an error
"it's a bug, not a feature"
agentzh i think i've heard that in my college textbooks. :)
but not quite sure. 17:38
gaal agentzh: unix's hard links are very charming too
they mean for example that it's fine to delete an open file.
you're actually "unlink"ing it, not deleting it
agentzh my instructor ever told us to compare symbolic links and hard links in english.
gaal the filesystem reclaims space just when the last reference is removed. 17:39
ah, okay, so you know about all this stuff :)
agentzh gaal: still quite fuzzy, since i've never tried out them in practice. :) 17:40
ayrnieu (incidentally, all files are 'hard links'. Even symbolic links are 'hard links'.)
gaal TimToady: on my ln, you can even drop the '.' if fakename is is the same as truename
ayrnieu but, geez, a reference counting mechanism? GC's are very efficient these days :-)
17:41 alinbsp joined
gaal disks being what they are, you can be assured of eventual global destruction 17:41
17:43 Southen joined
agentzh oh, help! Juerd has redirected /var/www to /home/Juerd/feather. How can i change that to ~audreyt/pugs/docs/feather? 17:43
buubot heave!
gaal agentzh: redireted, with ln? sec, logging in 17:44
agentzh k
svnbot6 r12624 | fglock++ | * Pugs::Compiler::Rule - 'regex' closures are emitted by v6.pm, if it is available
agentzh also, please redirect /var/www/syn to ~agentzh/public_html/syn... 17:45
not sure if it's possible...anyway...
gaal 1. fixed 17:47
agentzh gaal: what command did you use?
gaal 2. in principle yes, but not with the actual dir being a working copy
agentzh gaal: no problem. 2 is trivial. 17:48
gaal agentzh: sudo ln -s /home/audreyt/pugs/docs/feather/ /var/www
agentzh: I had to 'sudo rm /var/www' first
agentzh gaal: thanks!
ah
gaal you can also use -f on ln
agentzh -f means force?
gaal but only do that when you're sure you know what you're doing. 17:49
yes.
agentzh okay
forget about 2.
i have a better way to do that now.
:)
gaal okay :)
agentzh gaal++
i can finally go to sleep soon. :) 17:50
gaal good night soon :)
agentzh gaal: i was getting the error "File exists" when ln'd them. 17:51
so ``rm'' is very important.
gaal yes, because '/var/www' existed. hennce the rm/-f
agentzh got it. :)
gaal be very, very careful with 'rm' when root. there's no easy undelete. :) 17:52
agentzh is shamed to keep asking extremely basic questions on #perl6.
gaal: i hear that. :)
17:53 zgh joined
gaal agentzh: my first O'Reilley book was "Unix Power Tools". I think I still have the copy somewhere. It's really recommended, although by now probably outdated. 17:53
see if you can find a copy!
agentzh thank you! i'll take a look! 17:54
gaal (there was a brief section -- maybe a page or two? -- saying Perl rocks but that there was no room in the book for it, and that I should look out for the (then pink) Programming Perl volume. Turned out to be pretty good advice...
17:55 Limbic_Region joined
gaal ) 17:55
ayrnieu I'd imagine subtitles on many of its other sections with the text "If you already know Perl, you can probably safely skip to the next section." 17:56
17:56 Eidolos joined
agentzh perl++ 17:57
gaal not necessarily; it wasn't a programming book per se. It had nifty things like notifying you when your boss logged in and stuff like that. Wow, I miss the unix era
I mean when everybody used to log on to the same box 17:58
^ today that seems really trivial, like while who | grep, sure, but it was lots of fun when I first read it 17:59
anymoose enough nostalgia!
ayrnieu there are still such systems :-) with thin clients and whatnot.
agentzh in our lab, we students still have to log onto the same box in order to try out UNIX. :) 18:00
gaal networks with .rhosts eq '+ +'...
agentzh it's fun to see the big old machine runs out of resource every 1 hour or so... :)
18:01 zgh joined
agentzh i remember i even wrote a perl 4 script for the instructor to cleanup the shared resources with force. 18:01
or they'll have to restart the machine. :)
18:03 spoop joined
agentzh at that time, everyone in the lab was appreciating the power of perl. :) 18:03
18:06 zgh joined 18:11 zgh joined
svnbot6 r12625 | fglock++ | v6 - 'sub term:<a> {b}' compiles to $::_V6_GRAMMAR::term{'a'} 18:11
r12626 | agentz++ | [docs/feather/syn/index.html] 18:14
r12626 | agentz++ | - used relative urls in links.
r12626 | agentz++ | - made the page a bit pretty.
18:16 zgh joined, rindolf joined
agentzh i'd point /var/www to my ~agentzh/feather. 18:21
18:22 zgh joined
agentzh because i need to generate synopses into /var/www/syn/ without sudo. 18:23
TimToady well, stealing all of /var/www might be construed as antisocial, so I'd just symlink /your/location/of/syn to /var/www/syn 18:25
agentzh TimToady: thinking... 18:29
TimToady: but there's also a syn/ under audreyt's working copy.
18:30 zgh joined
agentzh that's the problem. 18:30
or just remove syn/index.html from the pugs repos.
okay 18:31
will do
rindolf Hi TimToady
agentzh so /var/www will still points to ~audreyt/pugs/docs/feather.
integral hmm, I remember mentions of packrat parsing in here last year sometime, anyone know why pugs didn't go with it? 18:33
18:35 zgh joined
agentzh integral: Audrey said there were already not much bracktracking in the Perl 6 parser so packrat parsing is not very beneficial. 18:35
*was
integral ah, thanks :)
agentzh np :) 18:36
svnbot6 r12627 | agentz++ | [docs/feather] 18:38
r12627 | agentz++ | - moved syn/index.html to syn_index.html
r12627 | agentz++ | - removed the syn subdir so that i won't have to steal
r12627 | agentz++ | the whole /var/www. TimToady++
r12628 | fglock++ | v6 - some fixes to category parser/emitter 18:41
18:42 zgh joined
fglock why do categories need names like 'term:<...>' instead of ':term<...>' ? 18:49
18:50 zgh joined 18:55 zgh joined
agentzh yay, so feather.perl6.nl/~agentzh/syn and feather.perl6.nl/syn effectively share the same directory now. 18:55
lambdabot Title: Index of /~agentzh/syn
agentzh symbolic links++ 18:56
Juerd: the feather homepage is automatically updated every few minutes again. thanks to gaal++'s help. :) 18:57
Juerd agentzh: Maybe it's nice to add a timestamp to the smartlinked synopses
agentzh++ gaal++
agentzh Juerd: there's a timestamp in the form of HTML comments. :)
Juerd agentzh: In the synopses, at the top; in the index, just after the second paragraph
agentzh: I think it's best if the timestamp is very apparent, so people immediately notice it if it gets out of date. You won't be watching it this closely forever, probably :) 18:58
agentzh Juerd: okay. will add that tomorrow...
Juerd Thanks
agentzh 2:56 AM here... :)
Juerd Whoa
Go to bed, you
agentzh sleep now &
18:58 agentzh left
Juerd Good night :) 18:58
19:02 zgh joined 19:07 ilogger2 joined
svnbot6 r12629 | fglock++ | * v6 - merge ( grammatical categories + Data::Bind ) 19:08
TimToady fglock: infix:<x>() would call the infix operator, whereas :infix<x>() would attempt (infix => 'x').()
in addition term: forces compile-time evaluation of the "subscript", I think, while :term{$x} wouldn't. 19:10
but whether or not it does that, I think it conveys the intent to the reader much better
also adverbs are allowed in places where terms aren't 19:11
also it would potentially break the autoquoting of => if it had to look on the left for syntactic categories. 19:13
and if it didn't do that, we'd have to explain why :infix was different from infix=>
fglock TimToady: so it is meant to parse as adverb postcircumfix 19:14
s/<ws>/plus/
is infix: => 'or' valid syntax? 19:16
such as: (infix: => 'x').() 19:18
19:18 kanru joined
fglock or even: (infix:<x>).() ? 19:18
or [ (infix:<x>) ] @a 19:20
coffee time here & 19:25
TimToady syntactic categories have nothing at all to do with pair notation 19:34
they're only a way to extend function names to sneak non-identifier characters into the name. 19:35
infix:<stuff>() is exactly like foo() as far as usage goes. 19:36
the magic of making sure "stuff" ends up in the infix table only happens at declaration time. 19:37
(or equivalently, import time)
so (infix:<x>).() is no good in the same way that (foo).() is no good if you wanted to call foo() 19:38
if you want a reference, you have to say &infix:<x>, just as with &foo
excuse me, a Code object. :) 19:39
19:40 jferrero joined
fglock TimToady: thanks 19:51
19:54 peterpiper joined 19:56 peterpiper left 19:58 lollan joined 20:10 renormalist joined 20:21 mdiep joined 20:24 bernhard joined 20:31 Khisanth joined
svnbot6 r12630 | fglock++ | * v6 - Pugs::Grammar::P6Term compiles! 20:33
Eidolos begins building Pugs.. 20:35
21:03 vel joined, rindolf joined, xinming joined, justatheory joined, marmic joined, ludan joined, weinig|away joined, buu joined, awwaiid joined, Juerd joined, Grrrr joined, marcus_ joined, clkao joined, Yappo joined, obra joined, Frances joined, LCamel joined, notsri joined 21:11 mdiep joined 21:19 Revision17 joined
svnbot6 r12631 | markstos++ | Link index page to the very useful "Differences" document, and also add a comment that it would 21:36
r12631 | markstos++ | be nice to auto-HTMLize this page as well.
21:39 cognominal joined 21:40 Khisanth joined
svnbot6 r12632 | markstos++ | Improve the titles on syn_index.html. I don't care were the content is stored, I care what it is. 21:42
21:44 Alias_ joined
Alias_ audreyt: ping? 21:44
The CPANPLUS_IS_RUNNING change is a mistake
(to M:I)
21:45 diakopter joined
Alias_ Because PERL5_CPANPLUS_IS_RUNNING does not actually mean that CPANPLUS is running 21:45
It reports false positives in some situations
21:48 markstos joined 21:52 diakopter joined 21:53 renormalist joined 21:55 mdiep_ joined
svnbot6 r12633 | markstos++ | Add a smart link to Differences.pod, in hopes that someone will make it work someday. 22:03
r12634 | markstos++ | Add mention of Global variables to Differences.pod, with another hopeful smart link... 22:06
22:28 mdiep joined 22:33 autark joined 22:36 mdiep_ joined 22:49 explorer__ joined
clkao win 4 22:52
23:00 larsen joined 23:01 Bit-Man joined 23:16 marksto1 joined 23:19 KxAx joined, eden_c joined
Alias_ clkao: Just to chase you up again, any thoughts after our YAPC::NA conversation on the idea of a svk gui? 23:32
Are you interested? Should I be waiting? Is it just on the low priority list?
clkao it's to be prioritised 23:34
23:36 eden_c left
Alias_ meaning... that it is pre-triage, and you have no idea if it will be done sooner or later, but it will probably be done? 23:36
obra I think "many people would like it, but many people would like a pony"
Alias_ Well, "everyone on Windows" is included in the "many" :) 23:37
But I agree, it's a pretty luxurious feature
high work certainly
So my answer is "Maybe, but don't expect it any time soon" 23:38
?
23:40 KxAx left
obra We've talked to at least one customer about it but aren't currently comissioned to develop it 23:42
23:53 hexmode joined
svnbot6 r12635 | lwall++ | Now reports "not ok" for when backticks throws exception. 23:57
r12635 | lwall++ | File finder now rejects filenames containing /-p5/.
r12635 | lwall++ | (Should look in files for 'use v6', really...) 23:58

Powered by Cro and the Raku® Programming Language.

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

Thank you!