Parrot 2.11.0 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Onward and upward with Google Code-In | Please test all nwellnhof/* branches | Merge html_cleanup
Set by moderator on 24 December 2010.
nwellnhof s/to/two/ 00:00
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#1923) fulltest) at 2_11_0-369-gaa07b32 - Ubuntu 10.10 i386 (g++-4.5) 00:04
00:12 Myhrlin left
dalek rrot/nwellnhof/unicode_io: b0db14f | nwellnhof++ | src/io/buffer.c:
Flush write buffer in Parrot_io_readline_buffer

Now that we call it directly...		 00:14
00:15 Myhrlin joined
cotto_work Is Kristian Inanov here? 00:22
Yuki`N dukeleto, looks like the CI project will probably take up most of the time I have until the end of GCI. 00:31
I'm not quite sure if it's very conducive to giving GCI credit for, either. :P 00:32
Looks like matt and kapace bumped up MappedByteArray without me, eh. 00:33
Actually no, doesn't look like it. I'll get working on that then, if someone can approve, socghop.appspot.com/gci/task/show/g...9360069576 00:35
cotto_work Yuki`N: done 00:37
Yuki`N Thanks. 00:38
00:41 kennym left 00:51 Matt_ joined
Yuki`N cotto, mappedbytearray is at 82% on my machine. 00:58
aloha, coverage? 00:59
aloha Yuki`N: coverage is cv.perl6.cz or tapir2.ro.vutbr.cz/cover/cover-results/
Yuki`N It's less there, hmm. 01:00
cotto_work Yuki`N: that's not updated continuously.
Yuki`N Ah. 01:01
Not sure how I can complete the 30% at this rate.
rfw Yuki`N: gonna have to cover all 112% of mappedbytearray 01:02
Yuki`N oh dear. 01:03
cotto_work rfw: That's how it has to be. 100% just isn't good enough.
rfw cotto_work: parrot is 112% functional!
Matt_ Yuki`N: Yes - I brought the coverage up to 82% today. 01:04
The task said it's at < 10% but it was actually at 40% when I claimed the task 01:05
Yuki`N I see.
I'll bring it as close to 100% as possible.
Matt_ cotto_work: Can you look at this: github.com/parrot/parrot/pull/57
whiteknight bah, we should be deprecating, removing, and cutting with a dull knife the Scalar PMC, not increasing test coverage 01:14
rfw i think i'll stop working on darter for a bit and go back to test coverage 01:16
01:16 elmex left 01:17 elmex joined
Matt_ whiteknight: then its a good thing I didn't go past +20% cause I was gonna do double the task required 01:19
cotto_work Matt_: will do, though probably later this evening. 01:20
Yuki`N I wish there was a way to make mmap() fail on purpose.
Unfortunately my system will never run out of memory.
Nor is it a good idea to make it do so.
cotto_work especially not as part of the test suite
rfw is make quickcover still broken? 01:21
Matt_ rfw: It is for me on GCC 4.5.1 01:22
but make quickcover_new works great
rfw make: *** No rule to make target `quickcover_new'. Stop. 01:23
do i need to pull in a branch?
Matt_ it should work with the latest master
did you maybe do a realclean and forget to configure again?
rfw nope
same thing
oh 01:24
whoops
pulled from wrong master
heh
Yuki`N Let's see what magic I can accomplish with my 13 new exception-handling tests. 01:25
rfw someone: www.google-melange.com/gci/task/sho...9361526246 01:33
dalek website: jkeenan++ | On Roadmaps and Teams
website: www.parrot.org/content/roadmaps-and-teams
cotto_work rfw: done. That'll be an interesting one. PMCProxy is unusual. 01:34
01:35 kid51 is now known as kid51_at_dinner
rfw urk, how do i get pmc keyed again in pir? 01:36
just [key]?
01:38 davidfetter left
cotto_work $P0 = foo['bar'] 01:40
that's it
rfw okay i must be doing something else wrong 01:41
01:45 davidfetter joined
Yuki`N These share_ro VTABLEs are quite annoying. 01:49
cotto_work whiteknight: ping 01:50
whiteknight pong
cotto_work whiteknight: do you see a compelling reason for share_ro not to die?
Yuki`N cotto, I was able to get MappedByteArray from 82% to 96%, it's not quite 30% but can I claim completion anyway? 01:51
cotto_work sure
Yuki`N The only lines that are untested are when mmap fails (which is almost impossible to reliably replicate) and something that can never happen via PIR.
rfw Class 'Array' not found 01:52
am i doing something wrong
new $P3, 'Array'
Yuki`N try ['Array']
rfw did that too 01:53
just got Class ['Array'] not found
cotto_work We no longer have an Array PMC.
rfw oh
how do i declare an array?
cotto_work use one the specialized array types 01:54
rfw uh what are they?
never mind got it 01:55
whiteknight cotto_work: I see a compelling reason for share_ro to die
kill it with fire
cotto_work src/pmc/fixed*array and src/pmc/resizable*array
whiteknight then poke it with a sharp stick
cotto_work sweet
Yuki`N whiteknight, will do.
cotto_work whiteknight: can you make a gci task?
whiteknight as far as I am aware it's basically a hold over from the old threading system and STM system
Yuki`N My fork of parrot is quite dirty for some reason 01:56
whiteknight STM was ripped out, and I want to rip out our current threads too
Yuki`N It's got all sorts of commits not in master.
I've submitted a pull request for MappedByteArray, and also work to socghop.appspot.com/gci/task/show/g...9360069576 01:59
cotto_work plobsing, nwellnhof ping 02:00
Yuki`N: did you read docs/project/git_workflow.pod? 02:01
Yuki`N nope.
rfw whee i think i finised 02:02
finished
02:02 elmex left, elmex joined
cotto_work Yuki`N: every time you don't read git_workflow, dukeleto kills a kitten 02:03
Yuki`N Lol.
It doesn't really apply to me as much, as I don't have direct access to the master parrot repository.
Instead my case is much more complex.
cotto_work Yuki`N: you can fork and clone 02:04
that's the beauty of git, among other things
Yuki`N Yeah.
git checkout bobby/tables # make sure we are on the branch
git rebase origin/bobby/tables # get the latest sql injections
who wrote this
rfw hahahaha 02:05
cotto_work git blame
Yuki`N dukeleto. :(
rfw Yuki`N: well, if whiteknight wrote it, it would probably be git checkout LOL/CATZ
whiteknight what did I write? 02:06
rfw [15:04:41] <@Yuki`N> git checkout bobby/tables # make sure we are on the branch
whiteknight oh, right
rfw i'm kind of confused - how do i change a 'Class' into a 'PMCProxy'? 02:07
can i just cast it directly?
i.e. replace Class with PMCProxy 02:08
Yuki`N Now to see if I can clean up my history at all. 02:09
It's really quite strange. 02:11
Github says there are 24 commits to be pulled in, but only 1 changed file. 02:12
I don't know if I should just rebase the commits out.
rfw Yuki`N: did you do 24 commits on one test 02:13
Yuki`N Nope. 02:14
rfw then, what
Yuki`N The 23 other commits aren't even mine.
rfw oh
did you merge someone else's crap in or something
Yuki`N Yeah. 02:15
I just
Rebased it all out. :/
rfw :/ 02:16
rfw rebases Yuki`N
Yuki`N whiteknight, since you're the one who usually handles the pull requests, github.com/parrot/parrot/pull/58
And then approve socghop.appspot.com/gci/task/show/g...9360069576 whenever
Kristaba Can someone explain me how to add a file in the project (to add it in the make and headerizer process)? 02:17
02:18 gbacon joined
Yuki`N You can edit the makefile template in config/gen/makefiles/root.in 02:18
You can run headerizer on individual .c files if you want, `perl tools/dev/headerizer.pl path/to/file.c' 02:20
Not quite sure how you're *supposed* to do it, though.
whiteknight Kristaba: sure. Adding a file has two basic steps: Create the new file, and add it to the makefile in config/gen/makefiles/root.in 02:21
rfw arghwhywon'tyouworkyoustupidtest
whiteknight Setting up a file with headerizer is also two steps: Add tags at the top of the file to tell headerizer where the function prototypes should go, and add tags to the header file to tell where to put them 02:22
rfw how do i typecast in parrot?
whiteknight rfw: what do you mean?
rfw $P3 = new ['PMCProxy']; $P3 = $P0 doesn't seem to work
it just doesn't want to be a pmcproxy
whiteknight rfw: the "=" operator is sort of stupid
rfw what do i use instead? 02:23
whiteknight there are two opcodes: set and assign. I think "=" is assign
rfw ah, so i want "set"?
whiteknight you may want to try set $P3, $P0
try i
it
or it may be the other way around
try both
I get confused
rfw :<
dalek rrot/gci_mappedbytearray_coverage: e66d2fd | Yuki`N++ | t/pmc/mappedbytearray.t:
Almost fully cover MappedByteArray.
 02:24
kid51_at_dinner rfw: ping 02:30
02:30 kid51_at_dinner is now known as kid51
rfw kid51: pong 02:30
kid51 I got your message about --silent. 02:31
You are partially correct.
rfw ah
kid51 --silent does silence the output of the individual configuration steps
but it doesn't silence the 'pre-' and 'post-' messages.
rfw oh yeah, i noticed
Yuki`N whiteknight, I'm going to make a branch and get rid of all the share_ro's. 02:35
With share_ro removal I also increase the Object PMC's coverage a good amount.
cotto_work Yuki`N: it'll help with a lot of PMCs. 02:36
whiteknight Yuki`N: you want that to be a GCI task so you can get points for it?
cotto_work I'd like to do a bit of digging before we ditch it.
just to be sure it's as useless as I think it is
Yuki`N whiteknight, sure, if it's not too much trouble. 02:37
dalek rrot: e66d2fd | Yuki`N++ | t/pmc/mappedbytearray.t:
Almost fully cover MappedByteArray.
Yuki`N cotto_work, yeah, it'll be branched.
I added one test to object, though, that's in my master right now I believe.
whiteknight, I'll probably take a GCI task for that, and a GCI task for object since I wrote (a) test for object, but the other pmc's that increase are just whatever. :P 02:38
rfw okay, this really really doesn't like me
Yuki`N I think string will reach 99% or 100% without share_ro.
whiteknight rfw: what's up? Sounds like you need a hand
rfw whiteknight: i can't seem to ever call the parents method
$P3 = new ['PMCProxy']; $P3 = get_class $P0; $P4 = new ['FixedPMCArray']; $P4 = $P3.'parents'() is what i have right now 02:39
cotto_work share_ro is definitely an STM leftover
rfw and $P3 is definitely a PMCProxy
so i don't understand why it doesn't work
Matt_ can someone go through this really quick: www.google-melange.com/gci/task/sho...1720#c4002 02:40
cotto_work Yuki`N: do you know what you need to do to nuke share_ro or should I spell it out a bit in the gci task? 02:41
I don't mind if it helps, but there's also no need to bother telling you things you alread know. 02:42
Yuki`N cotto_work, besides removing all the VTABLE implementations I'm not sure.
cotto_work *already
Yuki`N There's some master list of VTABLEs, isn't there?
Kristaba whiteknight: I tried to add my files into root.in, but 'make headerizer' can't headerizer these files
cotto_work it's basically that and removing it from src/vtable.tbl
Yuki`N ah ok
whiteknight Yuki`N: www.google-melange.com/gci/task/sug...332443?s=0 I don't think that task is public yet. When cotto or somebody else makes it public you can take it
Kristaba: what error does it give you?
Yuki`N Ok.
cotto_work published 02:43
Yuki`N Is there a special trac bit you need to do stuff besides comment and make tickets? 02:44
e.g. closing them.
Claimed btw.
Kristaba whiteknight: I have no error, but the files aren't headerized after a 'make headerizer'
whiteknight rfw: wait, what are you trying to test?
rfw whiteknight: method parents of PMCProxy
cotto_work Yuki`N: you should be able to do all that
Yuki`N Last time I looked at an open ticket I couldn't take assign, or stuff like that. 02:45
cotto_work you can now
Kristaba whiteknight: Maybe I didn't add the files name in the good place into root.in
cotto_work I updated your permissions.
whiteknight rfw: try $P0 = new ['Integer'], $P1 = get_class $P0, $P2 = $P1.'parents'() 02:46
Kristaba: there is a long list in there somewhere of all the files
rfw whiteknight: what parents does integer have that i can assert against?
whiteknight rfw: I believe Integer inherits from Scalar, currently
rfw ah okay
dalek rrot/gci_scalar_coverage: 85e9439 | (Matt Rajca)++ | t/pmc/float.t:
Increased test coverage for Scalar PMC
 02:47
Kristaba whiteknight: Yes, in the "Dependencies" section? 02:48
whiteknight INTERP_O_FILES 02:49
line 441
Add your new file to that list. Then you need to make realclean and reconfigure 02:50
rfw: There are two types of things: PMCProxy is the "Class" for built-in types. Class is for objects defined in PIR 02:53
so get_class on a built-in type returns a PMCProxy
rfw whiteknight: ah
well
i'm getting class on PMCProxy now
and checking if a parent is Class
it seems to work
whiteknight right, get_class on PMCProxy will return a different PMCProxy 02:54
It's the PMCProxy PMCProxy
and the parent of that should be Class, I think
rfw is there a PMCProxy PMCProxy PMCProxy
whiteknight no, doesn't recurse infinitely
though that would be funny
02:55 davidfetter left
Yuki`N Apparently there's also a share VTABLE. 02:55
02:56 bluescreen left
whiteknight share? 02:56
Yuki`N Yeah.
whiteknight KILL IT
Kristaba whiteknight: Perfect, thank you ;)
Yuki`N RIGHTO. 02:57
whiteknight Kristaba: We don't do this too often, so I'm sure we forget the details
Yuki`N++
rfw yes i win 02:58
96% coverage!
Yuki`N rfw, :P
rfw oh wait
wrong file!
Yuki`N Time to make test!
rfw 90.3%
aloha 0.903
whiteknight rfw++ 03:00
Yuki`N whiteknight, "Badly balanced PMC source" is an error I'm getting.
whiteknight aloha, coverage?
aloha whiteknight: coverage is cv.perl6.cz or tapir2.ro.vutbr.cz/cover/cover-results/
rfw whiteknight: i was looking at fixedpmcarray by accident, heh
03:00 bluescreen joined
whiteknight Yuki`N: That usually pops up when there aren't matching braces 03:00
make sure you got all the { and } pairs 03:01
cotto_work goes home
Yuki`N Oh I accidentally deleted one to many }'s
rfw whiteknight: can you github.com/parrot/parrot/pull/59
and www.google-melange.com/gci/task/sho...9361526246
dalek rrot: 46c8406 | jkeenan++ | / (3 files):
Implement option of fully silent configuration.

Expand scope of '--silent' command-line option to handle pre- and post-configuration messages. Cf.: trac.parrot.org/parrot/ticket/1897.
  rfw++ for spotting previous partial implementation..		 03:02
rrot: a3bd233 | jkeenan++ | t/pmc/mappedbytearray.t:
Merge branch 'master' of git@github.com:parrot/parrot
rfw whiteknight: could you close 03:04
nwellnhof whiteknight: pong
whiteknight nwellnhof: pong
dalek rrot/gci_pmcproxy_coverage: 74eae86 | Whiteknight++ | / (21 files):
Merge branch 'master' of github.com:parrot/parrot
 03:05
rrot/gci_pmcproxy_coverage: eb883bf | (Tony Young)++ | t/pmc/pmcproxy.t:
Increasing PMCProxy test coverage
rrot/gci_pmcproxy_coverage: d2e987f | Whiteknight++ | / (100 files):
Merge branch 'master' of git://github.com/rofflwaffls/parrot
rfw how do i change my name with dalek
nwellnhof whiteknight: you pinged a while before
rfw: on github, probably
rfw nwellnhof: hm but Yuki`N's github name isn't Yuki`N 03:06
so i don't know how he does it
dalek rrot: d2e987f | Whiteknight++ | / (100 files):
Merge branch 'master' of git://github.com/rofflwaffls/parrot
rrot: 5e1a75c | Whiteknight++ | t/pmc/pmcproxy.t:
Merge remote branch 'origin/gci_pmcproxy_coverage'
whiteknight nwellnhof: I don't remember why. Probably safe to ignore
nwellnhof sorry, it was cotto
dalek TT #1897 closed by jkeenan++: Implement 'silent' configuration
TT #1897: trac.parrot.org/parrot/ticket/1897
nwellnhof cotto: pong
Matt_ whiteknight: can you close this too: www.google-melange.com/gci/task/sho...1720#c4002 03:07
Kristaba Oh, another question :D
dalek rrot: 19d4723 | Whiteknight++ | t/pmc/float.t:
Merge remote branch 'origin/gci_scalar_coverage'
Kristaba I have to change do_sub_pragmas calls to PackFile_fixup_subs, but I'm not sure of how to transform the old "PackFile_ByteCode *self" argument
Yuki`N rfw, in credits.
rfw Yuki`N: but i am in credits 03:08
Yuki`N CREDITS, your irc nick I believe is what dalek puts.
Kristaba Simply put it in interp->code ?
Yuki`N Make sure your full name is correct.
It might not be in trunk.
rather
master.
rfw github.com/parrot/parrot/blob/master/CREDITS
line 1057-1060
oh 03:09
should be U
whiteknight: could you change my A: rfw to U: rfw
cotto_work nwellnhof: do you see any value in share_ro?
It's on the chopping block and I want to make sure I'm not missing anything. 03:10
dalek rrot: b3e5f5e | Whiteknight++ | CREDITS:
Fix entry for rfw++
cotto_work really goes home. Will backscroll
rfw do i get free karma for getting my nick changed? \\o/
whiteknight rfw: you've earned plenty of karma 03:12
karma rfw?
aloha rfw? has karma of 0.
rfw karma rfw
aloha rfw has karma of 11.
nwellnhof cotto: i think it's used by the thread implementation
Yuki`N At this point I know that my karma goes directly to N lol.
karma Yuki`N
aloha Yuki`N has karma of 0.
Yuki`N karma N
aloha N has karma of 24.
rfw karma C
aloha C has karma of 150.
Yuki`N harr harr
TT1918 contains my findings. 03:14
It's like they weren't even there.
Checking the new coverage % of object.pmc now.
kid51 cotto: This GCI task should either be deleted or closed: www.google-melange.com/gci/task/sho...9315789586
dalek rrot/nwellnhof/unicode_io: 2e28508 | nwellnhof++ | / (10 files):
Implement get_chars in ByteBuffer and MappedByteArray

This new method works for any encoding.
rrot/nwellnhof/unicode_io: 70427c1 | nwellnhof++ | / (6 files):
Convert some other code to Parrot_find_encoding_by_string
kid51 Can you or dukeleto take care of that?
rfw holy crap the role pmc is huge
whiteknight rfw: and the kicker is that it's mostly unused, mostly untested, mostly unloved 03:16
rfw whiteknight: Title: Evaluate possibly-dead functions in src/packfile/api.c 03:18
what does this involve?
whiteknight let me look. Link? 03:19
rfw www.google-melange.com/gci/task/sho...9372316240
whiteknight rfw: There are a few functions in that file that aren't covered by any tests, and I suspect they are dead 03:20
moderator Parrot 2.11.0 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | More gci tasks | Merge html_cleanup | Get Plumage working | Review current embedding api | Study EventHandler | Run examples_tests, test HLLs before merges 03:20
whiteknight so the task is to search around for them and see if they are dead and can be removed 03:20
rfw ah
requested
whiteknight it's a research task, so you don't actually have to remove any code
Yuki`N dalek is usually slow with the TT notifications. hmm 03:21
rfw i'm probably going to remove code anyway and see if it explodes
whiteknight (you can if you want)
that's a good research technique
it's called "slash and burn"
rfw Segment_new is definitely being used 03:22
dalek rrot/nwellnhof/unicode_io: 113b96f | nwellnhof++ | src/ (2 files):
Fix two compiler warnings
whiteknight rfw: it's being referenced from one place that I can see, but I don't think it's ever actually being called
and the coverage numbers support that idea
rfw hm
whiteknight: i'm getting matches for it 03:23
in all the object files
dalek TT #1918 created by Yuki`N++: Remove share and share_ro VTABLEs
TT #1918: trac.parrot.org/parrot/ticket/1918
Yuki`N There we go.
whiteknight you still might want to try removing it to see what happens
rfw i don't think i can remove it safely
since it's being registered with something 03:24
doesn't look quite clean to do
whiteknight okay. It's your call 03:25
rfw okay, i can remove Parrot_switch_to_cs_by_nr and PackFile_remove_segment_by_name safely
whiteknight awesome. 03:26
okay, I need to head off to bed now. I'll talk to you guys later
rfw night whiteknight 03:27
Yuki`N Night.
Whooh!
aloha, coverage
whiteknight goodnight
03:27 whiteknight left
Yuki`N aloha, coverage? 03:28
aloha Yuki`N: coverage is cv.perl6.cz or tapir2.ro.vutbr.cz/cover/cover-results/
Yuki`N object is up from 83% to 93%!
Kristaba OMG, I have a strange build error... 03:31
"/home/leo/GCI/parrot/blib/lib/libparrot.so: undefined reference to `CONST_STRING' "
rfw oh i lie
they can be removed
i'm a horrible horrible liar 03:32
cotto ~~
Yuki`N Kristaba, you should get a file along with that error. 03:33
cotto, could I take credit for a 5% increase of object.pmc, even though it really resulted because of the share_ro removal?
cotto if you're on the share_ro task, sure 03:34
rfw cotto: could you look at www.google-melange.com/gci/task/sho...9372316240 03:36
Kristaba Yuki'N: Oh, which file?
cotto kid51, deleted
Yuki`N Kristaba, the error should point you to a file. 03:37
kid51 cotto: gracias
Yuki`N I assume you've left out an include somewhere.
Kristaba The complete error is : cc -o miniparrot frontend/parrot/main.o src/null_config.o src/longopt.o \\ 03:39
Yuki`N cotto, I requested claim for socghop.appspot.com/gci/task/show/g...9360674520
Kristaba -Wl,-rpath="/home/leo/GCI/parrot"/blib/lib -L/home/leo/GCI/parrot/blib/lib -lparrot -lm -L/usr/lib -licuuc -licudata -lm -ldl -lm -lpthread -lcrypt -lrt -lgmp -lreadline -fstack-protector -L/usr/local/lib -Wl,-E
/home/leo/GCI/parrot/blib/lib/libparrot.so: undefined reference to `CONST_STRING'
collect2: ld returned 1 exit status
03:39 LaVolta joined
cotto Yuki`N, done 03:39
dalek TT #1919 created by rfw++: Removal of dead functions in src/packfile/api.c 03:40
TT #1919: trac.parrot.org/parrot/ticket/1919
Yuki`N cotto, submitted for review.
>jumpyshoes 176 03:41
ffs jumpyshoes.
rfw cotto: could you look at my removal thing too
Yuki`N: if google says what jumpyshoes is doing is fine I'M GOING TO DO THAT TOO
Yuki`N Lol. 03:42
We have Darter.
rfw yes, four points per person
not much
Yuki`N Feels much more intense though `-`
rfw think i'm going to submit another patch to OBATA Akio
Yuki`N OBATA-sama. 03:43
cotto Yuki`N, approved 03:45
Yuki`N Thanks~
cotto rfw, looking
rfw thanks
cotto rfw, can you expand on what you mean by "safe to be nuked"? Does that mean they're defined but never used, only used from unreachable/unreached branches, etc 03:47
rfw cotto: they're not used ever it would seem
i can find no references to those functions whatsoever
cotto mention that in the ticket then. Also, why do you say that PackFile_Segment_new should be replaced with NULLfunc and PackFile_Annotations_dump? Reasoning is important to communicate. 03:50
03:50 Matt_ left 03:53 TonyC left
rfw cotto: i'm not sure about the PackFile_{Segment,Annotations}, actually 03:54
they seem to have specific implementations but they're not being called
03:54 TonyC joined
cotto rfw, they might be called indirectly 03:55
rfw actually 03:56
i think there should be tests written for them if they are indeed called
dalek rrot: fbbcdad | nwellnhof++ | / (7 files):
Fix more compiler warnings
 04:00
nwellnhof incredible, the commit hash consists only of letters 04:01
Kristaba Yuki`N: Ok, it was a complex problem due to the c2str system and an include I left out, thank you ;) 04:02
plobsing nwellnhof: that has ~0.1% probability of happening. was bound to happen eventually. 04:05
nwellnhof plobsing: lol, i just made the same calculation. but even the first eight characters are letters. that's a 0.03% probability. 04:06
plobsing yeah, but only the first 7 are displayed
cotto NERDS 04:07
04:07 Kristaba left
Coke that's not a valid sha. 04:21
rfw you know i can'ta even tell how most of PackFile is being called 04:22
dukeleto ~~ 04:24
cotto rfw, it's pretty indirect 04:25
rfw yeah, i can see that now
Yuki`N I think I just got Opcode up from 37% to at least 90%, hah. 04:27
cotto PackFile_funcs_register accepts a constant and a struct of function pointers. All the functions are called in terms of those constants and function pointers, iirc.
plobsing yes, it's another vtable pattern 04:28
cotto there might be exceptions, but that't the general pattern 04:29
Yuki`N Yes, Opcode up to 93%. 04:32
dukeleto, can you approve socghop.appspot.com/gci/task/show/g...9360974537 04:34
dukeleto Yuki`N++
Yuki`N: accepted 04:36
Yuki`N I have a pull request for those tests plus some a test from the Object class. github.com/parrot/parrot/pull/60 04:37
Object PMC, rather.
Can someone explain how ResizablePMCArray is autovivifying? 04:44
I see no code that supports that claim.
Coke gets kcachegrind working on his OSX box. 04:48
Yuki`N Never mind, I found it
dukeleto Yuki`N: it is deep down the rabbit hole
Yuki`N Indeed it is.
rfw cotto: trac.parrot.org/parrot/ticket/1919 04:50
god that code is painful
Yuki`N Haha. 04:54
dukeleto, can you close? socghop.appspot.com/gci/task/show/g...9360974537
rfw or dukeleto, could you look at that ticket 04:55
cotto rfw, like a boss
rfw oh, cotto's still there
cotto approved
well-earned, I'm sure
rfw hooray
Yuki`N How does headerizer work, really?
rfw that goddamn dynamic dispatch 04:56
Yuki`N Lol.
I need a header put in because the file won't build.
But apparently you need an .o file before you can headerize?
cotto Yuki`N, I don't think headerizer cares about the actual object file. It just uses that because that's what the Makefile deals with. 04:57
Yuki`N Ohhh
That makes sense now.
And my build continues on its merry way.
cotto could you close the link above?
cotto Yuki`N, ok. Are those in a branch or in master in parrot's repo? 04:58
i.e. has someone pulled them?
dalek rrot: ac94044 | nwellnhof++ | src/hash.c:
Remove a Parrot_const_cast
Yuki`N cotto, Nobody's pulled them yet. 04:59
04:59 theory left
cotto ok 04:59
reviewing now 05:00
dalek rrot: 8432080 | Yuki`N++ | t/pmc/object.t:
Test another Object PMC edge-case.
 05:02
rrot: 001af17 | Yuki`N++ | t/pmc/opcode.t:
Increase coverage of the Opcode PMC.
cotto accepted
er, approved
whatever that thing you asked for was
Yuki`N Ah, thanks.
xD
Can you take care of accepting this one too? socghop.appspot.com/gci/task/show/g...9373964971 05:06
I'm taking care of that pesty autovivification.
cotto done 05:07
05:09 kid51 left
dalek rrot/encapsulate-main: 51ac88d | plobsing++ | src/packfile/api.c:
always run :load subs, no matter how something has been loaded
 05:14
rrot/encapsulate-main: 0f6de1b | plobsing++ | / (25 files):
add :main tags to benchmark examples
05:16 nwellnhof_ joined
Yuki`N Killed it 05:17
with fire.
dalek tracwiki: v14 | nwellnhof++ | BuildWarnings 05:19
tracwiki: trac.parrot.org/parrot/wiki/BuildWa...ction=diff
05:19 nwellnhof left, nwellnhof_ is now known as nwellnhof
Yuki`N Things look good so far. 05:24
05:24 plobsing left
Yuki`N Fulltest only failed because I forgot to change the plan. 05:24
05:26 gbacon left
cotto Yuki`N, branch? 05:30
Yuki`N cotto, pull request on github. 05:31
05:37 rurban_ joined 05:40 rurban left, rurban_ is now known as rurban, nwellnhof left
Yuki`N cotto, socghop.appspot.com/gci/task/show/g...9373964971 05:48
cotto Yuki`N, great. I'll accept that. Have you tested with Rakudo and/or Partcl? 05:59
Yuki`N Nope, not yet.
I'll clone the rakudo repo now I suppose.
cotto done 06:00
ok. Make sure to install parrot too. perl Configure.pl --prefix=/path/to/install/location
and make sure that the executables are in your path so that Rakudo's Configure.pl can find them.
Yuki`N Yeah. 06:01
I guess I'll take the other Opcode task. 06:05
Still build Rakudo apparently.
*building 06:06
cotto It looks like we've had at least a 1.9% increase in overall test coverage over the course of gci. 06:11
Yuki`N Rakudo doesn't build. 06:14
But that's just because of api functions
cotto Yuki`N, what do you mean? 06:18
Yuki`N um
aloha, nopaste?
aloha Yuki`N: nopaste is is nopaste.snit.ch (works with the script in $_PARROT/tools/dev/nopaste.pl)
Yuki`N pastebin.com/jQnJuykp 06:19
Compile issues.
cotto indeed 06:20
possibly from an old parrot?
Yuki`N Shouldn't be.
cotto many thing shouldn't be
Yuki`N How can I make sure it's not from an old parrot? 06:21
cotto nuke the install, run reconfig, install again
Yuki`N Heh, I didn't install to a prefix.
cotto also, make sure that there's nothing in rakudo's parrot_install dir. It builds parrot there if you don't tell it otherwise.
Yuki`N Ok, nuked the install, and installing this time to a prefix 06:27
06:28 GeJ left
Yuki`N LOL DAMNIT 06:32
I HATE YOU LINUX
Good thing it's just a vm.
dalek rrot: f2c7203 | Yuki`N++ | / (2 files):
Kill autovivification in Fixed/ResizablePMCArray.

Signed-off-by: Christoph Otto <christoph_git@mksig.org>		 06:33
cotto That was premature. 06:34
Yuki`N No.
I installed to ~/parrot_install
And of course for it's not a bash script so the ~ didn't get expanded. 06:35
So
I ls'd
I saw ~ as a directory.
I `rm -rf ~'
lol
TIME TO RECREATE MY USER.
cotto live and learn
dalek rrot: 61be0ca | nwellnhof++ | config/auto/git_describe.pm:
Change git-describe regex

The hashes returned from git-describe can be longer than seven chars if they're ambigious.		 06:38
sorear Yuki`N: i'm suprised parrot works at all with a ~ in the path 06:42
parrot has path reparsing issues; I've seen reports of spaces and hanzi causing problems 06:43
Yuki`N Hah. 06:44
Aha, this clone of parrot is going to take a while. 06:47
cotto slow connection?
Yuki`N Yeah, like 180kb/s down
I'm actually surprised the whole world didn't explode
Receiving objects: 54% (186957/346215), 49.39 MiB | 179 KiB/s oh my 06:51
cotto I may end up reverting that if rakudo starts failing spectests. 06:58
06:59 JimmyZ joined
Yuki`N Apparently. 07:00
My git-compiled version of parrot is too old. 07:01
I am confused.
cotto lolwut?
Yuki`N I have no idea. 07:02
It says I need RELEASE_2_11 and I have 2_10.
cotto and you definitely don't have an old parrot installed? 07:05
Yuki`N [nlum@yukiko rakudo]$ perl Configure.pl --parrot-config=/home/nlum/parrot_install/bin/parrot_config
Reading configuration information from /home/nlum/parrot_install/bin/parrot_config ...
===SORRY!===
Parrot revision RELEASE_2_11_0-343-g0264a39 required (currently RELEASE_2_10_1-1339-g39795d6)
Oh my fork apparently doesn't have tags. 07:06
Oh, git. 07:08
07:11 JimmyZ left
cotto Coke, ping 07:13
Yuki`N Trying to build Rakudo again. 07:17
cotto, it's still running spectests but they've all passed so far. 07:38
cotto Yuki`N, they'll take a while 07:42
still running here too
partcl is broken, so it's not clear your change breaks it more 07:43
Yuki`N ignorecase is broken
I don't think that's related to my change though.
07:45 Yuki`N left 07:48 fperrad joined 07:52 Yuki`N joined
Yuki`N Two tests failed so far. 07:52
cotto Yuki`N, how fast is your machine? 08:02
Yuki`N Um.
It's a VM.
2 cores of my i7.
cotto "not very"?
doesn't sound too bad
Yuki`N Yeah, not very.
I'm almost done. 08:03
Running sech.t
trig functions. `-`
cotto me too
exactly now
08:03 contingencyplan joined
Yuki`N How do you pronounce "sinh" and "tanh" lol 08:05
cotto I don't.
Yuki`N I'm on integration. 08:06
Actually these tests are single-threaded.
so
pastebin.com/Gdk4kj9L 08:08
cotto looks like your machine is faster than mine
Yuki`N So all tests actually passed.
And I unfortunately must sleep now. 08:09
cotto 'night
08:09 Yuki`N left 08:41 perlite left
Kapace_ well, that was a really long lunch :O. anyways, that code at the end of parrotlibrary is need to test destroy() I found it from some other test... anyways I'll take another look at it tommorow.. 09:01
Night...
cotto Kapace_, the code doesn't get covered otherwise? 09:14
09:23 sorear left 09:24 TimToady left 09:46 rfw left 09:52 TimToady joined, sorear joined
dalek rrot: 5089dca | cotto++ | / (2 files):
remove unused function Parrot_switch_to_cs_by_nr
 09:56
rrot: ab1f1f9 | cotto++ | / (2 files):
remove unused function PackFile_remove_segment_by_name
rrot: c40e82e | cotto++ | t/ (2 files):
[t] fix FPA test plan
10:02 kennym joined 10:27 LaVolta left
dalek rrot: c2dc001 | cotto++ | t/tools/pbc_dump.t:
[t] add some test coverage for annotation dumping
 10:35
10:37 M_o_C joined 10:40 fbrito joined, fbrito left, fbrito joined 10:41 GeJ joined
cotto hi fbrito 10:48
fbrito cotto hello :)
cotto Happy Really Old Year!
fbrito thank you :D 10:49
can't believe that we are going to 2011 10:50
next year will be the 10th anniversary of the WTC attack. 10 years :o. I remember as if it were yesterday 10:51
cotto It looks to be a good year, anniversaries notwithstanding. 10:59
fbrito :D 11:03
cotto fbrito, how's the sub task going? 11:15
fbrito started it this morning
so far, so good :)
have to run quickcover again to see how much I have already increased
quickcover_new is awesome :) 11:36
81.9% to 88.3%... hm :P not yet 11:45
cotto so close, yet not so close 11:52
fbrito hm... t/pmc/sub.t is have tests written in PIR 12:04
ah, nevermind. I found what I was looking for :) 12:05
I have to write a test for a Sub function written in C... that's going to be challenging 12:07
cotto fbrito, what are you testing? 12:08
fbrito tapir2.ro.vutbr.cz/cover/cover-resu...b-pmc.html
print_sub_name 12:09
cotto trace PARROT_SUB_TRACE_CALL_FLAG in PIR should do it 12:11
just include interptrace.pasm
I'm not sure if that code is useful, but it shouldn't be tough to exercise. 12:12
fbrito ok, I will have a look. thank you 12:13
cotto np 12:20
'night
fbrito good night
ehhe, 9am here :)
cotto 4am here 12:22
cotto sleeps
fbrito msg cotto I am done with my Sub PMC task. Please take a look on it when you have some time :). Link: www.google-melange.com/gci/task/sho...9361644285 12:41
aloha OK. I'll deliver the message.
13:13 redicaps joined
GeJ Happy New Year everyone! 13:20
fbrito Thank you! 13:22
The same to you
13:24 M_o_C left 13:37 rurban_ joined, M_o_C joined 13:40 rurban left, rurban_ is now known as rurban 14:05 contingencyplan left
dalek rrot/gci_remove_autovivification: cf06c9f | Yuki`N++ | t/pmc/object.t:
Test another Object PMC edge-case.
 14:06
rrot/gci_remove_autovivification: 05fd57c | Yuki`N++ | t/pmc/opcode.t:
Increase coverage of the Opcode PMC.
rrot/gci_remove_autovivification: 45f9597 | Yuki`N++ | / (2 files):
Kill autovivification in Fixed/ResizablePMCArray.
rrot/gci_remove_autovivification: c90627c | Yuki`N++ | t/pmc/fixedpmcarray.t:
Fix plan.
rrot/gci_remove_autovivification: 39795d6 | Yuki`N++ | / (15 files):
Merge branch 'master' of git://github.com/parrot/parrot
rrot/gci_sub_coverage: 2a7826e | fbrito++ | t/pmc/sub.t:
[t] Add tests to Sub PMC
 14:15
rrot/gci_sub_coverage: 5ef9834 | fbrito++ | t/pmc/sub.t:
[t] Add more tests to Sub PMC
rrot/gci_sub_coverage: 99a78e9 | fbrito++ | t/pmc/sub.t:
[t] Add even more tests to Sub PMC
rrot/gci_sub_coverage: f03d78b | fbrito++ | t/pmc/sub.t:
[t] Increase test coverage of Sub
rrot/gci_sub_coverage: 5fd2b05 | fbrito++ | t/pmc/sub.t:
[t] Remove wrong test that I have wrote
14:17 kid51 joined
dalek rrot: 2a7826e | fbrito++ | t/pmc/sub.t:
[t] Add tests to Sub PMC
 14:17
rrot: 5ef9834 | fbrito++ | t/pmc/sub.t:
[t] Add more tests to Sub PMC
rrot: 99a78e9 | fbrito++ | t/pmc/sub.t:
[t] Add even more tests to Sub PMC
rrot: f03d78b | fbrito++ | t/pmc/sub.t:
[t] Increase test coverage of Sub
rrot: 5fd2b05 | fbrito++ | t/pmc/sub.t:
[t] Remove wrong test that I have wrote
rrot: cf65268 | Whiteknight++ | t/pmc/sub.t:
Merge remote branch 'origin/gci_sub_coverage'
14:18 whiteknight joined
fbrito whiteknight: Hi! Happy New Year :) 14:18
and thank you very much for reviewing my work
whiteknight fbrito: happy new year!
no, thank YOU for doing the work! 14:19
kid51 I have a git question 14:21
whiteknight kid51: what question?
kid51 If I have a branch which is tracking the remote 'origin', and I say 'git pull', should I expect that it pulls in all recent commits to origin in addition to any recent commits to my branch? 14:22
(Background info: I have an older version of git (1.5.6.5) in which I created this branch by saying: 14:24
git checkout --track -b tt1908/moretests_target origin 14:25
whiteknight kid51: I think git pull pulls for all branches
kid51 But at $job -- where we've been using git for several months now, with a more recent version -- when I am working in a branch and say 'git pull', I don't seem to get changes merged into the branch. 14:26
whiteknight: So, in that sense, it has a much broader, more potent effect than saying 'svn up' in a branch?
, i.e., a Subversion branch? 14:27
whiteknight i think so
kid51 How then does that differ from 'rebasing' (an operation I have never yet had to perform)? 14:29
fbrito kid51: yesterday I found a interesting link about merge, rebase and cherry-pick: stackoverflow.com/questions/1241720...e-workflow 14:30
an interesting* 14:31
kid51 fbrito: That is interesting, but that's more concerned with what happens "down the line" from where I am right now. 14:33
I'm talking about what happens when I'm working in a branch *before* the point where I'm ready to merge. 14:34
If what I think is happening is indeed happening, then I doubt I would ever need to worry about rebasing-vs-cherry-picking, etc.
But that's not how I'm accustomed to work (coming from a heavy use of SVN branches) 14:35
I don't *really* want to be pulling commits from master into my branch all along.
dalek rrot: 63d7ce6 | Whiteknight++ | /:
Merge remote branch 'origin/gci_remove_autovivification'
 14:37
14:37 LaVolta joined 14:38 LaVolta left
Coke cotto: pong. 14:40
fbrito seen cotto 14:44
aloha cotto was last seen in #parrot 2 hours 21 mins ago saying "4am here".
fbrito he went to bed
14:48 M_o_C left
Coke fbrito: he ping'd me. I pong him. that's how it's done. 14:50
It also would have had the side effect of having aloha deliver any messages to me.
14:51 M_o_C joined 14:53 gbacon joined
fbrito whiteknight: when you have some time, can you please accept my claim request on this task? :D www.google-melange.com/gci/task/sho...9361561487 15:01
whiteknight done 15:02
dalek TT #996 closed by whiteknight++: Socket.pmc should implement readline 15:14
TT #996: trac.parrot.org/parrot/ticket/996
TT #1901 closed by whiteknight++: Parrot_io_get_last_file_position should be removed in the next deprecation ...
TT #1901: trac.parrot.org/parrot/ticket/1901
rrot: e56a22f | Whiteknight++ | DEPRECATED.pod:
Add DEPRECATION notice about Scalar PMC
 15:15
fbrito whiteknight: hm, it seems that your claim request acceptation didn't work 15:16
whiteknight let me check 15:17
fbrito: done, again
msg cotto when you get a chance can you look at TT #1870? If that's the route we want to take I would like to get in the deprecation notice before 3.0 15:18
aloha OK. I'll deliver the message.
fbrito whiteknight: nice! thank you
whiteknight msg Coke Can you take a quick look at TT #657 and see if it's still a problem with the new makefile and the dependency checking mechanisms? Thanks 15:21
aloha OK. I'll deliver the message.
whiteknight Is there anybody here who builds Parrot on Win32? 15:22
15:24 plobsing joined
Coke whiteknight: it's not fixed, but it's a duplicate of #682. 15:27
I'll close it and point to the dup.
whiteknight Coke++ 15:28
dalek TT #1894 closed by whiteknight++: EventHandler's VTABLE_invoke unused and crashes . 15:31
TT #1894: trac.parrot.org/parrot/ticket/1894
TT #657 closed by coke++: libparrot.dll missing entry points
TT #657: trac.parrot.org/parrot/ticket/657
Coke fill_params is a big function.' 15:48
fbrito yeah! just increased code coverage of Task PMC from 86% to 98.5% 15:59
15:59 M_o_C left
whiteknight nice 16:00
dalek TT #1920 created by whiteknight++: Deprecate Threads 16:04
TT #1920: trac.parrot.org/parrot/ticket/1920
16:06 Patterner left, Psyche^ joined 16:07 Psyche^ is now known as Patterner
PacoLinux this site seems interesting : www.1024cores.net/home/parallel-com...on-problem 16:19
16:20 Matt_ joined
dalek rrot/tt1908/moretests_target: 87e332f | jkeenan++ | config/gen/makefiles/root.in:
Add 'moretests' make target: test + examples_tests + codetest (except perlcritict.t).
 16:33
16:35 redicaps left 16:37 whiteknight left
dalek rrot/tt1908/moretests_target: b198652 | jkeenan++ | config/gen/makefiles/root.in:
Typographic cleanup.
 16:39
rrot/tt1908/moretests_target: 4f3ddb7 | jkeenan++ | DEPRECATED.pod:
Merge branch 'master' of git@github.com:parrot/parrot into tt1908/moretests_target
kid51 msg whiteknight Feedback welcome on trac.parrot.org/parrot/ticket/1908#comment:7 16:49
aloha OK. I'll deliver the message.
kid51 msg cotto Feedback welcome on trac.parrot.org/parrot/ticket/1908#comment:7
aloha OK. I'll deliver the message.
fbrito cotto: when you have some time, please take a look on this task: www.google-melange.com/gci/task/sho...9361561487 17:13
msg cotto when you have some time, please take a look on this task: www.google-melange.com/gci/task/sho...9361561487 17:14
aloha OK. I'll deliver the message.
fbrito Happy New Year everyone! :D.
I am leaving right now
17:14 fbrito left
dalek Some 'plobsing' person just gave me a 11604 byte commit packet. They need to be more careful in the future. 17:27
plobsing dalolz
17:31 plobsing_ joined 17:35 plobsing left 17:39 theory joined
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#1938) fulltest) at 2_11_0-427-ge56a22f - Ubuntu 10.10 i386 (g++-4.5 with --optimize) 17:47
17:48 Matt_ left 17:53 kid51 left 17:57 Matt_ joined 18:00 Matt_ left, Matt_ joined
dalek rrot/encapsulate-main: 23f62f8 | plobsing++ | t/pmc/orderedhash.t:
add :main tag to tests added in recent merge
 18:06
plobsing_ damn gci kids. adding tests, tramplin' on lawns, and such.
18:11 fperrad left 18:12 whiteknight joined
Matt_ whiteknight: Do you have some time for a quick review: www.google-melange.com/gci/task/sho...4833#c2002 18:16
18:39 autark left 18:53 autark joined 18:57 jan left 19:02 davidfetter joined
dukeleto Year.new().happy! 19:06
Kapace_ morning, dukeleto, is it 2011 over there already? 19:11
dukeleto Kapace_: in my mind, yes :) 19:22
Kapace_: i still have about 12.5 hours of 2010 left here
Matt_ dukeleto: can you do a quick review :P www.google-melange.com/gci/task/sho...4833#c2002 19:23
whiteknight plobsing: do we have an ETA on encapsulate-main merge?
dukeleto Matt_: i can't do a review now, but i can accept claim requests 19:24
Kapace_ www.google-melange.com/gci/task/sho...9360615067
dukeleto Kapace_: accepted 19:25
Kapace_ thanks
anyone know how to set a custom_free or _clone _func?
whiteknight msg cotto whether we deprecate and remove the current thread implementation now or not, I think we need to get serious about designing a replacement for it. I've got a few ideas. I want to make sure they jive with lorito plans
aloha OK. I'll deliver the message.
whiteknight Kapace: what do you mean?
Kapace_ tapir2.ro.vutbr.cz/cover/cover-resu...t-pmc.html shows that destroy() can call a custom_free_func() 19:26
I haven't seen how to set one anywhere in other tests etc
cotto ~~ 19:27
Coke rakudo: say "hi"
p6eval rakudo f99d08: ( no output )
Coke ... is that our fault?
dalek rrot: 9596abd | Whiteknight++ | NEWS:
two quick news items
cotto I'm popular today. 19:28
Kapace_ morning cotto
cotto hio Kapace_
Kapace_ cotto: from last night, yes I think destroy isn't called on ParrotLibrary if those last few lines aren't in
(according to my quickcover results) 19:29
19:29 M_o_C joined
dalek rrot/gci_error_handling: 56de7e3 | (Matt Rajca)++ | src/packfile/api.c:
Use Parrot_ex_throw_from_c_args to report errors
 19:30
cotto Kapace_, ok. that's suspicious.
Kapace_ it should call destroy() if those lines aren't in? gc isn't picking them up? 19:31
cotto It should be destroyed when the interp shuts down.
dalek rrot/gci_task_coverage: cccac93 | fbrito++ | t/pmc/task.t:
[t] Add new test to Task PMC
 19:32
rrot/gci_task_coverage: 8f87d11 | fbrito++ | t/pmc/task.t:
[t] Add new test to Task PMC [2]
rrot/gci_task_coverage: b9b6a9f | fbrito++ | t/pmc/task.t:
[t] Add new test to Task PMC [3]
rrot/gci_task_coverage: 27625da | fbrito++ | t/pmc/task.t:
[t] Add new test to Task PMC [4]
rrot/gci_task_coverage: 88a870f | fbrito++ | t/pmc/task.t:
[t] Exercise mark VTABLE on Task PMC
rrot/gci_task_coverage: 9fe4cfa | fbrito++ | t/pmc/task.t:
[t] Fix test description on Task PMC
rrot/gci_make_targetnames: 1bb6425 | (Paul Johnson)++ | / (3 files):
Changed cover to fullcover and quickcover to cover
 19:36
19:36 jan joined
Kapace_ cotto: I have to do something similar with managedstruct 19:37
cotto whiteknight, ping
Kapace_ null $P0, sweep 1 is sure to call destroy(), but I agree that it should happen automatically
whiteknight cotto: pong
cotto whiteknight, do you know why ^ is necessary?
whiteknight cotto: what is necessary?
cotto Kapace_> null $P0, sweep 1 is sure to call destroy(), but I agree that it should happen automatically 19:38
whiteknight I'm not sure I understand what that means
cotto Apparently that code is needed to convince the gc to call VTABLE_destroy on a PMC for some tests.
19:39 km0r3 joined
pjcj dukeleto: Somehow 1bb6425 is attributed to me, but it's really from wesjdj - he was using my machine, but on his login, so I'm not exactly sure how the checkin was attributed to me. 19:42
19:42 kennym left
pjcj in any case, neither cover (quickcover) nor the fullcover (cover) targets seem to be working properly anyway - we'll investigate further later 19:44
off to rutsch now ...
19:58 km0r3 left
dukeleto pjcj: that sounds odd 20:00
Coke pjcj: guessing you have a global .gitconfig. 20:05
20:10 fbrito joined, rfw joined 20:17 fbrito left
Kapace_ should I continue trying to set a custom destroy function or give up at 28.2% cover increase? 20:18
cotto Kapace_, which pmc? 20:20
Matt_ whiteknight: Can you close the task: www.google-melange.com/gci/task/sho...4833#c2002
Kapace_ cotto: managedstruct
20:20 whiteknight left
Kapace_ Matt_: don't you hate it when that happens, lol 20:21
cotto I'm working through any gci tasks in need of review now. 20:25
Kapace_ I can submit with 28.2% increase..
cotto msg whiteknight Did the task to switch cover and quickcover get merged? 20:30
aloha OK. I'll deliver the message.
20:35 davidfetter left 20:49 plobsing joined 20:53 plobsing_ left
Matt_ Is it safe to remove the dead VTABLES mentioned here: trac.parrot.org/parrot/ticket/1914 ? Its the only way of increasing the code coverage for the Capture PMC. 20:59
Kapace_ Matt_: you've been able to build rakudo?
for some reason I can't, so I can't test rakudo built without those functions 21:00
Matt_ Kapace_: Ill try now
Kapace_ are there any other HLLs that should be tested and reported to the ticket? 21:01
Coke lua is a good one to test. 21:03
partcl used to be. 21:04
Kapace_ :( lua build fails in the same place as my previous rakudo build attempt 21:08
21:08 lucian joined 21:09 lucian left 21:10 rfw left
plobsing msg whiteknight: (re: encapsulate_main) I'm mopping up some examples tests. after that, nqp-rx needs to be updated. I've tried a proper fix a couple of times, with no success. I'll put it to the list. 21:11
aloha OK. I'll deliver the message.
Matt_ Kapace_: my build is still going 21:13
Kapace_ yeah, takes awhile eh? 21:14
21:16 rfw joined
Matt_ Kapace_: no errors here 21:20
Which functions are we checking for removal?
Kapace_ Matt_: they're all listed in the ticket, and probably defined/exists/etc_keyed_str 21:21
Matt_ yeah I noticed all of those are 'dead' too 21:22
Coerage could easily go up to the 90% if this is all fixed
*90's
Kapace_ yeah, can I have a task too plz? :P
Matt_ There's like 3 for the Capture PMC, enough for it to go around :) 21:23
Kapace_ ok 21:24
mikehh btw - t/src/embed.t - TODO passed: 3 in --optimize builds, not in builds without --optimize
I think it is probably down to an ASSERT or something like that (optimized builds do not enforce ASSERTS) 21:28
Matt_ Kapace_: are we just checking if rakudo builds or doing testing on it too?
21:28 M_o_C left 21:29 chromatic joined
Kapace_ Matt_: I would do the tests too 21:29
cotto that's a good idea
chromatic GCI idea: write many small benchmarks which isolate NQP-rx features.
Matt_ kk building rakudo with the new parrot, then `make test` 21:30
chromatic has the impression that careful and precise benchmarking could lead us to 30-50% NQP-rx performance improvements.
plobsing it is slow enough to seem reasonable 21:33
cotto chromatic, are you thinking of using something similar to the existing test infrastructure wrapped in a timer? 21:34
Coke chromatic: hey. 21:35
cotto if that's the case, we could write an example and create a batch of tasks based on that. 21:36
chromatic We need to test individual rules and features in isolation. If the easiest way to do so is repurposing existing tests, let's do that.
Coke, I saw some benchmark figures from you. fill_params at 9% seems reasonable. 21:37
21:37 rurban_ joined
Matt_ cotto: Should we test anything other than Rakudo. build went fine after commenting out the VTABLE's. Testing is underway... 21:37
chromatic We could make fill_params faster by removing the indirection necessary to handle both va_list and Context parameters/arguments, but that's a bigger change.
Matt_ cotto: Tests passed: cl.ly/1x2F1z072u3T3s1s1k2V 21:38
\\cc Kapace_
Kapace_ Matt_: lua if you can
plobsing boo va_list. why won't it die?
Matt_ how exactly do you get Lua to run on Parrot? 21:39
Kapace_ ask google or www.ohloh.net/p/lua-on-parrot 21:40
plobsing github.com/fperrad/lua.git
Kapace_ maybe this link is better: github.com/fperrad/lua
chromatic va_list exists for Parrot_ext_call(), or whatever it is now.
Matt_ kk. I thought it was more officially supported
21:40 rurban left, rurban_ is now known as rurban
Coke chromatic: is there a way with kcachegrind to see what is taking time /in/ a function? or is just (as I expect) function-level info. 21:41
plobsing Matt_: it has much core developer support because it is an HLL that tends to keep working for some reason
Coke (or some other cachegrind tool)
chromatic Coke, if you've built with -g you can.
perl Configure.pl --maintainer --optimize="-O3 -g" --ccflags="-march=native" --prefix=/home/chromatic/dev/tmp/parrot
Matt_ Kapace_: how did you try building Lua 21:42
Coke what does arch=native do?
Kapace_ Matt_: ./setup.pir, once you've cloned it
Matt_ kk
plobsing Matt_: you'll need to have parrot *installed*
and in your PATH 21:43
Kapace_ OK, this is really strange:
Matt_ ok. Was just gonna ask if you can specify where Parrot is installed
Kapace_ I have a fresh clone of rakudo, and i did perl Configure.pl --gen-parrot
which should get a clean copy of parrot, and use it
Coke that installs it in the ./parrot-mumble-something.
Kapace_ but I still get the same build errors
21:44 kennym joined
Coke Kapace_: if you already have a copy of parrot in place, it's not as clean as you might think. 21:44
try removing the parrot* directories and trying again.
Kapace_ from /usr/local etc?
chromatic arch=native uses the best possible optimizations for your processor.
Coke chromatic: ... seems like that should be a default. 21:45
danke.
I regularly build with --optimize --ccflags="-g", which ends up being -O3 -g, so I'm halfway there.
Matt_ Kapace_: I didnt use --gen-parrot
I specified the parrot_config file instaed 21:46
its in the README in the root of the git repo
Kapace_ ok ill give that a shot
Coke Kapace_: If you are using --gen-parrot, the source is checked out into the local directory and installed to another local directory.
sorear Coke: when compiling on a 386 compatible, gcc defaults to only using 386 instructions, since you might want to give the binary to someone else
Coke but if you do it multiple times, the checked out copy is not thoroughly cleaned before rebuilding.
sorear Coke: -march=native means "I don't plan to publish the binary or downgrade my CPU after compiling"
Matt_ Kk. the test for lua failed. Not sure if its related 21:47
chromatic -march=native also means "You know, you might be onto something with this SSE thing."
sorear Coke: it's also possible to explicitly specifiy a minimum CPU, for instance -march=pentium2
Matt_ "Parse errors: Unknown TAP token: "sh: luac: not found"" 21:48
Kapace_ doesn't sound related, rebuild with the functions uncommented or whatever to make sure, I guess
plobsing Matt_: you need "real" lua installed for testing for contrast-and-compare purposes 21:49
Coke native isn't valid on my platform.
ah well.
21:49 contingencyplan joined
sorear Coke: powermac? 21:49
chromatic Darwin, likely.
Coke x86 mac.
bah, nopaste is still down. 21:50
Permissible names are: armv2, armv2a, armv3, armv3m, armv4, armv4t, armv5, armv5t, armv5te, armv6, armv6j, iwmmxt, ep9312.
chromatic The -g is the important part of what I posted anyhow.
Coke yes, but faster is better.
-mtune=native looks like it might be how to spell that locally. 21:51
sorear no, -mtune is something else
what version of gcc do you have?
Coke 4.2.1
... the docs seem to indicate it's what chromatic said, actually. 21:52
chromatic -march implies -mtune, no? 21:55
Matt_ Kapace_: same results. I guess its not being used? That one test failure seems unrelated.. 21:58
Kapace_ ok Matt_, report your findings to the ticket, maybe that will convince them to remove the functions :) 22:00
Matt_ done 22:03
22:05 Yuki`N joined
dalek rrot: 2fb6564 | kapace++ | t/pmc/managedstruct.t:
Got managedstruct coverage up to 88.2
 22:14
rrot: e4b3bac | cotto++ | t/pmc/managedstruct.t:
[t] whitespace fix
cotto Kapace_, accepted. thanks 22:15
mikehh rakudo (f99d081) - builds on parrot (2_11_0-428-g9596abd) - make test PASS, make spectest_smolder[(#1941), roast (ee3b704)] FAIL - Ubuntu 10.10 i386 (g++-4.5 with --optimize)
t/spec/integration/advent2009-day18.rakudo - Parse errors: Bad plan. You planned 6 tests but ran 0.
27,067 ok, 6 failed, 614 todo, 1,855 skipped and 0 unexpectedly succeeded
Kapace_ cotto: thanks, I really have to kick that bad habit of whitespace
Yuki`N Oh cotto, can you approve socghop.appspot.com/gci/task/show/g...9360981098
Kapace_, what os are you running?
Kapace_ Yuki`N: Ubuntu 10.10 22:16
Yuki`N Using gedit?
Kapace_ I'm using gedit
mikehh Kapace_: you need to run make codetest - it will pick up any coding standard problems
Kapace_ yeah :/
Yuki`N github.com/dinkel/gedit-whitespace-remover
Kapace_ oh that might be useful!
Yuki`N Removes trailing whitespace every time you save.
Kapace_ cool
mikehh I use Kate and have that setting there 22:17
i.e. remove trailing whitespace on save
Matt_ cotto: trac.parrot.org/parrot/ticket/1914#comment:2 22:18
can we proceed with the removal?
cotto Matt_, sure. Just do it in a branch. 22:21
Matt_ cotto: Here is the new branch: github.com/mattrajca/parrot/commit...ad_vtables 22:28
not sure if you want a pull request yet
cotto no need 22:29
22:29 whiteknight joined
Matt_ www.google-melange.com/gci/task/sho...0454733#c1 22:29
Here is the task 22:30
this brings up the coverage to 88% from 77%
There are other dead VTABLE's which Kapace_ will investigate
(coverage should go up to the high 90's once they're all fixed / removed)
dalek rrot/gci_capture_cleanup: 2816668 | (Matt Rajca)++ | NEWS:
Merge branch 'master' of git://github.com/parrot/parrot
 22:33
rrot/gci_capture_cleanup: a4750d8 | cotto++ | src/packfile/api.c:
Merge branch 'master' of github.com/mattrajca/parrot into gci_capture_cleanup
rrot: b3df151 | Whiteknight++ | t/pmc/managedstruct.t:
Merge branch 'master' of github.com:parrot/parrot
rrot: 5b39b44 | Whiteknight++ | t/pmc/task.t:
Merge remote branch 'origin/gci_task_coverage'
 22:34
rrot: 2662ff6 | cotto++ | t/pmc/task.t:
[t] typo fix in test name
 22:38
cotto task requests accepted 22:40
Matt_ work submitted: www.google-melange.com/gci/task/sho...9360454733 22:41
cotto done 22:43
rfw what was the thing about the object pmc's share_ro again? 22:48
Kapace_ rfw: burn it with fire?
rfw oh
bleh so what do i do with this object pmc coverage thing 22:49
22:49 kennym left
rfw cotto: ping 22:50
whiteknight 119 closed tasks, and 10 days left to go 22:51
you GCI students are kicking ass
rfw or whiteknight
what do i do about share_ro
whiteknight rfw: BALEETED
rfw oh
hm it's still in master 22:52
whiteknight kill it
with a shovel
rfw do i get points for it?
lolol
rippan out
cotto rfw, pooooonnnnnngggggg 22:54
rfw cotto: too late :D
cotto oh noes 22:55
dalek rrot: 15df436 | cotto++ | / (3 files):
Merge branch 'gci_make_targetnames'
cotto whiteknight, are you making sure to delete branches as the get merged? 22:56
pjcj Coke: ah yes, that's probably it - thanks 22:58
Yuki`N rfw, no it's all ripped out in a branch 23:00
:P
rfw Yuki`N: fjdlsag
Yuki`N I should test building rakudo with that branch actually.
rfw wanted to do that task too :(
Yuki`N Upping my Fedora VM to 4 CPUs
In preparation for super-fast parallel building. 23:01
23:01 kennym joined
rfw # Method 'get_user_id' not found for invocant of class 'OS' 23:06
lolwhat
heh but it's right there
cotto rfw, pmc2c drop stuff on the floor if it isn't formatted just right. 23:08
dukeleto, ping
chromatic whiteknight: GCI idea: write many small benchmarks which isolate NQP-rx features. 23:09
rfw yeah it looks like get_user_id is being dropped 23:11
cotto rfw, is this in master?
rfw no
the get_user_id branch
heh so is can_execute 23:12
cotto looking into it now 23:13
rfw i think it's because can_execute isn't doc'd properly
aha fixed 23:14
# ./parrot: symbol lookup error: /root/parrot/runtime/parrot/dynext/os.so: undefined symbol: Parrot_get_user_id 23:15
whiteknight: plz2fix :(
cotto rfw, can you push your tests? 23:17
it looks fine to me
rfw cotto: haven't written any yet
cotto ah
rfw well
haven't finished writing any
but i'm getting undefined symbol when i call get_user_id
pushing 23:19
Yuki`N cotto, rakudo builds with my branch remove_share_ro
Running the spec tests now.
cotto Yuki`N, great 23:20
I'm not going to miss that.
rfw cotto: github.com/rofflwaffls/parrot/comm...ba068a72be 23:22
Yuki`N I'm using.
86% of my RAM.
I've never used that much before.
cotto you're welcome
Yuki`N Maybe it's because I'm running 2VMs, and one of them is doing some intense git cloning and the other is running spectests. 23:23
Which apparently don't get affected by -j4
rfw my computer likes to die with one vm open
granted, the vm runs with 4 cpus
er, s/cpu/core/
Yuki`N Lol.
8 virtual cores, 6GB ram, hell yes. 23:24
I want another 6GB actually...
rfw i only have 4 cores and 4gb
heh
Yuki`N I'm apparently using 5.1GB ram now.
cotto That's cute. Apparently Parrot_get_user_id given a prototype but never a definition. 23:26
rfw cotto: it is
hold on
./config/gen/platform/generic/uid.c:Parrot_get_user_id()
i think it's generated at runtime
errr
compile time
Yuki`N Duubious. 23:27
rfw yeah
cotto ah. probably need to reconfigure
23:29 whiteknight left
cotto may have a fix 23:30
there you go 23:31
rfw where do i go
tadzik happy new year parrots!
cotto gist.github.com/761425
sorry. nopaste was down
rfw oh perl has a getpw* thing 23:33
why theh ell was i using id -u
oop, forgot to realclean 23:34
hm cotto, still broken 23:35
cotto rfw, after reconfig?
rfw yep 23:36
after realclean and Configure.pl
cotto which is the same as reconfig
odd
rfw i see it in platform.c 23:37
so in theory it should be fine
but it's not being linked into libparrot.so 23:40
oh i lie 23:41
wait, i think i know why
cotto what's your theory? 23:45
rfw well, i thought it was because i had a system libparrot.so but after removing it it's still broken
Yuki`N lol 23:46
rfw cotto: does it work on yours?
cotto nope
Yuki`N No tests have failed so far. 23:47
A few dubious results, with all subtests passing.
cotto Ah. Because it's using libparrot's exported symbols, it needs PARROT_EXPORT in Parrot_get_user_id. 23:49
well, on it 23:50
rfw let's try that 23:51
cotto That raises the question of whether we want to expose those functions though libparrot.
It fails another test with, but it can find the symbol.
23:51 kid51 joined
cotto hio kid51 23:51
rfw t/dynpmc/os.t .. ok
All tests successful.
yay cotto, it works
thanks 23:52
cotto np, though we shouldn't have only that one function marked PARROT_EXPORT.
I 23:53
I'll open an rfc ticket. I don't think there's much harm in exporting those functions, but it's good to see if anyone else has other ideas. 23:54

Powered by Cro and the Raku® Programming Language.

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

Thank you!