parrot.org/ - clean up those smolders for the release!
Set by moderator on 20 October 2008.
dalek r32840 | infinoid++ | trunk: 00:00
: [rakudo] Remove trailing whitespace.
diff: www.parrotvm.org/svn/parrot/revision?rev=32840
Coke tene: can you double check 58460?
chromatic Coke, I believe I did.
In fact, I've been using that.
pmichaud if #58934 is fixed, does that mean we could close #31133 also? 00:01
00:01 dmknopp joined
particle Spider Missing Aboard International Space Station 00:03
Coke yah; it's not resolved, but we kind of avoid the issue.
00:03 Ademan joined
pmichaud isn't the issue in 31133 ultimately that we couldn't see failures in separate runcores? If we can now do that (via make fulltest), I think we can close 31133. 00:04
Coke RT#59912 ; tcl is working again, I have no outstanding issues there.
Tene Coke: I didn't provide any way to reproduce the bug.
Coke: I vote to reject that ticket.
Coke pmichaud: yup. so we don't really need the requested feature now.
Tene: which, yours?
Tene Coke: 58460, mine 00:05
pmichaud Coke: okay, closing #31133.
Coke 60616 was applied, neh?
pmichaud 638.
00:07 rhr joined 00:09 AndyA joined
pmichaud bwahahaha -- I can close #42369! 00:14
637. 00:17
Infinoid pmichaud++
pmichaud dinnertime here - back later.
00:21 apeiron joined
Infinoid allison: is www.parrot.org/smolder the URL we will use? if so, I will update PLATFORMS now. 00:21
dalek r32841 | chromatic++ | trunk: 00:26
: [GC] Made the concurrency scheduler a constant PMC. This means that it never
: gets collected by the end-of-the-world GC run before all tasks and timers and
: events (which depend on it for their destruction) get collected and destroyed.
: See RT #60622 for one example of that.
: Of course, as it's a constant PMC now, it always needs its live flag cleared
: just before it gets marked as live, otherwise it won't have a chance to perform
: its custom marking of the PObjs to which it refers. This is a workaround, but
: the real fix is to figure out safe and consistent order of destruction, and
: that's a much larger patch.
: I also tidied some related code.
diff: www.parrotvm.org/svn/parrot/revision?rev=32841
allison Infinoid: don't update until it's working 00:27
Infinoid ok, I'll use the big ugly URL for now, it's better than the old smoke link. 00:28
dalek r32842 | infinoid++ | trunk: 00:29
: [PLATFORMS] Swap the old smoke link for the new smolder link.
diff: www.parrotvm.org/svn/parrot/revision?rev=32842
chromatic Must be at 634 by now. 00:30
Whiteknight chromatic++ 00:32
nice fix
chromatic "nice"
I'll take brilliant or clever, but not clean.
Want a fun toy to debug? 00:33
./parrot t/pmc/exception_26.pir
./parrot t/pmc/exception_26.pbc
Remove .tailcall from the PIR
Whiteknight I gots to merge in that calling_convention branch tonight, I reckon
00:33 kj joined
Whiteknight mmhmm 00:33
chromatic ./parrot -o t/pmc/exception_26.pbc t/pmc/exception_26.pir 00:34
./parrot t/pmc/exception_26.pbc
:tailcall in PBC breaks the lexical association, somehow.
pmichaud 634.
Infinoid I'm still seeing a whole bunch of type conversion warnings during build, but test, testC and codetest still pass on x86_64. 00:37
particle: any objection to my updating docs/parrot.pod to clarify that win32 support only means x86? 00:38
particle no, please do 00:41
dalek r32843 | chromatic++ | trunk: 00:42
: [config] Removed an extra GCC warning accidentally committed in r32797. This
: makes GCC report a lot more warnings, but we haven't made a serious push to fix
: them yet.
diff: www.parrotvm.org/svn/parrot/revision?rev=32843
allison Infinoid: okay, working now, see www.parrot.org/smolder
dalek r32844 | infinoid++ | trunk: 00:43
: Update docs/parrot.pod: Under Supported Platforms, "win32" specifically means x86.
diff: www.parrotvm.org/svn/parrot/revision?rev=32844
Infinoid odd, I get a different page than the one smolder.plusthree.com/app/public_pr..._reports/8 gives me
/details/8 vs /smoke_reports/8 00:44
Whiteknight has the release gone out yet?
allison Infinoid: different how?
chromatic It hasn't gone out yet. 00:45
I want to touch up the Storable problem first.
Infinoid allison: /details/8 gives me a start date and some tags; /smoke_reports/8 gives me a list of smoke reports in reverse chronological order
allison Infinoid: ah, well that's the link I put in
dalek r32845 | infinoid++ | trunk:
: [PLATFORMS] Update smolder link to the parrot.org shortened version.
diff: www.parrotvm.org/svn/parrot/revision?rev=32845
r32846 | chromatic++ | trunk:
: [config] Made fulltest Makefile target stop at first test failure; moved distro
: tests to the end of the fulltest target list, as we don't pass all of those
: yet. See RT #58934.
diff: www.parrotvm.org/svn/parrot/revision?rev=32846
Infinoid well, whatever you link to, PLATFORMS now links to you, so you can leave it or adjust it however you like :)
allison Infinoid: if we want 'smoke_reports' instead of 'details', I'll do that instead
Infinoid: changed 00:46
00:46 TiMBuS joined
Infinoid allison++ 00:47
ok, I'm off, back later
dalek r32847 | chromatic++ | trunk: 00:58
: [lib] Downgraded Storable dependency to 2.12, per RT #59660. If this fails, at
: least it'll give us more data to decide which version of Storable to require
: and where.
diff: www.parrotvm.org/svn/parrot/revision?rev=32847
00:58 Whiteknight joined
dalek r32848 | chromatic++ | trunk: 00:59
: [t] Marked :immediate Class non-thawing from PBC test as TODO per RT #60648.
diff: www.parrotvm.org/svn/parrot/revision?rev=32848
01:00 ruoso joined
dmknopp does rakudo support string interpolation? Some thing like print "hi $var"; i thought I read that it did. 01:03
dalek r32849 | chromatic++ | trunk: 01:06
: [t] Marked .tailcall and lexical map non-thawing from PBC tests as TODO per RT
: #60650.
diff: www.parrotvm.org/svn/parrot/revision?rev=32849
01:06 apeiron joined
pmichaud dmknopp: yes, rakudo supports string interpolation. 01:09
it even supports closure interpolation.
dmknopp oh cool.
pmichaud rakudo: say "I found { 3 + 1.5 } dollars.";
polyglotbot OUTPUT[I found 4.5 dollars.ā¤]
dmknopp nice
what is it called in Perl 6, I was digging thru the code 01:10
is it called interpolation?
pmichaud you're looking in the code for it?
dmknopp I didnt see any references to it
pmichaud no, because it's not so much interpolation as it is concatenation
dmknopp well I wanted to add that to Cardnial
and I have no idea how to even start adding it to Cardnial 01:11
pmichaud rakudo-parse: "I found { 3 + 1.5 } dollars."
polyglotbot No output (you need to produce output to STDOUT)
pmichaud rakudo-parse: say "I found { 3 + 1.5 } dollars."
polyglotbot No output (you need to produce output to STDOUT)
pmichaud hrm.
dmknopp hmm.. ok so you split the string on { }
I was doing a grep for interpolation
chromatic ATTN: last chance to get NEWS updates and quick bugfixes in before the release.
dalek r32850 | chromatic++ | trunk: 01:12
: [t] Skipped t/op/lexicals.t when running with the -r flag, per RT #60652.
diff: www.parrotvm.org/svn/parrot/revision?rev=32850
nopaste "pmichaud" at 72.181.176.220 pasted "parse of interpolated string" (102 lines) at nopaste.snit.ch/14636
chromatic Sadly I had to file three bugs, so we're up to 636 or so again. 01:13
pmichaud it's not really a split, it's more like a scan and certain characters cause us to escape to other parsers
we can't just split on { } because the parts inside might also have curlies.
dmknopp at what point is that parse done? at grammar parse time? 01:14
01:14 bacek_ joined
pmichaud rakudo: say "Conditional { if 1 { "yes" } else { "no" } } result."; 01:14
polyglotbot OUTPUT[Conditional yes result.ā¤]
pmichaud note that the double quotes inside the curlies didn't terminate the string. :-)
dmknopp ahh ok 01:15
pmichaud yes, it's done as part of the parse. In rakudo's grammar it's the quote_expression rule.
that's currently implemented in PIR (src/parser/quote_expression.pir, I think)
dmknopp yep, found it thanks 01:16
pmichaud I'm done with bugfixes in trunk for the night -- I'm afraid I might break something for the release. :-|
dmknopp tene added that file for Cardinal, but I wasnt following how the grammar parse broke off into the other parsing done in quote_expression.pir
ill look into it more thx :-) 01:17
pmichaud in rakudo, it just calls the <quote_expression> subrule.
dmknopp yeah I wasnt sure of the correct terminology for Perl6 01:18
I noticed a note about Boyer-Moore string matching in the PGE docs, was this going to be used for longest token matching? 01:20
pmichaud we don't just match strings -- we match regular expressions.
i.e., longest token matching has to be able to treat xy*z as being part of the longest token match.
Tene dmknopp: please update NEWS with your cardinal work this month.
dmknopp right so Boyer-Moore wouldnt help 01:21
TimToady and BM has nothing much to do with LTM; it's for scanning for strings, not anchored strings
pmichaud if we do Boyer-Moore, it would probably be as a pruning optimization.
dmknopp stepehen: ok ill update NEWS
pmichaud but that's not likely anytime soon.
01:22 Whiteknight joined
dmknopp right I wasnt following the docs on LTM but I have worked w/ Boyer-Moore before. I wrote a version in java, I was wondering what the need or requirements were for PGE 01:22
pmichaud yes, the docs for PGE are very old right now -- they probably need updating. 01:23
probably when I get around to refactoring it later this month.
dalek r32851 | pmichaud++ | lex3: 01:27
: Re-syncing branch for lexicals work.
diff: www.parrotvm.org/svn/parrot/revision?rev=32851
r32852 | dmknopp++ | trunk: 01:35
: [Cardinal] updated NEWS to include Cardinal changes
diff: www.parrotvm.org/svn/parrot/revision?rev=32852
r32853 | dmknopp++ | trunk: 01:37
: [Cardinal] updated NEWS to include Cardinal changes
diff: www.parrotvm.org/svn/parrot/revision?rev=32853
pmichaud oooh, I need to get the deprecation for Closure into DEPRECATED.pod 01:39
chromatic: am I too late for that?
Coke No. =-) 01:42
pmichaud (writing furiously) 01:43
bacek_ rakudo: say eval 'my $a = 1' 01:45
polyglotbot OUTPUT[1ā¤]
chromatic It'll probably be another hour or two.
bacek_ o! It fixed!
pmichaud okay, good.
on #60648, it's my understanding that things marked :immediate *aren't* stored in the resulting PBC. 01:46
however, the immediate sub can return an object which will be frozen into the resulting PBC
see pmc/subs.pod:138 01:48
and pdd19_pir.pod:429 01:49
I'll reply to the ticket with those.
chromatic That sounds somewhat right, but when I traced the code, the :immediate sub did get thawed. 01:52
dalek r32854 | pmichaud++ | trunk:
: [DEPRECATED]: Deprecate the Closure PMC in preparation for the lexicals branch.
diff: www.parrotvm.org/svn/parrot/revision?rev=32854
pmichaud the immediate sub itself, or what it produced?
phrased differently, then -- the sub may be stored in the packfile but it doesn't automatically execute when loaded. 01:53
chromatic It definitely gets stored, but it doesn't execute. 01:54
pmichaud okay. I think that's normal behavior.
in order to get it to execute upon load or init one would add the :load and/or :init flags. 01:55
chromatic I'm not sure about :load even.
pmichaud well, to have the :pbc execute that sub upon being run from the command line, one would use :init 01:56
to have it execute upon being loaded via load_bytecode, one would use :load
to have it do it for all three cases (during compile, when run from command line, when loaded via load_bytecode), use all three flags.
chromatic Right. 01:57
01:57 jimmy joined 02:05 cotto joined
dalek r32855 | pmichaud++ | lex3: 02:18
: Bring lex3 branch (from trunk) up-to-date with latest lex2 changes.
diff: www.parrotvm.org/svn/parrot/revision?rev=32855
chromatic particle, can you review PAUSE and make sure that PARROTRE owns all of the module files? 02:26
The release is starting. 02:28
dalek r32856 | chromatic++ | trunk: 02:31
: [include] Replaced a magic constant with symbols for clarity's sake.
diff: www.parrotvm.org/svn/parrot/revision?rev=32856
r32857 | chromatic++ | trunk: 02:55
: [distro] Updated version numbers, news, and other miscellaneous files for the
: 0.8.1 release.
diff: www.parrotvm.org/svn/parrot/revision?rev=32857
02:56 wubo joined
chromatic Tarball created and branch tagged. Commit away! 03:06
dalek r32858 | chromatic++ | RELEASE_0_8_1:
: [distro] Tagged release 0.8.1
diff: www.parrotvm.org/svn/parrot/revision?rev=32858
Patterner Congratulations.
chromatic Thanks. 03:14
jimmy Congratulations too. 03:20
03:20 kj joined 03:25 Psyche^ joined
jimmy and is there any problem at www.parrotvm.org/svn/parrot/revision?rev=32857 ? 03:25
An error occured.
03:29 kid51 joined
moderator Parrot 0.8.1 "Tio Richie" Released! | www.parrot.org/ 03:34
GeJ chromatic: congratulations indeed. 03:34
03:35 Andy joined
chromatic Any parrot.org admins around? 03:43
allison yo 03:44
chromatic: need an admin?
chromatic I'm trying to create a story, but all I can create is a scratch. 03:46
allison ah, just a sec, will bump you up... 03:47
chromatic: you now have god-like powers on parrot.org 03:48
chromatic Just like in real life! 03:49
allison hmmm... who is user "Arthur Dent"?
chromatic Who are the other god-like admins with the laying on of hands and the burning, itching and the eyeballs, and the hey lady! 03:50
I'm updating the release manager guide to say "YOU NEED THESE POWERS A FEW DAYS IN ADVANCE NEXT TIME" to remind myself.
allison currently you, particle, jeff, coke, and me
we should go ahead and grant 'editor' access to all release managers at the beginning of a 6 month release cycle 03:52
kid51 chromatic: Regrettably, that test in t/pmc/pmc.t is now failing with a different SIGNAL error: rt.perl.org/rt3/Ticket/Display.html?id=60622 03:54
chromatic Hm. That shouldn't be. 03:55
kid51 Did you test it on Darwin PPC ... or just on Intel? 03:56
chromatic I know what the problem is and I reproduced it with a different test on x86-32 Linux. 03:57
kid51 Sorry to rain on your parade ... errr, your release. 03:58
chromatic S'okay.
I haven't uploaded a tarball yet.
Anyone else have Darwin PPC around?
I'm looking at you, Allison.
allison sorry, I'm Intel
chromatic (and can I say, Drupal, how in the world could you possibly break unfiltered HTML?)
allison but there's a Darwin PPC laptop in the office
chromatic On my bookshelf, four feet behind my chair, right now? 03:59
kid51 I don't insist that you hold up the release for this ... because even if we were to fix this, we still have that failure in t/pmc/packfile.t that predates this one.
allison chromatic: somewhere like that, with a big Perl sticker on the top
chromatic RT #60068? 04:00
kid51 Yes. 04:01
chromatic kid51, I'll poke at RT #60622. If it's a quick fix, I'll put the patch in the tarball.
kid51 Well, I have to go to sleep pretty soon, so I probably will only be able to re-test sometime tomorrow. 04:02
chromatic Dear Drupal, welcome to the Internet. I would like to introduce you to HTML. PLEASE READ THE RFC AND THE GRAMMAR BEFORE YOU GO MANGLING UNFILTERED HTML. Love, chromatic.
I have a Darwin PPC laptop now. I can try it.
kid51 These 2 failures can be seen in this smolder report: smolder.plusthree.com/app/public_pr...tails/8259
04:03 elmex_ joined
allison chromatic: there is a flag to turn off the Drupal magic 04:03
chromatic Is there a flag to say YOU CANNOT NEST TAGS IN WHITESPACE ALLOWED INSIDE A TAG IN HTML! REGULAR EXPRESSIONS DO NOT KEEP DUTCH PEOPLE COOL! 04:04
allison chromatic: eh?
purl Speak up, sonny!
chromatic I'm looking for a flag to disable the automatic linebreaking. 04:05
I know how to use a <p> tag.
kid51 must sleep
purl $kid51->sleep(8 * 3600);
allison chromatic: I think I disabled that one
tewk_ <pre>? 04:06
purl <pre> is pretty hard to fuck up.
allison tewk: yeah, I usually wrap the NEWS section in pre, and then just delete the extra line breaks in the paragraphs 04:07
chromatic The problem is, I've used <p> tags appropriately. I know how to write HTML. You might say I'm even a professional, considering I'm a professional editor of a website.
Drupal doesn't give me the option of NOT adding linebreaks whenever it encounters a newline, which, in HTML, is just another not-special whitespace character. 04:08
allison I can turn off auto-linebreaks, but it'll break all our other pages
pmichaud I'm so glad I kept HTML out of my wiki markup. :-) 04:09
allison chromatic: okay, I just gave you a new format "Plain HTML" that does no filtering
chromatic Thanks! 04:10
Storable is broken on this laptop.
Do you have a non-system Perl installed somewhere?
allison chromatic: the non-system perl should be found if you type 'ppath' 04:11
chromatic I'll try that. Thanks.
allison chromatic: (it's my alias for temporarily adding the perl in /opt/perl to the path) 04:12
chromatic There's no /opt/perl
allison then, no, it doesn't have an alternate perl installed (maybe I wiped the laptop) 04:13
(I've been compiling perl in /opt/perl for over 10 years, so there's no chance it's anywhere else) 04:14
chromatic I suppose I could root the box, fix the CPAN configuration, install Storable, then build Parrot and try to debug the problem... or I could have dinner and watch some TV and call it a good night. 04:15
allison chromatic: the phrasing gives away your inclination :) 04:16
chromatic What, just because I think Steve Jobs and Jonathan Ives should be tried at the Hague for war crimes against people who know POSIX? 04:17
japhb Congrats on Parrot release! 04:19
chromatic allison, I can't fix this laptop without sudo. 04:26
Dinner wins.
dalek r32859 | kjs++ | trunk: 04:27
: [pirc] work on tests. Remove a newline in PASM output. Loads of tests should be written (anyone bored? :-)
: I'm planning to set up tests structured as the grammar rules, so that it will become easier to check whether grammar features have tests (same order of tests as rule order).
: Also, -E and -H options should be tested, which output preprocessed and heredoc-preprocessed PIR code, respectively. That way, the heredoc processor and macro bits of the grammar can be tested.
diff: www.parrotvm.org/svn/parrot/revision?rev=32859
r32860 | chromatic++ | trunk: 04:28
: [distro] Added minor updates from the 0.8.1 release.
diff: www.parrotvm.org/svn/parrot/revision?rev=32860
04:41 johbar_ joined
dalek r32861 | kjs++ | trunk: 04:44
: [pirc] clean up some grammar bits. More work needed.
diff: www.parrotvm.org/svn/parrot/revision?rev=32861
allison chromatic: you should have sudo 04:56
lathos There will be a SQLite extension by the end of the day. 05:03
allison lathos: excellent! 05:04
purl zwooshes
lathos I'm really astonished at how cool the PIR-Perl6 interface Just Works. 05:05
kj when was #ps? Did I miss it again? :-(
kj is still confused by the time zone difference
dalek r32862 | kjs++ | trunk: 05:23
: [pirc] add some comments, fix typo. Minor stuff.
diff: www.parrotvm.org/svn/parrot/revision?rev=32862
r32863 | kjs++ | trunk:
: [squaak] fix a test. kindof. needs more investigation; something wrong with for-loop and step size of 2.
diff: www.parrotvm.org/svn/parrot/revision?rev=32863
r32864 | allison++ | pdd22io: 05:32
: [pdd22io] Implement 'buffer_size' method on FileHandle PMC.
diff: www.parrotvm.org/svn/parrot/revision?rev=32864
05:32 ff-wonko joined 05:34 ff-wonko joined
allison kj: yes, missed it by almost 12 hours (and yes, trying to make meetings around time zones is always a pain) 05:34
kj 12 hours.. :-S Well I've got a good excuse. I was seeing off Jonathan :-)
cotto Is SHIM_INTERP for cases when the interp arg isn't used? 05:35
kj cotto: I believe so.
allison cotto: yes 05:36
cotto ok. Thanks. 05:37
kj allison: why is the INTVAL in: RETURN(INTVAL buffer_size);
is there a description somewhere of the pseudo-C syntax for PMCs ? 05:38
allison that's PCC METHOD syntax 05:39
kj: it always specifies a type in the return
kj: looking for docs...
kj ok, thanks. Is that because it may be autoboxed ? 05:40
so if leaving out, it might be boxed into a Integer PMC?
allison kj: docs/dev/pccmethods.pod 05:42
kj allison: ah thank you.
allison kj: well, it's more because a PCC METHOD can return different types depending on various conditions 05:43
05:44 ff-wonko joined
allison kj: so, for example, the same method might return an INTVAL if one piece of information is found, but an integer array (PMC) if some other piece of information is found 05:44
kj: so, the C code that handles the return needs explicit instructions on how to handle the return value
kj so, suppose a PCC METHOD has an if-statement; in one branch it can return an INTVAL, in another some other value (array, NUMVAL, etc)? 05:45
allison kj: yes 05:47
kj; it can also return multiple values, like RETURN(PMC *status, INTVAL count)
kj: you can think of it as the signature of the return 05:48
kj right, I like the fact that there's a parameter list in AND out.
PCC METHODs are continuation-based as well, no? 05:49
allison kj: yes, also continuation-based
kj neat. So basically you can do ANY thing in a PCC METHOD that you can also do in PIR 05:50
05:54 Psyche^ joined
lathos We are now *close* to a SQLite extension, since I have written a wee generic extension Makefile system thingy. 05:58
dalek r32865 | kjs++ | trunk: 06:04
: [docs] .get_results for exception handling only receives the exception object now; message can be retrieved from that.
diff: www.parrotvm.org/svn/parrot/revision?rev=32865
r32866 | kjs++ | trunk:
: [pirc] udpate TODO.
diff: www.parrotvm.org/svn/parrot/revision?rev=32866
r32867 | kjs++ | trunk: 06:06
: [docs] change .Sub into 'Sub'
diff: www.parrotvm.org/svn/parrot/revision?rev=32867
jimmy www.parrotvm.org/svn/parrot/revision?rev=32857 there is a error 06:21
can anyone fix it?
lathos A 'p' NMI return value does *not* appear to stash the pointer it gets back into the pmc_data member of an UnManagedStruct. 06:22
NCI rather
Weird, the code looks right though. 06:23
Code looks very right. 06:24
06:28 Tene joined
GeJ jimmy: where's the error? 06:32
purl GeJ: Invalid argument
chromatic GeJ, SVN::Web is showing an error. 06:33
jimmy yes. 06:34
lathos gdb shows that NCI *is* stashing the pointer correctly, but it's getting lost somewhere on the way to Perl.
And because it's an unmanagedstruct I can't easily debug it.
jimmy An error occured 06:35
Missing @@ line before @@ -1 +1 @@ at line 5 of string
GeJ Oh, ok. I was looking at a locally mirrored repo through a local trac, and couldn't spot the error.
chromatic lathos, what are you trying to do with the pointer once you have it in an UnManagedStruct? 06:37
lathos Use it as a parameter to another NCI function. But by the time it gets there, the pmc_data has become 0x0.
chromatic Hm. Is the PMC the same between those two calls? 06:38
I mean, it's still the same pointer. 06:39
lathos It's the same PMC. The pmc_data has gone missing though. I'm going to through some watchpoints at it and see what happens.
Ack, no, it's not the same PMC 06:41
I think it's alloc'd me a fresh one. 06:42
chromatic That's good and bad. That's good, because I was worried. That's bad, because it's not working for you. 06:43
lathos st->val->_ptrs->_pmc_val changed. OK, this is progress. 06:48
06:50 uniejo joined
lathos It changed in Parrot_find_pad. 06:50
No, that can't be right.
chromatic I wouldn't expect you're using lexicals. 06:51
lathos Yeah, it's because the variable moved about in memory, I was trying to be too clever. 06:53
OK, let's use a P signature instead and see if that helps. 07:01
07:01 davidfetter joined
chromatic Lower-case p works for me with SDL. 07:04
lathos Yes, but it doesn't work for me.
Are you using SDL from Perl 6?
chromatic No. 07:05
Did you see the OpenGL examples?
They use Perl 6 bindings.
lathos Oh, *bugger*. Entirely my fault. 07:14
Well, partly my fault. It now works nicely from PIR. 07:15
It's still not passing the right value back to Perl 6 though.
chromatic What's Perl 6 getting? 07:16
lathos By the time I pass the database handle back into prepare, it's an empty UnManagedStruct. 07:18
Is there a way from PIR to get at raw pointers for debugging? 07:19
chromatic You can use get_pointer, which should cast the pointer to an INTVAL. 07:20
... though you'd need an opcode which calls VTABLE_get_pointer, now that I think about it. 07:21
lathos UnManagedStructs, you can't really do anything at all with.
Unless I use NCI to pop them into a bit of C and mess with them there.. 07:23
chromatic That's true.
lathos OK, this now tells me exactly what I already knew. Perl6 gets a different value from the one returned by NCI. 07:33
But now I know it's a problem within the NCI argument system. 07:34
07:35 jimmy joined, flw joined
flw hi all 07:36
lathos Hoho. No, it's a problem with *assignment* in Perl 6. 07:38
jimmy hello flw
ꈑę˜Æę•£čŠ±ē‰§äŗŗ
lathos This code works: SQLite::pmc_check(SQLite::open("test.db")); 07:39
This code does not: $db = SQLite::open("test.db");
SQLite::pmc_check($db);
flw čæ™é‡Œä¹ŸåÆä»„č®²äø­ę–‡ēš„吗ļ¼Ÿ
jimmy åŸŗęœ¬äøåÆ仄
07:39 galf joined
cotto chromatic, ping 07:41
07:42 bacek_ joined
lathos The good news, I guess, is that this code works: SQLite::step(SQLite::prepare(SQLite::open("test.db"), "CREATE TABLE foo (bar)")); 07:42
chromatic pong 07:43
cotto I'm removing the seed arg to string_hash. key_hash_STRING also uses a seed. Should I remove the seed arg from there too, or just not use it?
(key_hash_STRING calls string_hash) 07:44
chromatic Don't use it for now; we can remove it later.
cotto I'm not one to complain about having less work. 07:45
lathos OK, let's SVN up, see if that helps.
Nope. Grr. 07:47
jimmy hello cotto.
cotto hi jimmy 07:49
jimmy hello flw
this is cotto
flw is chinese too. he is good at c and perl. 07:50
07:50 Maddingue joined
cotto welcome, flw 07:51
07:51 ff-wonko joined
jimmy i recommend him to join the developing of perl6 07:52
07:52 ff-wonko joined
jimmy flw: ping 07:53
07:53 galf left
flw pong 07:54
07:54 galf joined
flw hello cotto. 07:54
my english is poor.
cotto Don't worry about it. 07:56
Just do your best.
flw :) 07:57
jimmy rakudo: <Welcome Flw>.join(' ').say; 07:58
polyglotbot OUTPUT[Welcome Flwā¤]
jimmy hello purl
purl niihau, jimmy.
jimmy nopaste?
clunker3 pasta.test-smoke.org/ or paste.husk.org/ or nopaste.snit.ch:8001/ or rafb.net/paste or poundperl.pastebin.com/ or paste.scsys.co.uk/
purl i heard nopaste was at nopaste.snit.ch/ (ask TonyC for new channels) or rafb.net/paste or poundperl.pastebin.com/ or paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl or at www.extpaste.com/ or paste.scsys.co.uk (for #catalyst, #dbix-class, #moose and others)
flw rakudo: <Welcome Flw>.join(' ').say; 08:00
polyglotbot OUTPUT[Welcome Flwā¤]
flw hello purl
chromatic lathos, have you used parrot perl6.pbc --target=PIR sqlite.p6 > sqlite.pir 08:01
jimmy what is it?
purl it's it!
lathos Ah, that's a good plan. 08:02
chromatic jimmy, instead of compiling Perl 6 code to PIR and running it, the Perl 6 compiler prints the PIR code and exits.
jimmy you means '<Welcome Flw>.join(' ').say;' ? 08:03
lathos Code looks right, again: 08:04
$P24 = $P22($P23)
$P25 = "infix:="($P21, $P24)
$P26 = $P20($P25)
chromatic Question being what infix:= does
lathos Without the assignment:
$P23 = $P21($P22)
$P24 = $P20($P23)
The wrong thing. :)
chromatic I'm going to say a word and it's going to scare you: typemap. 08:05
lathos But yes, that tells me what I already know: Perl 6 assignment is busted.
jimmy chromatic: hello , you means '<Welcome Flw>.join(' ').say;' ?
chromatic jimmy, I thought you were asking what --target=PIR does. 08:06
Oh yes, a new Coverity run!
Results/KLoc: 0.08. A new record for us.
jimmy chromatic: what is the final compilerļ¼Ÿ pirc or PIR? 08:07
chromatic Right now, IMCC.
We hope that pirc will replace it in the near future.
jimmy and PIR? 08:08
purl i think PIR is a bofh at tufts or www.pir.net or Passive Infra Red or Parrot Intermediate Language (and almost an acronym) or the Price is Right. or www.parrotcode.org/docs/art/pp001-intro.html
jimmy i means languages/PIR.
chromatic As far as I know, languages/PIR was an experiment that no one's working on seriously right now. 08:09
Is that what you had in mind when you were asking about bootstrapping earlier?
jimmy yes
chromatic Once pirc can replace IMCC, languages/PIR has a chance of bootstrapping. 08:10
jimmy and i want PIR can come true. 08:11
chromatic pirc cannot generate PBC, which is what Parrot actually runs.
08:12 Zaba joined
jimmy Is it will in the future? 08:12
err. 08:13
08:13 flw joined
jimmy Is it will generate PBC in the future? 08:13
chromatic Yes.
jimmy and so is PIR, I think. it is stirring. 08:14
08:15 iblechbot joined
chromatic If we can allow pirc to generate PBC, languages/PIR can use the same code. 08:15
jimmy i have a question. where is binary code, and compile src/ codes. 08:16
I try to find it. 08:17
is it libparrot.a ?
s/and/after/ 08:19
chromatic I'm not sure what you're asking. Almost everything in src/ gets compiled into blib/lib/libparrot.so
08:19 Maddingue joined
jimmy got it, thanks chromatic. in windows ,it is libparrot.a 08:20
chromatic I would expect libparrot.dll, unless you're using Cygwin... but yes, it's all about the same thing. 08:21
jimmy I'm just using windows now. 08:22
for working.
dalek r32868 | allison++ | pdd22io: 08:24
: [pdd22io] Add a 'mode' method for checking the read mode of the FileHandle.
diff: www.parrotvm.org/svn/parrot/revision?rev=32868
08:59 Zaba joined 09:10 particle joined 09:20 barney joined 09:51 particle joined
barney ordered the book 'Extending and Embedding PHP' 09:57
jimmy where is the book> 09:58
10:15 tomyan joined 10:24 ff-wonko joined
barney jimmy: I searched on Amazon 10:26
jimmy good book. though it is english. 10:31
10:31 Ademan joined
barney Supporting extensions is high on my TODO list. 10:36
But I fear that many extensions use internal knowledge of the Zend Engine 10:37
jimmy yup, it is extremely important.
most does not i think 10:38
i.e. APC
cotto barney, is your eventual hope is that Pipp will be able to use native PHP extension libraries? 10:44
s/hope is/hope/
10:47 tomyan joined 10:48 tomyan joined 10:49 tomyan joined
barney cotto: Yes. The build system might be hairy. The extension API should be doable with NCI. 11:01
cotto should be fun 11:02
that'll really give NCI a good workout 11:04
barney and reduce the number of functions that need to be implemented in PIR 11:09
lu_zero hi 11:13
purl privet, lu_zero.
barney cotto: Do you mind if I reformat the Pipp Grammar to the code layout used by Rakudo? 11:20
cotto go ahead. It could use some cleanup. 11:21
barney k 11:24
11:29 Alias joined
Alias Ping? 11:29
szbalint pong
Alias What version of Perl are most people using to build parrot 11:31
I want to have a quick shot at building it again with Perl::Dist
szbalint I'd guess 5.8.8 11:32
Alias And is Rakudo included with the parrot build, right? 11:33
barney a lot of 5.10 in smolder.plusthree.com/app/public_pr..._reports/8
With Parrot 1.0, Rakudo should become a separate dist. 11:34
Alias Sure, but right now it's inside it? 11:35
The plan is to see if I can quickly reuse the same toolkit I use for Vanilla Perl to make a Vanilla Rakudo
Windows installer, etc etc
barney The source is in the Parrot repos and right now you build it with: cd languages/perl6 && make 11:36
Alias I promised to look into it at YAPC::NA, but the toolkit previously hasn't supported abuse of this magnitude yet
OK, let me play with it and see
Out of curiosity, are any of the other languages usable? 11:37
Should I try to just target Rakudo? Or try for build Parrot plus several languages
barney depends on your tuits 11:38
Most langs are working, but are incomplete
Pipp, Cardinal, Pynie, Partcl are actively worked on 11:39
Alias Cadinal is Ruby... so screw them :) 11:42
What's Pipp 11:43
purl Pipp is Parrot's PHP
Alias ?
ah
Even scarier :)
Does lolcode work? Yes?
Alias ponders Strawberry Lolcode :)
szbalint that'd be great
CAN I HAS CHEEZBURGER?
barney hq9plus works too 11:45
Alias wassat? 11:46
11:47 jimmy joined
barney www.esolangs.org/wiki/HQ9_Plus 11:49
dalek r32869 | bernhard++ | trunk: 11:53
: [hq9plus] HQ9+ got deleted from Wikipedia, so get rid of the link
diff: www.parrotvm.org/svn/parrot/revision?rev=32869
r32870 | bernhard++ | trunk: 11:58
: [Pipp] Set variables in functions to lexical scope.
szbalint wikinazis--
dalek diff: www.parrotvm.org/svn/parrot/revision?rev=32870
cotto night 12:05
jimmy night 12:12
12:20 mberends joined
dalek r32871 | bernhard++ | trunk: 12:30
: [Pipp] Normalize heredoc delimiters.
: Minor cleanup in test scripts.
diff: www.parrotvm.org/svn/parrot/revision?rev=32871
12:32 wubo joined 12:47 flw joined 13:11 Lorn joined 13:16 apeiron joined
Alias Does parrot support dmake? 13:22
13:22 ChrisDavaz joined
Coke not intentionally. 13:26
gmake & nmake I think are the big two. 13:27
but if it's an easy update, glad to help out there.
13:28 masak joined
pmichaud should we s/parrotbug@perl.org/parrot-tickets@parrot.org/ in the repository? 13:44
er, @lists.parrot.org
masak 'ticket' is more neutral, which is both good or bad, IMHO 13:45
s/or/and/
Coke (vanilla) I often use strawberry to build parrot and partcl, no problems. 13:59
Alias Coke: Ya, so I'm told
dmake?
purl dmake is a borland thing or at search.cpan.org/dist/dmake/ or <reply>it's just like make, except it's totally gay
Coke allison: I would be curious if a vanilla partcl worked.
Alias or the gnu-make
Coke mingw32-make
Alias ok 14:00
Coke (if you run the configure.pl with strawberry perl, that's the make it recommends.)
Alias ah
I need to automate it, Perl::Dist doesn't allow human input
dalek r32872 | bernhard++ | trunk: 14:03
: [Pipp] Reformat the grammar.
diff: www.parrotvm.org/svn/parrot/revision?rev=32872
Alias Is there a flag to Configure.pl that says "just do what you think is best" 14:06
14:07 flw joined
dalek r32873 | pmichaud++ | trunk: 14:10
: [rakudo]: Update objectref semantics for non-Rakudo objects.
diff: www.parrotvm.org/svn/parrot/revision?rev=32873
Coke I think that's "no flag" 14:11
do you want to just end up with a built parrot?
perl Configure.pl && mingw32-make ?
14:12 gryphon joined
Alias Coke: OK 14:13
dalek r32874 | bernhard++ | trunk: 14:23
: [Pipp] require whitespace after 'echo', in <echo_statement>.
diff: www.parrotvm.org/svn/parrot/revision?rev=32874
14:24 tomyan joined 14:28 iblechbot joined
dalek r32875 | pmichaud++ | trunk: 14:29
: [rakudo]: spectest-progress.csv update: 220 files, 4578 passing, 41 failing
: Failure summary:
: S03-operators/increment.rakudo aborted 41 test(s)
diff: www.parrotvm.org/svn/parrot/revision?rev=32875 14:30
14:31 tomyan left 14:32 tomyan joined 14:34 particle1 joined
dalek r32876 | bernhard++ | trunk: 14:39
: [Pipp] Add tests for user defined functions with multiple args
diff: www.parrotvm.org/svn/parrot/revision?rev=32876
14:47 ff-wonko joined
lathos pmichaud: Still no joy: 14:49
PMC 0x2641740 data pointer 0x2d1a860
PMC 0x26417e8 data pointer 0x0
Alias ok, first attempt at Vanilla Parrot underway 14:54
It will probably die horribly...
dalek r32877 | bernhard++ | trunk: 14:55
: [Pipp] A less stupid implementation of testing functions.
diff: www.parrotvm.org/svn/parrot/revision?rev=32877
PerlJam Alias++
Alias Meh, we'll see
PerlJam Trying is just as important as succeeding sometimes. :) 14:56
Alias If anyone wants to play with it, svn.ali.as/cpan/trunk/Perl-Dist-Parrot
The dmake test should attempt to build a distribution
Dependencies are of course pretty heavy, and must be run on Windows
pmichaud lathos: can you see if the second PMC is an ObjectRef PMC? 15:10
lathos How do I do that? 15:11
Oh, I'll try printing it.
pmichaud that actually won't work -- it'll delegate to its reference. :-)
lathos Ah.
pmichaud from PIR, try an isa $P0, 'ObjectRef'
from C, look at the vtable->whoami structure
15:12 rdice joined
PerlJam lathos: btw, mjd mentioned the other day that you're engaged. Congrats. 15:15
lathos Yes. Thanks.
I've been engaged for about six months now, but we had the formal ceremony on Saturday. 15:16
pmichaud in the meantime, if you just want to get "something that works", you could try doing $db := SQLite::open("test.db") instead of assignment.
lathos Am I an objectref? no 15:17
The := form does work, yes. Shame though. 15:19
pmichaud well, we'll undoubtedly get assignment to work at some point also -- you're the first person to really try to use a non-Rakudo object. :-) 15:21
I'll be happy to play with it some more -- but today is really devoted to getting lexicals to work properly. 15:22
lathos Sure.
pmichaud also if I can get the SQLite code you're using ... :-)
lathos := will work for me for now.
Should I check it in? I've made a ext/SQLite3 in my local copy. 15:23
pmichaud sounds good to me
lathos It's also significant because I've made a way of doing Makefile.PL's for Parrot extensions.
pmichaud yes, that's definitely significant
ext/ sounds like a good place to put it for now
PerlJam lathos++ yes, definitely check it in. Then I won't have to continue the work I was doing; I can just use yours :) 15:24
dalek r32878 | simon++ | trunk: 15:32
: Beginnings of a SQLite extension. Also contains a way to generate Makefiles for Parrot extensions. My C, as always, needs work, and my Make-fu even more.
diff: www.parrotvm.org/svn/parrot/revision?rev=32878
15:35 tomyan joined
dalek r32879 | particle++ | api2export: 15:36
: [api2export] merge in trunk changes to branch
diff: www.parrotvm.org/svn/parrot/revision?rev=32879
15:40 bacek joined
masak how will fork() work in Rakudo? will it create a new Parrot instance? 15:51
pmichaud I have no idea. 15:52
masak :)
pmichaud that's really going to depend on what Parrot does.
masak yeah, I guess.
PerlJam won't rakudo just call parrot's fork op ? :-)
lathos Depends what Perl 6 fork semantics require, too. 15:53
15:59 gryphon_ joined 16:00 clunker9_ joined, jimmy joined 16:05 jhorwitz joined
Infinoid is parrotbug switched over to trac yet? 16:06
particle1 no 16:07
particle neither the utility, nor the email addr
don't know which you mean
Infinoid well, if we can switch over the email address, I think they both will work 16:11
dalek r32880 | bernhard++ | trunk: 16:13
: /home/bernhard/devel/Parrot/trunk/svn-commit.2.tmp
diff: www.parrotvm.org/svn/parrot/revision?rev=32880
Infinoid starts moving his tickets over 16:14
16:14 bacek joined 16:20 ruoso joined
pmichaud afk, lunch 16:25
dalek r32881 | pmichaud++ | lex3:
: Refactor context tracing and do "make headerizer" to bring stuff up-to-date.
diff: www.parrotvm.org/svn/parrot/revision?rev=32881
r32882 | bernhard++ | trunk:
: [codingstd] Fix svn properties and trailing spaces.
diff: www.parrotvm.org/svn/parrot/revision?rev=32882
16:27 Limbic_Region joined
jonathan arrives home. 16:30
lex3?!
particle welcome home, jonathan
jonathan It's...colder than San Francisco. 16:31
particle my wife was in hawaii on saturday, with highs in the 80's. she's now in dc, with highs in the 30's. 16:32
masak yes, welcome home, jonathan
dalek r32883 | particle++ | api2export: 16:35
: remove invalid/old svn:mergeinfo data
diff: www.parrotvm.org/svn/parrot/revision?rev=32883
16:38 tomyan left
dalek r32884 | particle++ | api2export: 16:39
: [api2export] merge in trunk changes to branch
diff: www.parrotvm.org/svn/parrot/revision?rev=32884
jonathan particle: *those* 30s are cold. :-)
particle yes, indeed :)
jonathan It's not much better here. 16:40
dalek r32885 | particle++ | trunk:
: remove invalid/old svn:mergeinfo data
diff: www.parrotvm.org/svn/parrot/revision?rev=32885
cognominal particle, make that in centigrade, please :) 16:46
particle convert 80 f to c
purl 80 Fahrenheit is 26.6667 Celsius.
cognominal I mean in Celcius
particle convert 30 f to c
purl 30 Fahrenheit is -1.11111 Celsius.
cognominal reads about the diff between centigrade and Celcius 16:48
particle jonathan: your name is notably absent here: trac.parrot.org/parrot/wiki/ParrotRoadmap ;) 16:53
jonathan particle: Awesome! Does that mean I've got no responsibility? ;-) 16:59
confound heh
16:59 hercynium joined 17:00 register joined
cognominal if jonathan was more often there, there would be no need of roadmap 17:01
jonathan particle: Do I have an account for Trac? 17:03
particle iunno, but i'm sure you can create one
think it's trac.parrot.org/parrot/register 17:04
jonathan particle: OK, I didn't ask what I meant to.
What I meant to ask was, does my other parrot.org account work here?
Or are they separate?
particle yes, no
jonathan Ah.
jonathan wonders why he couldn't sign on with those details then 17:05
particle hrmm, maybe they are separate
you mean is the drupal id the same as trac id? i don't know that.
they just might be different
jonathan Yes.
OK.
I'll register
done 17:06
particle average price for a gallon of gas in nj is $1.97 now. wow. 17:07
jonathan kjs and I were *amazed* how cheap it was to re-fill the tank after we took a rental car for a roadtrip on Monday.
particle :) 17:08
jonathan US certainly has cheap gas.
cognominal particle, be happy not to pay taxes like in Europe
jonathan I was looking at the options at the rental place thinking..."whoa, this mpg is shite compared to what you'd find in Europe"
But with gas prices like that, no wonder nobody cares...
particle ayep 17:09
jonathan wonders who is responsible for the user documentation
allison jonathan: we're all responsible for the user documentation 17:10
jonathan allison: OK, great. 17:11
I figured as much.
dalek r32886 | tewk++ | trunk:
: [NCI] added v_fff to config/gen/call_list/misc.in, fixex RT #60662
diff: www.parrotvm.org/svn/parrot/revision?rev=32886
jonathan I wondered if anyone was going to co-ordinate the effort.
And try to make sure it really happens.
Or make sure two people don't go and write user docs for the same thing. 17:12
17:12 hercynium joined
jonathan particle: But my name in a couple of places. 17:12
*put
particle thanks, i see that :)
particle is almost done with testing _API->_EXPORT in trunk 17:13
jonathan particle: Is it just a straight renaming?
particle yep, but skipping some files
jonathan ok
particle make codetest now, then commit 17:14
feh, the sqlite3 files are failing 17:17
dalek r32887 | Whiteknight++ | trunk: 17:18
: [Book] Update about the first half of chapter 4, removing cruft and old information, updating examples, improving flow.
diff: www.parrotvm.org/svn/parrot/revision?rev=32887
17:20 apeiron_ joined
dalek r32888 | particle++ | trunk: 17:22
: #60624: [DEPRECATED] PARROT_API changes to PARROT_EXPORT
diff: www.parrotvm.org/svn/parrot/revision?rev=32888
particle allison: you have a % status for i/o refactor work?
85%? 17:23
purl hmmm... 85% is M::B got lucky the last year or so
cognominal forget 85% 17:30
purl cognominal, I didn't have anything matching 85%
pmichaud lex3 is my branch where I re-synced with trunk. 17:35
jonathan pmichaud: Ah, OK. HOw's the lex work going? 17:36
I guess I could read backscroll and find out, but I'm lazy..
pmichaud jonathan: just got really restarted on it this morning. 17:37
spent the last couple of days cleaning up things for release.
I'm reworking the context tracing a bit so I can better see what's going on. 17:38
particle pmichaud: you using svn 1.5 yet?
jonathan OK
Nasty.
pmichaud particle: still 1.4.6 here.
re-sync with trunk is really easy for me.
(1) create new branch (lex3) from trunk 17:39
(2) svn merge old branch to new branch
particle you should really upgrade, then resync is: svn merge svn.perl.org/parrot/trunk
PerlJam idly notes that svn 1.5.3 comes with Intrepid
pmichaud PerlJam: sure, but Intrepid has a bunch of other broken stuff for me.
particle and merge to trunk is: svn merge --reintegrate svn.perl.org/parrot/branches/branchname
pmichaud eh, mine seems just about as easy. :-)
particle has lex2 been deleted? 17:40
pmichaud not yet, I'm keeping it around a bit just in case I abandon the lex3 branch.
PerlJam particle: and if he were using git, all the merges would just be "git merge ..." :)
dalek r32889 | particle++ | api2export: 17:42
: remove api2export branch, changes merged with trunk
diff: www.parrotvm.org/svn/parrot/revision?rev=32889
17:44 gryphon joined
pmichaud hmmm, perhaps I should re-re-sync with trunk, since PARROT_API is now changed. 17:44
dalek r32890 | pmichaud++ | lex3: 17:52
: [gc]: restore Parrot_pop_context to how it is in trunk for now.
diff: www.parrotvm.org/svn/parrot/revision?rev=32890
r32891 | pmichaud++ | lex4:
: Create new branch to re-re-sync lex with trunk.
diff: www.parrotvm.org/svn/parrot/revision?rev=32891
jonathan places a bet on getting to double figures in numbers of lex branches
pmichaud I hope it doesn't live _that_ long. 17:53
jonathan Me too!
moritz proposes counting in binary to support jonathan's bet
NotFound ETOOMANYLEX
pmichaud I was just thinking that I could get up to 36 just using a..z :-)
I could also keep track of them modulo 10 :-) 17:54
Coke I miss the ability to rebase.
(only thing I liked about clearcase)
PerlJam Coke: you too can use git :)
Tene rebases parrot patches regularly with git-svn 17:55
moritz pmichaud: should I unfudge the bigint tests that pass for me? 17:58
pmichaud moritz: I'll fudge them a bit later today, before the next spectest-progress update.
moritz pmichaud: ok
pmichaud I'm going to go ahead and refactor increment.t to move the overflow tests into their own test. 17:59
I'm also thinking we might want to update fudge so that we could have
17:59 chromatic joined
pmichaud #?bigint skip "Not available w/o bigint" 17:59
and then be able to tell fudge to honor directives having either "#?rakudo" or "#?bigint"
it would give us the ability to run tests on a platform-specific basis, instead of just by the specific interpreter being used. 18:00
moritz I vote against it - these fudges would be very hard to maintain
rather make bigint availabilty a policy for now 18:01
or maybe make the information available through %?VM
pmichaud in which case we always have to fudge to the lowest common denominator
(or have the tests depend on %?VM, which amounts to the same difficulty of maintenance, yes?)
moritz that's hard if your machine is not configured to be on the lowest common denominator 18:02
jonathan wonders if we can bundle some big int support with Parrot.
pmichaud jonathan: eventually I suspect that's what we'll do.
moritz pmichaud: not exactly, because it's orthogonal to the implementation
jonathan Rather than relying on people to have a certain library, which may or may not but likely will be a pain for Win32 folks. ;-)
pmichaud jonathan: if Parrot doesn't offer native bigint support in the Integer PMC then we will likely implement Perl6Int
(that does.)
moritz: I don't see how maintaining #?bigint would be any more/less difficult than what you propose. 18:03
it moves the decision of "what tests to include" out of the test file and into the command line options for fudge 18:04
if we do it based on %?VM, then the test file has to ask the VM "do you support bigints" and every VM has to know how to answer yes/no.
moritz pmichaud: I'm not yet sure how you think that #?bigint and #?rakudo would interact
pmichaud if I put #?bigint into a file, then fudging for "bigint" would skip those tests, yes? 18:05
particle pmichaud: how about #?parrot<bigint> skip ...
moritz pmichaud: yes, but that would disable all #?rakudo fudges
pmichaud: (unless you process it twice, that is)
pmichaud so, I'm proposing that fudge be able to handle "rakudo|bigint"
particle er, #?rakudo<bigint> i mean 18:06
pmichaud which then processes "#?rakudo" and "#?bigint" directives -- i.e., those matching the pattern "rakduo|bigint"
dalek r32892 | kjs++ | trunk:
: [docs] move docs/art to docs/user/pir/
: + update MANIFEST accordingly.
diff: www.parrotvm.org/svn/parrot/revision?rev=32892
particle actually, it's nobigint you're looking for 18:07
pmichaud okay, nobigint
particle and this probably belongs on #perl6
pmichaud or even 32bitint
moritz 's gotta go now, see you later
particle since there are other implementers there
pmichaud particle: yes, but I'm not planning to do this immediately -- it's just for discussion and reflection at the moment.
particle ok 18:08
pmichaud in some sense I may just move the test out of the regression suite for now.
pmichaud files his first trac ticket.
particle we could kick it around on the phone today, to get folks thinking
dalek r32893 | kjs++ | trunk:
: [docs] add svn:mime-type property to docs/user/pir.
diff: www.parrotvm.org/svn/parrot/revision?rev=32893
pmichaud perhaps. somehow that's not a conversation I'm eager to have.
particle can we get an svn backing store for the parrot.org wiki? 18:10
it'd be nice to have something to ack through to make wiki updates easier (e.g. PARROT_EXPORT change) 18:11
dalek r32894 | allison++ | pdd22io: 18:14
: [pdd22io] Sanitizing names and cleaning up temporary files in FileHandle tests.
diff: www.parrotvm.org/svn/parrot/revision?rev=32894
r32895 | allison++ | pdd22io: 18:17
: [pdd22io] Missed a 'break' statement in the read-mode parsing switch, meaning
: all 'write' filehandles were also 'append'.
diff: www.parrotvm.org/svn/parrot/revision?rev=32895
r32896 | kjs++ | trunk:
: [docs] remove docs/art directory.
diff: www.parrotvm.org/svn/parrot/revision?rev=32896
mj41 hi, still too many branches I think? tt.ro.vutbr.cz/report/pr-Parrot/ 18:18
allison particle: just about ready to merge io 18:20
particle: can you retest windows?
particle i just did svn up, i'll run tests
give it ~15m, i'll get you results 18:21
dalek r32897 | allison++ | pdd22io: 18:26
: [pdd22io] Enable buffering at the same time as setting buffer size in FileHandle.
diff: www.parrotvm.org/svn/parrot/revision?rev=32897
r32898 | pmichaud++ | lex4:
: re-sync lex branch with trunk (PARROT_API -> PARROT_EXPORT)
diff: www.parrotvm.org/svn/parrot/revision?rev=32898
Coke boggles at patrick.
Coke knows branches are cheap, but damn. 18:27
dalek r32899 | allison++ | pdd22io:
: [pdd22io] Clarifiying semantics of setting 'buffer_size', and fixing test to match.
diff: www.parrotvm.org/svn/parrot/revision?rev=32899
particle allison: all tests pass in pdd22io with msvc for r32899 18:44
allison excellent
18:52 gmansi joined
Coke kicks off the 3.5 hour spec test run. 18:54
zzzzzzzzzzzzzzzzzz
dalek r32900 | allison++ | pdd22io: 18:58
: [pdd22io] Some PDD cleanup based on I/O implementation.
diff: www.parrotvm.org/svn/parrot/revision?rev=32900
Coke allison: you around? 18:59
19:02 mberends joined 19:07 Theory joined 19:08 Lorn_ joined 19:13 sjansen joined
allison Coke: yup 19:14
Coke design q, sorry if it's a rehash: 19:18
in partcl, I load a parrot library .pbc; it puts things in the top level parrot namespace. my HLL top level namespace is parrot's top level 'Tcl' namespace; how do I get at the newly loaded sub? 19:19
past conversations included creating a 'parrot' namespace in my tcl HLL that pointed back to the parrot top level.
so that in tcl ::parrot::_config would correspond to parrot's top level '_config' 19:20
(I tried grabbing the root_namspace, and shoving it into tcl's HLL namespace 'parrot' slot during tcl startup, but this causes any subsequent load_bytecode to hang, presumably because it's trying to walk the loop. 19:21
(design q because while getting this working now is nice, knowing the plan is nicer) 19:23
I vaguely recall more recent conversations that involved me having to manually wrap the parrot library with my own HLL library.
Tene Coke: you could get the root namespace and walk it, installing all entries (except tcl) into a new namespace you create 19:28
That's a pretty simple iterator, but it won't reflect future modifications to the root namespace
You could explicitly rewrite access to ::parrot in your compiler 19:29
Coke all of those involve a lot more work (or are static), compared to the possibility of installing an alias.
the last is the best of the 3, and will probably be my fallback. 19:30
Tene Or you could invent new syntax for accessing other namespace roots
That's Perl's solution. use Foo :from<other>;
Coke I'm not trying to rewrite tcl; I'd rather (ab)use existing syntax. 19:31
Tene nods. 19:32
Coke I wonder if the easiest way to implement #3 is to subclass the Namespace obj. 19:33
(and put that one subclassed object into Tcl's 'parrot' slot.
(rather than just copying in parrot's top level NS obj.)
Tene I'm told that the namespace pdd needs significant review and revision. 19:34
Coke: wouldn't that potentially have the same looping problem that copying does? 19:35
Coke not if my subclass of NS was smart enough to not report on itself. 19:37
so ask it for ::parrot::_config, get _config from parrot. ask it for ::parrot::tcl and get "no" 19:38
dalek r32901 | pmichaud++ | trunk:
: [rakudo]: Update test_summary.pl to summarize each synopsis section.
: * Also grab plan count from test output if it's available.
diff: www.parrotvm.org/svn/parrot/revision?rev=32901
particle pmichaud++ can you post those results when you get them? 19:42
nopaste "pmichaud" at 72.181.176.220 pasted "spectest regression, by synopsis" (14 lines) at nopaste.snit.ch/14640
Tene Coke: That's not asking it for itself, though, that's asking it for its parent. 19:44
Which I guess could work too.
Coke pmichaud: int? 19:52
purl int is item => .post_circumfix:<[ ]>, Str => Item .post_circumfix:<{ }>
Tene purl: forget int
purl Tene: I forgot int
Coke purl: initiate self destruct 19:53
purl Coke: i'm not following you...
pmichaud Coke: those are tests that are in t/spec/integrated -- they aren't really categorized by synopsis (yet?)
Coke partcl? 19:54
purl partcl is tcl on parrot or code.google.com/p/partcl
Coke partcl is also slow, help make it go faster.
purl okay, Coke.
Tene Coke: just throw more hardware at it. 19:55
cotto Try CERN. I heard they have a partcl accelerator. 19:56
Infinoid t/spec/S16-io/basic-open.rakudo has been running for half an hour, I think it's stuck in a loop. 19:58
any debugging info I should get before I kill it?
(looks like the rest of spectest_regression went through normally.) 19:59
pmichaud phone. 20:00
TimToady phone
drat
register mmm 20:33
does anyone know a way to infer declarations automatically from c code?
NotFound register: a C compiler 20:34
register mmm
a c compiler
compiles definitions
it does not infer anything
20:34 masak joined
register i am looking for a tool to desume declarations 20:34
from a bunch of c code definition
jonathan register: Like .c => .h? 20:35
register exactly jonathan
i am working on ticket 38929
but there is a hell to do manually
jonathan I understand what you mean then, but don't know what of anything to do that.
register :D
jonathan Well
Don't we have a PGE grammar written for C? ;-)
chromatic Headerizer does that, per our coding standards.
jonathan Oh, yes. 20:36
:-)
register i read that headerizer works on object files
i need to strip declarations from code
not from objects
chromatic But it doesn't.
NotFound Object files? You need a wizard in the literal meaning of the word.
chromatic The API is misleading.
headerizer infers the name of a .c file, given the name of an .o file. 20:37
jonathan by s/\\.o/.c/ ?!
chromatic Don't ask me why.
jonathan :-)
chromatic: Welcome to Parrot. ;-)
register why? 20:38
:D
pretty strange
tewk_ if PGE were faster, ncigen would be the tool to use.
chromatic It may have been to simplify Makefile rules, but that's my best guess.
register ah ok
tewk_ it has a full c parser
jonathan If speed isn't critical, that would be a way to go. 20:39
NotFound tewk_: for some values of "full"
tewk_ NotFound: provide a test case that doesn't parse and I'll fix it.
I've been adding GNU C extensions.
NotFound tewk_: parsing a header does not solve all problems. You can have a typedef'ed or declared parameter that maps to an int in your platform, and to a long in other platform. 20:40
tewk_ I started parsing parrots own header files, but PGE couldn't handle it :( 20:41
NotFound: ncigen parses fully pre processed source c files.
register i try to use the headrizer
i will modify it to read directly a source file 20:42
and will try to use in on the headers
NotFound tewk_: yes, and the signatures generated on that base will work in your platform, but may not work in others.
register in the jit
that contain a lot o c code
hope it works
else i will continue doing it manually
moritz re 20:45
jonathan fw
moritz jonathan: do you have a rakudo day this week? or next week? 20:47
jonathan moritz: I need to catch up on my non-Rakudo/Parrot clients tomorrow. 20:51
I may or may not fit some hacking in on Friday.
Next week, will certainly have one.
moritz ok
jonathan Do you have specific wants?
moritz not really
jonathan OK. :-)
moritz clean up the rakudo side of pmichaud's lex merge perhaps ;) 20:52
(assuming we see it until then)
jonathan Yeah, hopefully pmichaud finds the remaining issues without too much pain.
GeJ Good morning everyone
jonathan morning GeJ
masak GeJ: morn' 20:53
register who wrote the jit?
masak jonathan: are list slices difficult to implement? or are you saving that for the S09 sprint? 20:54
jonathan masak: Not so bad, but I think they're less useful without list context. 20:58
pmichaud plans, IIRC, to do that soon (before he starts on protoregexen)
masak oh -- we don't have that yet? :)
jonathan ;) 20:59
Not really.
masak I thought `my @a = 1,2,3` was list context.
jonathan Oh oh!
I meant list *assignment*.
masak ah.
jonathan Sorry, jetlagged.
masak np.
jonathan That's why I'm currently labelling photos rather than writing code.
masak good plan.
tewk_ register: I probably modified it last, but I didn't write it
register tewk may i chat with you in private? 21:00
Coke (holy crud tcl spec test still running. guess I shouldn't have nice'd it!) 21:12
masak No more Mr Nice guy.
s/g/G/ 21:13
nopaste "coke" at 72.228.52.192 pasted "why doesn't this print BOO?" (15 lines) at nopaste.snit.ch/14642 21:19
Coke anyone see why that nopaste doesn't work? 21:21
cotto Coke, I think the key is being passed as a Key PMC. 21:25
bacek Coke: you probably need 'NameSpace' declared.
cotto It works if you rename the function to get_string_keyed 21:26
particle rats, you beat me to it. cotto++
cotto++ # also, you're doing excellent work mentoring new contributors 21:27
cotto thanks.
It's nice to help someone else on occasion. 21:28
Coke bacek ; it already is decalred. it's the namespace PMC.
no?
cotto: hurm. 21:29
particle cotto: keep it up!
Coke cotto: isn't that what get_string_keyed_str means?
ah.
get_string_keyed :vtable isn't invoked either. 21:30
if someone could get that little sample working, that'd help out tcl a bit.
particle coke it works with get_string_keyed 21:31
c:\\Users\\particle\\dev\\parrot\\trunk>parrot ns.pir
BOO
.sub get_string_keyed :vtable
nopaste "cotto" at 96.26.202.243 pasted "working version of Coke's script" (15 lines) at nopaste.snit.ch/14643
bacek .sub '' :vtable('get_string_keyed') 21:32
Coke: this one works
tewk_ how do I create a new wiki page on Trac?
cotto It probably involves getting around the 500. 21:33
moritz you don't until someobdy fixes the 500
Coke we have 3 wikis now. yay! 21:34
(sarcasm)
cotto++
21:34 PacoLinux_ joined 21:35 PacoLinux_ joined
particle tewk: create a WikiLink item on the parent page, save the page, then click the link 21:35
tewk_ Ok I think I figured it out
thanks
register is looking for basic CAGE tasks involving C that are on the ROADMAP to 1.0 21:37
ideas?
particle yes
DEPRECATED.pod has a number of c functions that need to be removed
Coke cotto: if that syntax doesn't invoke the _str variant... why do we -have- the string variant?
particle: the ones that are left can't easily be removed. 21:38
(at least the MMD ones.)
I already removed all the unused ones.
particle coke: the namespace pmc is a mess 21:39
chromatic tewk_, it'd be nice to fix up some of the compilation warnings in C -- especially the conversion and width ones. 21:40
particle did allison run out of tuits, i thought pdd22 was going to merge?
tr/,?/?./ 21:41
allison particle: testing on linux
particle gotcha
allison and every other platform I can find, the more the merrier
particle who's working on calling conventions? you and whiteknight?
allison particle: yes 21:42
particle: he's doing the merge on that one
jonathan Will start doing bytecode stuff next week, esp annotations. 21:43
Will try and keep us ahead of schedule on that one.
register I am not really looking for a cage task
but simply for something
that is well defined and closed
something where one, at the begigging, could just use c and linux/windows programming skils
a cage task could be a solutions 21:44
but anything else is fine
chromatic register, how do you feel about cleaning up some compilation warnings?
tewk_ chromatic: trac.parrot.org/parrot/wiki/LlvmJitExperiment is a summary of my llvm vision with register, feel free to add to and improve.
Tene register: a nice task for someone new to parrot is cleaning up deprecated items or compilation warnings.
register that should be pretti easy
pretty 21:45
chromatic It's not very sexy work, but it's useful, it produces cleaner code, and it's a nice way to start to get familiar with the code.
register sorry for my english
yeah
Tene There are plenty of c-level items that need tobe removed, iirc
register that was also my purpose
getting familiar with the code
before
moritz register: many of us are not native speakers, no need to apologize
register passing to something more serious
Limbic_Region moritz - you are not a native speaker - is your primary means of communication ESP ? 21:46
Limbic_Region ducks
Tene register: another option is to look through open tickets for something you think you can resolve.
Coke anyone looking for more tests to run in about-to-be-merged-branches, code.google.com/p/partcl/wiki/PartclSource
chromatic I spend a fair amount of time adding white space, aligning = signs, moving pointers to the variables and not the types, and just making the code slightly easier to read.
moritz Limbic_Region: what's "ESP"?
Coke ESP?
purl i guess ESP is the encapsulation/encryption protocol used by IPSec, IP protocol 50 or okra or not okra or Encapsulating Security Payload or Extra-Sensory Perception or an overloaded acronym or Extra-Special Pants or Eternal Spinning Pinwheel or Egregiously Sanctimonious Perl
Limbic_Region extra sensory perception 21:47
but explaining kinda deflates the joke
chromatic - there isn't a tidy utility that does that? 21:48
chromatic Is there? I didn't know of one.
Limbic_Region I just assumed - something like perl tidy for C
chromatic I'm not sure one exists that's as picky as I am, but that would be a good start.
particle perltidy aligns on = if you want it to
it even kinda works on c code
Limbic_Region given that C is easier to parse - I would just assumed
anyway
particle lint will do that won't it? 21:49
dalek r32902 | allison++ | pdd22io:
: [pdd22io] Generated headers for STDIO utility functions.
diff: www.parrotvm.org/svn/parrot/revision?rev=32902
particle no, not lint... what's it called? hrmm
chromatic indent? 21:50
purl indent to intend!
cotto Coke, get_foo_keyed is called for all keyed lookups. In some PMCs (*array) it dispatches to get_foo_keyed_int, and in some (Hash) get_foo_keyed does all the work.
Namespace PMCs extend Hash, so Namespaces act the same way in this case. 21:51
tewk_ chromatic: I thought I saw you use a vim macro or something to aligh = assignments, care to share? 21:54
chromatic There's a Vim package, I think called Align. Let me see. 21:56
Yes, it's Align by Charles E. Campbell, Jr. 21:57
Coke cotto: if it's all internal, what's the point of having a vtable for it?
chromatic vim.sourceforge.net/scripts/script....ipt_id=294
tewk_ chromatic++
cotto Coke: I can just tell you what happens, not why. 21:59
Coke ew. when I do a get_hll_namespace, it converts the HLL (tcl) into an hll_id and then does a keyed int lookup
Tene nice 22:01
cotto chromatic++ 22:03
chromatic Eh?
purl Speak up, sonny!
chromatic For sending an email about exploratory testing? 22:04
Coke hurm. how does one dynamically create a namespace? 22:06
(that is, not with the .namespace directive)
cotto for that Align macro 22:07
chromatic Oh good, because if it were for the other thing, I'm changing my SSH keys immediately.
Coke tries .namespace ['foo'], and then a .sub that does get_namespace, and then tries to assign to result of that. boom. 22:08
nopaste "coke" at 72.228.52.192 pasted "slightly more realistic attempt at proxying a namespace (still fails)" (68 lines) at nopaste.snit.ch/14644 22:13
Coke there you go, cotto, have fun. =-) 22:15
cotto ETOOMUCHFUN 22:16
Infinoid ENOSUCHTHING 22:17
Coke -> 22:18
cotto can someone confirm that make cover is broken?
solarion EBADE
(/usr/include/asm-generic/errno.h) 22:19
22:20 davidfetter joined
Infinoid solarion: you're welcome. :) 22:21
22:21 Whiteknight joined
solarion Infinoid: :) 22:22
Infinoid cotto: I'm seeing quite a few errors from that 22:24
22:25 alvar joined
cotto Infinoid++ #thanks for confirming 22:25
dalek r32903 | allison++ | pdd22io: 22:34
: [pdd222io] Bringing the pdd22io branch up-to-date with trunk r32902.
diff: www.parrotvm.org/svn/parrot/revision?rev=32903
chromatic make cover is running for me, but it fails some tests. 22:36
PerlJam I get gcov2perl not found. Where do I get that? 22:37
Infinoid I get various test passes and failures interspersed with src/exception.c build failures.
dalek r32904 | allison++ | pdd22io:
: [pdd22io] Adding new files to the manifest.
diff: www.parrotvm.org/svn/parrot/revision?rev=32904
chromatic You might have to install a gcov package. 22:39
Whiteknight allison, you preparing to merge pdd22io?
dalek r32905 | moritz++ | trunk:
: [rakudo] add test file for RT #58818 to spectest.data
diff: www.parrotvm.org/svn/parrot/revision?rev=32905
Patterner PerlJam: Devel::Cover 22:40
purl Devel::Cover is cool or secretly a simple profiling tool it seems :)
alvar hi!
PerlJam oh, duh!
cotto make cover was working on this machine around r30000
PerlJam yeah, it's part of Devel::Cover (which I thought was already installed, but apparently not)
allison Whiteknight: yup 22:41
alvar I don't know if this is a known bug: on OS X the parrot test suite pops up some OS X crash report windows, but everything is/seems to be fine
allison Whiteknight: you about to go with the calling conventions branch?
Whiteknight okay, I was going to do the calling_conventions branch right now, but I'll let you go first
allison Whiteknight: nope, go ahead, I'm still going to run the coding standards tests
chromatic alvar, we have some segfaults marked as TODO, so they pass (because we expect them to segfault), but they pop up crash reporters on some systems. 22:42
Whiteknight I haven't run any tests yes, you go first, allison
chromatic We should probably fix those segfaults soonish.
alvar chromatic: yes; so it is known, good ;)
chromatic Maybe we can sic register on one, if we're feeling especially cruel!
cotto allison, should PObj_is_shared_TEST(x) be true if a STRING is shared between interpreters?
dalek r32906 | Whiteknight++ | calling_conventions:
: [calling_conventions] bringing branch up-to-date with trunk, in preparation for a merge
diff: www.parrotvm.org/svn/parrot/revision?rev=32906
allison Whiteknight: looks like the merge brought in some bugs, so I'll let you know how that goes 22:43
alvar and another question: what is the state of threads in parrot?
allison cotto: only if it's been marked as shared
alvar the docs seem to be a little bit incomplete ;) 22:44
allison alvar: we have an implementation of threads, but plan to make extensive revisions in the next year or so
cotto If it hasn't been marked, that's a bug?
alvar allison: OK, i saw this in the roadmap ;)
Whiteknight looks forward to working on the threading system
alvar allison: would it be a good idea to build something which relies on threads on the current implementation? 22:45
(or would it help to develop the new version? ;) )
allison cotto: well, depends on how it ended up shared between interpreters, but generally anything that is intended to be shared should be marked as shared, and anything that isn't marked as shared is probably only accidentally shared, so, yes, it can be considered a bug 22:46
alvar: you will likely run into some rough spots in the current threading implementation, but a real use for the threads implementation is a great way to shake out the bugs 22:47
alvar hehe ;)
i have to build a small server which uses some more or less big data structures for searching lots of words 22:48
22:48 dalek joined
allison alvar: so, I guess it depends on whether you're more of a "mountain bike" or "racing bike" kind of programmer 22:48
Whiteknight what's the difference?
...besides the number of crashes
alvar the current implementation is with mod_perl und pre-shared data, but this is ugly and I have to rewrite it and perhaps it would be better to use parrot then perl 5 ... ;)
Infinoid big gnarly wheels and a penchant for running over small woodland creatures
alvar allison: puh, what sdo you define asmountain bike or racing bike? ;) 22:49
s/sdo/do/
moritz I guess mountain bike programmer = likes obstacles and gets around/above them, racing bike programmer = wants fast results NOW 22:50
Infinoid the difference is in the number of yaks, probably.
allison alvar: "mountain bike" would be enjoying exploring rugged terrain, not minding a few bumps
particle the racing bike programmer wears a funny hat, too
chromatic If a mountain biker sees a goat, that's wonderful. If a racing biker sees a goat, something's wrong. 22:51
allison alvar: "racing bike" would be enjoying speed, with the way paved out for you so it's all smooth
alvar allison: hmmm, usually both is fine, haha ;)
Whiteknight allison: I've got a failure in t/compilers/pct/complete_workflow.t, so I cant merge yet
and I'm heading out to dinner soon, so I probably wont get it done before then 22:52
alvar so, that parrot is work in progress is not too much a problem ... and helping to make it better is always a good idea
moritz #
(sorry, weird c&p)
Whiteknight (heling to make it better)++ 22:53
chromatic (waving to goats as you fly by)++
allison Whiteknight: I seem to have headerizer problems, so can't merge yet either, but will likely merge while you're at dinner 22:54
"watch out for goats in the GC system"
bugs, features, and goats 22:55
Whiteknight it's my experience that headerizer usually causes problems. I try not to run it anymore
allison Whiteknight: in this case, I think it's that I have old headers, and just merged in a pile of new headers from trunk 22:56
Whiteknight eventually I'll go through it's output and fix all the missing function-level documentation snippets
allison ah... got it, PARROT_API changed to PARROT_EXPORT in trunk 22:58
but my new files are all still PARROT_API 22:59
chromatic You need a realclean to account for that.
allison chromatic: did that, but also have to change PARROT_API to PARROT_EXPORT in all my files 23:03
done and rebuilding
23:04 apeiron joined
chromatic Yes, you sent that line as I finished typing mine. 23:04
particle allison: just make sure you don't modify NEWS or DEPRECATED.pod with PARROT_EXPORT 23:06
allison particle: it was 3 files, so I just did it in vim, not perl -p -ie 23:07
23:07 rafl joined
allison chromatic: the joys of asynchronous communication 23:08
chromatic <Bool:True> 23:10
moritz <Bool:Tr<Bool:True>ue> 23:11
chromatic Dumb things I've personally said to Robin Milner: "Conjunctive types do seem useful." 23:12
dalek r32907 | allison++ | pdd22io: 23:20
: [pdd22io] Convert all PARROT_API to PARROT_EXPORT to match change in trunk.
diff: www.parrotvm.org/svn/parrot/revision?rev=32907
23:23 sjansen joined 23:41 bacek_ joined
dalek r32908 | allison++ | pdd22io: 23:47
: [pdd22io] Update FileHandle tests to use quotes now required around method names.
diff: www.parrotvm.org/svn/parrot/revision?rev=32908

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!