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.
00:00 nwellnhof joined
nwellnhof Matt: you should be able to generate coverage reports with the changes in branch nwellnhof/make_cover. 00:01
Matt_ nwellnhof: Thanks!
dalek rrot/encapsulate-main: 2b0abde | plobsing++ | runtime/parrot/library/PGE/Perl6Grammar.pir:
mark PGE main subs :main
 00:04
rrot/encapsulate-main: 5739ce1 | plobsing++ | lib/Parrot/Test/PGE.pm:
mark pge testing PIR with :main
Matt_ whiteknight: What about functions like Parrot_api_load_language ? 00:05
I think coverage will already increase +10% after commenting out Parrot_api_set_stdhandles and adding tests for get/set_compiler but I want to do a few more 00:06
nwellnhof i'm about to merge nwellnhof/string_unescape now... 00:08
rfw whiteknight: did it finish yet? 00:11
Coke during a bisect, should git st be saying I'm not on a branch? 00:15
dalek rrot: 2ceb0e8 | nwellnhof++ | / (19 files):
Merge branch 'master' into nwellnhof/string_unescape

Conflicts:
  \tDEPRECATED.pod		 00:16
rrot: b8b3c7d | nwellnhof++ | / (16 files):
Merge branch 'nwellnhof/string_unescape'
whiteknight rfw: I'm on a resource-constrained VM for my testing. It moves *very* slowly 00:17
rfw oh
heh
i make -j5 in my vm
dalek TT #1809 closed by nwellnhof++: Remaining string_* functions are deprecated 00:20
TT #1809: trac.parrot.org/parrot/ticket/1809
Matt_ whiteknight: Is there any good way of testing Parrot_api_add_library_search_path ? 00:24
whiteknight good question 00:25
dukeleto tadzik: i was away. did you make progress on that issue? 00:29
rfw whiteknight: in theory my misc.t test should cover the remainder of misc.c shouldn't it 00:30
tadzik dukeleto: nah
that seems just weird to me, I don't know where to look
dukeleto tadzik: what is the current problem? i lost it in the backscroll 00:31
whiteknight rfw, okay it's done. looking at results now
...no results 00:32
wtf
tadzik dukeleto:
whiteknight what's the name of the file it's supposed to generate? coverage.html? 00:33
tadzik wklej.org/id/445749/
repl works, running from file doesn't
dalek TT #1628 closed by nwellnhof++: Unescape parrot string to parrot string 00:36
TT #1628: trac.parrot.org/parrot/ticket/1628
whiteknight I won't make any more coverage-related GCI tasks if it's going to be such a hassle 00:41
dalek rrot: 20ddb40 | mikehh++ | examples/pir/life.pir:
replace concat_s_s/concat_s_sc opcode and tabs
rrot: ee1e105 | mikehh++ | examples/pir/readline.pir:
replace concat_s_s opcode and tabs
rfw whiteknight: should be 00:42
i ended up hacking around with lcov and genhtml because make cover did silly things 00:43
Yuki`N Some coverage tasks are easier than others, honestly.
The hardest part is really setting up the environment for `make cover' to run.
whiteknight yeah 00:44
Yuki`N I didn't have any trouble personally, but then again I also just installed packages from my distro's repos.
rfw Yuki`N: can you make cover current master's HEAD
Yuki`N Um, sure.
Let me boot up the vm.
rfw and check if any coverage reports are being generated for src/pmc
Yuki`N If you don't get any, in my experience.
nwellnhof rfw: i fixed a a lot of problems in my make_cover branch.
rfw nwellnhof: ah
Yuki`N You need to check to make sure that all tests are passing
And a make distclean sometimes helps. 00:45
whiteknight nwellnhof: I CAN HAZ MERGE NAO PLZ?
PLZKTHXBYE
rfw 4 points: reimplement parrot in lolcode
whiteknight IM IN UR BYTECODEZ INTERPRETIN UR BYTES 00:46
tadzik there was a PAST-to-Lolcode compiler once
whiteknight wtf, does make quickcover cause a complete reconfigure/rebuild? 00:47
I need some goddamned alcohol. gcov is angrying up my blood
Yuki`N There's already a HLL for lolcode on parrot, is there not?
nwellnhof whiteknight: it has to add some coverage options to gcc.
ok, i merge make_cover then... 00:48
rfw let's merge nwellnhof/make_cover into my fork
whiteknight rfw: once he merges, just do a pull on yours 00:49
Yuki`N Do you still need coverage information generated?
rfw whiteknight: too late \\o/
rfw watches parrot compile 00:50
whiteknight :)
Yuki`N Hmm, make distclean doesn't clean out t/src/embed/
Let me fix.
whiteknight Yuki`N++
dalek rrot: c11e91b | nwellnhof++ | config/ (3 files):
Merge branch 'nwellnhof/make_cover'
Tene Yuki`N: Yes, there is lolcode on parrot, but it has not been maintained for quite a while. I originally wrote it. 00:51
rfw CAN HAZ Tene
Tene rfw: No, you can not haz tene. There's only one tene, and I'm currently using it. 00:52
rfw :<
tadzik it works,$key
rfw okay now i'm getting 50% coverage
wtf
tadzik bah, suprise paste
Tene rfw: I'll be done with my tene around midnight tonight. I won't need it again until 9:00 AM, so feel free to stop by and borrow it, if you like. 00:53
rfw Tene: so what is a Tene anyway
Tene rfw: Me.
whiteknight blah. make cover fails here. It looks like all the file paths it is looking up are wrong 00:55
can't find file src/pmc/src/pmc/continuation.pmc
(and so on for all source files)
Kapace_ if you need it to work, just get a ubuntu 9.10 vm :P 00:56
whiteknight Kapace_: I'm on Ubuntu 10.10
Kapace_ same
rfw Tene: so how do i borrow your you
nwellnhof whiteknight: are you testing with my changes? 00:57
whiteknight nwellnhof: I'm pulling your changes now
nwellnhof it should fix the src/pmc/src/pmc/... problem. this was caused by changes in Devel::Cover. 00:58
Tene rfw: I was being silly, and suggesting that I didn't need myself while I was sleeping, and some other nonsense.
rfw well now you just ruined the joke!
Tene rfw: I've been very busy lately, but if you want help with lolcode on Parrot, feel free to harass me about it this weekend. 00:59
rfw hah i'm still _very_ new to the wrold of parrot
dukeleto Tene.new.clone.fork
whiteknight rfw: We're all new to it
Tene Every time I've made an offer of help to people here in the past several months, I haven't actually followed through, though, so I wouldn't be very hopeful.
dukeleto rfw: you are doing quite well for such a newbie ;)
rfw dukeleto: thanks
dukeleto rfw: you are rocking GCI, good luck! 01:00
rfw dukeleto: *almost*
01:00 fbrito left
rfw okay, coverage is completely ignoring my tests 01:01
Kapace_ rfw: hey, when you got to 100 points, did you unlock any cheat codes or something?
rfw Kapace_: nope
i tried up up down down left right left right b a though
Kapace_ darn..
lol
rfw FINALLY 01:03
YES
I HAVE NO IDEA WHAT I DID BUT IT WORKED
Yuki`N github.com/parrot/parrot/pull/50 01:05
Those are my modifications.
dalek rrot: 3cb17c1 | mikehh++ | examples/streams/Combiner.pir:
replace concat_s_sc opcode
rrot: 128eb76 | mikehh++ | examples/streams/Filter.pir:
replace concat_s_sc opcode
Yuki`N With the changes that nwellnhof made it looks like they won't merge very well.
Removing QUICKCOVER_DIRS also affects the clean targets and it doesn't look like you changed those. 01:06
I'll see if I can perhaps merge and then fix that.
dukeleto rfw: nice contra reference :) 01:07
whiteknight nwellnhof: with your changes I get the same problem
it can't find any of the source files
rfw dukeleto: sadly i suck at it
whiteknight no source src/pmc/src/pmc/bigint.pmc, etc 01:08
nwellnhof whiteknight: have you tried make realclean?
dukeleto rfw: that reminds me of my youth, playing atari and NES games
whiteknight nwellnhof: I thought so. Let me retry
nwellnhof make sure that there are no old .gcov files around. 01:09
rfw dukeleto: i enjoy playing retro games myself
Yuki`N whiteknight, take a look at my pull request. I have some clean-related makefile changes. 01:10
rfw okay my code coverage just went from 81.5% to 31%...
Yuki`N loooooooool.
whiteknight Yuki`N: okay, I'll do it in a little bit
rfw i swear parrot's cover build system is broken
Yuki`N It's a bit of black magic. :P 01:11
I think the real problem is the tremendous amount of output hides any problems that may exist.
i.e. failed todo tests print out in their entirety.
dalek rrot: d6c54c5 | nwellnhof++ | config/gen/makefiles/root.in:
Remove reference to QUICKCOVER_DIRS
 01:12
mikehh 4 more failures to fix in t/examples/catchall.t, but I will do that later
Yuki`N nwellnhof, aww you stole my commit.
mikehh needs sleep
Yuki`N nwellnhof, perhaps you can take care of my pull request then, it basically did that + fixed the fact that t/src/embed wasn't getting cleaned.
rfw, if you give me a copy of the test you wrote I can try it myself. 01:14
Is it on github?
nwellnhof Yuki`N: the only remaining commit is github.com/nol888/parrot/commit/26...4f234a4fdd ?
rfw Yuki`N: hold on my making -j5 is hanging everything else
Yuki`N nwellnhof, yes. 01:15
rfw and make doesn't want to die
Yuki`N Lol I only gave my VM 2 cpu's.
The i7 is a beast for virtualization. 01:16
rfw i gave mine 4 but i set processor affinities :\\
too bad it doesn't seem to care
Yuki`N Lol.
rfw oh tomoyo is the new selinux 01:17
Yuki`N Really?
rfw seems like it
developed by ntt (lol japan) 01:18
Yuki`N Well Tomoyo is pretty tsundere.
rfw okay my vm is conclusively broken 01:20
dukeleto tomoyo.sourceforge.jp/\\
blarg! tomoyo.sourceforge.jp/
Yuki`N nwellnhof, you broke my make cover. :( 01:21
I have 5 files. 01:22
and lots of gcov files like /.#src#pmc#unmanagedstruct.pmc.gcov 01:23
I'm even more confused now.
01:23 nwellnhof left
Yuki`N Hmmm. 01:24
Why is my make cover not working.
rfw argh
38.9% coverage
why the hell...
oh, i think my parallelized make is breaking it 01:27
Yuki`N I give up. 01:29
Something with make cover is now broken again.
Seems to be a gcc version barrier or something.
01:36 nwellnhof joined
rfw whiteknight: YES I WIN 01:38
File:\tsrc/misc.c
Coverage:\t100.0%
nwellnhof lost my internet connection... 01:39
rfw nwellnhof: make cover doesn't play nice with -jx
er
-jX*
hell, it tries to run tests before configure 01:40
dukeleto, whiteknight or whoever's there: www.google-melange.com/gci/task/sho...9341582405 01:41
nwellnhof rfw: i'll look into that.
Yuki`N nwellnhof, your fixes break my make cover.
I get a lot of ".#src#...gcno" files in my root 01:42
And only 5 files appear in the coverage report.
I rebased without your commit and trying make cover just in case.
nwellnhof Yuki`N: are there any error messages at the end of make quickcover? 01:43
Yuki`N Nope.
nwellnhof Yuki`N: are you running make -j? 01:45
Yuki`N nope.
nwellnhof Yuki`N: what's the output of make quickcover after running the tests? 01:48
01:50 kid51 joined
dalek rrot: 12e7238 | nwellnhof++ | config/gen/makefiles/root.in:
Fix make cover dependencies
 02:01
dukeleto has weighed in on the great gci-discuss jumpyshoes controversy 02:02
Yuki`N nwellnhof, it's normal 02:05
dukeleto, can you add me another task for increasing src/embed/pmc.c?
nopaste "kid51" at 192.168.1.3 pasted "make quickcover: broken at d6c54c50e335" (2899 lines) at nopaste.snit.ch/27513
Yuki`N I got it up to 95%, it was about 83% prior.
The relevant test additions are already in master.
rfw dukeleto: could you approve this: www.google-melange.com/gci/task/sho...9341582405
Yuki`N nwellnhof, kid51 has the exact same problem that I have apparently. 02:06
After rebasing out your commit, cover and quickcover work again, so it's definitely something that your commit broke.
rfw Yuki`N: i wonder why it works here :| 02:07
Yuki`N rfw, what's your gcc version?
rfw gcc (GCC) 4.5.1 20101125 (prerelease) 02:08
what's yours
Yuki`N 4.5.1 as well.
rfw odd
what's your gcov version
Yuki`N gcov is part of gcc.
rfw oh wait 02:09
duh
yeah
i'm silly
Yuki`N You can do make cover now? Weren't you using some hacked-together system before?
whiteknight dukeleto: what was the jumpyshoes controversy?
rfw whiteknight: groups.google.com/d/topic/gci-disc...discussion (yes i posted it) 02:10
whiteknight: also, unrelated: www.google-melange.com/gci/task/sho...9341582405 02:11
whiteknight I've looked at jumpyshoes before in the standings. It is *weird* how he is racking up so many points like that 02:12
nwellnhof: I made realclean and make quickcover again and this time it works. I must not have realcleaned before as I thought
rfw: I don't see any comments from dukeleto in there 02:16
rfw whiteknight: it has to be reviewed first
02:16 kennym left
whiteknight oh, okay 02:16
rfw could you look at my task :D
(and help me defeat the evil forces of jumpyshoes)
whiteknight rfw: I can't pull tonight, but I will approve the task 02:17
rfw ah okay 02:18
whiteknight I'll pull it tomorrow
rfw it should work (provided make cover isn't breaking)
02:18 nwellnhof left
whiteknight I trust you. Your work has been high-quality so far 02:18
but now I have to go to bed. Goodnight
rfw heh, thanks
Yuki`N Night.
rfw night
Yuki`N Hmm, waiting on some dfBSD mentors. 02:19
02:19 whiteknight left
kid51 'make quickcover' is still broken for me at commit 12e72382f0f (in much the same way as previous paste) 02:19
02:22 nwellnhof joined 02:23 AndChat| joined
nopaste "kid51" at 192.168.1.3 pasted "make quickover still broken at 12e723" (5612 lines) at nopaste.snit.ch/27514 02:23
dalek rrot: 401fcfb | nwellnhof++ | config/gen/makefiles/root.in:
Remove gcov -p option

This should fix make cover for older versions of Devel::Cover. But some gcov files will now be overwritten.
nwellnhof kid51: can you try 401fcfb541?
kid51 Among other problems with 'make quickcover', I'm getting files named like this in my top-level dir: compilers#imcc#debug.c.gcov
nwellnhof: Are you yourself trying these out before you commit? 02:24
nwellnhof kid51: that was expected. it's caused by the gcov -p option.
kid51 Even with 'quickcover', the full cycle after a new commit takes most of 15 minutes. 02:25
02:26 Matt_ left, gg411 left
kid51 make quickcover: still broken at 401fcfb541 02:36
nwellnhof: Would it be possible to back out your changes and rework them in a branch? 02:37
02:41 Matt_ joined
nwellnhof kid51: yes, we can do that 02:45
02:46 AndChat| left
kid51 nwellnhof: Thanks. 02:51
02:51 nwellnhof left 03:01 mtk left 03:03 Matt_ left
dalek rrot: 7188dd4 | nwellnhof++ | config/ (3 files):
Revert make cover changes
 03:09
03:13 kid51 left 03:23 nwellnhof joined
rfw wait what exactly is a roflbbq 03:27
sorear fnord. you are not cleard for that. 03:33
03:51 fbrito joined
davidfetter rfw, it's what you have after a long roflcopter ride 04:01
rfw will there be lmaonaise? 04:02
davidfetter of course :)
dalek TT #1907 created by coke++: Behavior of "exit" opcode changed since 2.11.0 04:26
TT #1907: trac.parrot.org/parrot/ticket/1907
04:44 Coke left 05:07 Coke joined 05:16 estrabd left 05:18 nwellnhof_ joined 05:22 nwellnhof left, nwellnhof_ is now known as nwellnhof 05:28 cotto joined
cotto ~~ 05:31
05:33 dd070 joined 05:35 rurban_ joined, dd070 left, Myhrlin left 05:37 rurban left, rurban_ is now known as rurban 05:41 nwellnhof left, Myhrlin joined
dalek rrot: 6dd7f77 | Coke++ | / (2 files):
check dependencies on new frontend/ TLD too

  (and make them all pass)		 05:43
rrot: 8f1f70d | Coke++ | t/src/checkdepend.t:
Don't skip directory that no longer exists.
rrot: 34dea34 | Coke++ | / (2 files):
Check deps for src/string/encoding

  (and make them all pass)
rrot: 9b71f41 | nwellnhof++ | / (3 files):
Experimental make target quickcover_new

New approach to fix make cover on different platforms. The script tools/dev/make_cover.pl is called from make to generate the report.		 05:46
Coke if I want to create a new upstream branch, ??
just start with a -b and then push it? 05:47
cotto dukeleto, ping 05:51
coke++ for dependencies cleanup work 06:00
GeJ Merry belated Christmas #parroteers.
cotto Good tidings to you, to you and your crazy machine. 06:01
GeJ Thanks cotto. 06:03
Coke BWAHAHAHAH 06:05
(apparently init::headers lets me rip out more than I anticipated) 06:06
dalek rrot/checkdepend_pmcs: cd5cbdd | Coke++ | / (14 files):
remove init::headers config step

hardcode include directory here instead of config step. Removing NONGEN headers entirely still passes checkdepend.t and the build works at a high -j.		 06:13
fbrito can anyone please accept my claim request on this task? :) www.google-melange.com/gci/task/sho...9228630126 06:20
rfw cotto: www.google-melange.com/gci/task/sho...9315789586 <-- this task is invalid: Configure.pl already has --silent implemented 06:24
cotto onde 06:25
rfw at least
i think it is
cotto er, fbrito's request was accepted
rfw, ping kid51 about that.
rfw ah okay
cotto I have some concerns about another of his tasks. 06:26
Yuki`N rfw, yeah I already found that.
rfw msg kid51 trac.parrot.org/parrot/ticket/1897 <-- running Configure.pl in --silent mode does, in fact, make it run silenty
aloha OK. I'll deliver the message.
rfw Yuki`N: did you tell him?
i think i'll do Title: Research embedding Parrot into Internet Explorer 8 06:27
Yuki`N rfw, I brought it up a while back but I don't think anyone paid attention ot it.
rfw Yuki`N: ah okay
cotto: www.google-melange.com/gci/task/sho...9052493089 <-- could you approve
fbrito cotto thank you :). I have also opened a pull request with my work on it 06:28
cotto but don't worry reviewing it now... I am going to bed 06:29
cotto rfw, done
rfw thanks
cotto have fun and don't use wikipedia as your primary resource
fbrito rfw: nice! a research task :)
rfw i won't :D
cotto As a starting point it's great. Just be sure to dig deeper.
GeJ Huuuum, tramadol! 06:37
dukeleto rfw: i am excited to see what you come up with :) 06:42
rfw dukeleto: well, i've already found IActiveScript
cotto looks like the parallel build is much better now. coke++ 06:43
fbrito i am going to bed. good night guys 06:46
Yuki`N dukeleto, can you add me another task for increasing src/embed/pmc.c?
I got it up to 95%, it was about 83% prior.
06:46 fbrito left
dukeleto Yuki`N: you have a link for the previous task? 06:53
06:53 Yuki`N left, Yuki`N joined
Yuki`N socghop.appspot.com/gci/task/show/g...9336706640 06:54
dukeleto Yuki`N: socghop.appspot.com/gci/task/edit/g...968685?s=0 07:02
Yuki`N Requested claim. 07:03
cotto Yuki`N, accepted 07:05
Yuki`N Submitted work.
cotto kid speedy 07:06
dukeleto Yuki`N: do you want to make that a pull request against the proper branch? 07:07
Yuki`N: then i will approve the task :)
Yuki`N: or is it already committed?
rfw cotto: pastebin.com/RwMH6exx <-- do i need any more content?
dukeleto Yuki`N: i see that is in your fork
Yuki`N It's already committed to master.
dukeleto Yuki`N: gotcha 07:08
Yuki`N: things move fast these days
Yuki`N Yeah. :P GCI.
Jumpyshoes is batching his tasks it seems.
dukeleto Yuki`N: i saw your very carefully-worded email to gci-discuss :) 07:09
Yuki`N Yes, indeed.
rfw i think i was being a bit too blunt with mine 07:10
cotto Yuki`N, done
dukeleto Yuki`N: task completed. go get another one :)
cotto dukeleto beat me 07:11
clever gir^H^H^Hduke
dukeleto cotto: yes, i am quick these days 07:12
Yuki`N :P
07:12 Myhrlin left
dukeleto msg fbrito bigint gci task completed 07:13
aloha OK. I'll deliver the message.
Coke cotto: better? 07:14
dukeleto parrot.git has 5 pull requets right now
Coke perhaps the encoding stuff was wonky.
cotto Coke, the parallel build seems much more reliable 07:15
dukeleto cotto: did you apply the git hook stuff already? can that pull request be closed?
cotto I just ran make -J9 && make reconfigure 16 times and it didn't fail
dukeleto, yes
I though I closed that 07:16
dukeleto cotto: github.com/parrot/parrot/pull/43
cotto dukeleto, is there a nice way to review all commits to a github repo?
rfw cotto: pastebin.com/BS2Rkg6r revised abit
cotto closed
Coke cotto: it's been failing at -j9 ? 07:17
cotto rfw, you could use a gist
Coke, occasionally
Coke if it ever fails, open a ticket so I can fix it. ;)
rfw oh
pastebin due to force of habit
cotto gists are nice for things that get updated frequently
just my preference, though
rfw gist.github.com/757003 there we go
yeah i would use gist more if gist.github.com was shorter than pastebin.com 07:18
cotto did you find anything about projects that have tried something similar? 07:19
dukeleto, are you looking through any of the pull requests? 07:20
rfw hm i looked at activestate's perlscript and pythonscript
i mentioned it somewhere near the top
cotto so you did
rfw personally i think using an activex object is better than getting into wsh's activescript 07:21
cotto perhaps I should review when my brain is more functional
rfw heh okay
dalek rrot: bad917c | fbrito++ | t/pmc/bigint.t:
[t] Add tests to BigInt
 07:40
rrot: 17c2706 | fbrito++ | t/pmc/bigint.t:
[t] Add test to decrease in BigInt
rrot: 400caf1 | fbrito++ | t/pmc/bigint.t:
[t] Add 'add' and 'i_add' tests in BigInt
rrot: 0502e9b | fbrito++ | t/pmc/bigint.t:
[t] Add 'multiply' and 'i_multiply' tests in BigInt
rrot: 2de15fa | fbrito++ | t/pmc/bigint.t:
[t] Add 'mod' and 'comp' tests to BigInt
dukeleto cotto: i have reviewed most of them when i completed the gci tasks they refer to 07:42
cotto: do you want to merge a few? 07:43
cotto yes. I was just going through outstanding pull requests
are they all ok to merge without reviewing? 07:46
dalek rrot: 6a49fd6 | Yuki`N++ | config/gen/makefiles/root.in:
Fix makefile clean target; properly clean generated files in t/src/embed/
 07:49
rrot: 2989842 | Yuki`N++ | config/gen/makefiles/root.in:
Fix straggling reference to QUICKCOVER_DIRS
07:51 Yuki`N left 08:25 theory left
dalek rrot/checkdepend_pmcs: fd73b15 | Coke++ | t/src/checkdepend.t:
Enable testing of src/pmc dependencies.
 08:42
rrot/checkdepend_pmcs: 05f1002 | Coke++ | config/ (2 files):
Avoid duplicate deps on PMCs; add minimum deps
rrot/checkdepend_pmcs: 47a8ea0 | Coke++ | config/auto/pmc.pm:
remove debug output
rrot/checkdepend_pmcs: 15b9d2a | Coke++ | config/auto/pmc.pm:
Minor fixups on calculated dependency paths.
rrot/checkdepend_pmcs: a2b22ae | Coke++ | config/auto/pmc.pm:
every PMC has this as a dep.
09:02 fperrad joined 09:21 rfw left 09:49 cognominal joined
dalek rrot/checkdepend_pmcs: a09555a | Coke++ | config/ (2 files):
move PMC_H_FILES into the pmc dep generator.

add some more dynamic reqs.
  * down to 4 failures in checkdepend.pl		 09:49
Coke that branch is down to 4 failures in t/src/checkdepend.t (and that's with testing all of src/pmc)
10:00 kennym joined
dalek rrot: b743497 | mikehh++ | examples/io/httpd.pir:
replace concat_s_s opcode
 10:23
rrot: b747d04 | mikehh++ | examples/library/ncurses_life.pir:
replace concat_s_s opcode
rrot: 0224f9d | mikehh++ | examples/pir/pirric.pir:
replace concat_s_s opcode
rrot: 8765817 | mikehh++ | examples/sdl/lcd/clock.pir:
replace concat_s_s opcode
Coke wonders why all those were not fixed BEFORE the opcode was removed. 10:27
mikehh Coke: because nobody runs fulltest 10:28
tadzik hmm, any Rubyists out there? 10:29
Coke if you're ripping something out, you should have to.
mikehh yeah - we should emphasize that in the docs somewhere
dalek rrot/checkdepend_pmcs: f8671f3 | Coke++ | config/auto/pmc.pm:
Track cases where our include includes.
 10:30
rrot/checkdepend_pmcs: 1f81d6e | Coke++ | config/auto/pmc.pm:
Fix very old typo

and the pmc2c code were doing 2 different things.
  - checkdepend.t passes again (now with no .c skips)
rrot/checkdepend_pmcs: 19002ba | Coke++ | config/auto/pmc.pm:
whitespace
rrot/checkdepend_pmcs: 987f414 | Coke++ | t/codingstd/pccmethod_deps.t:
Fix same bug that was cut/pasted from auto::pmc

Add note about merging this dependency check into checkdepend.t
rrot/checkdepend_pmcs: b44084d | Coke++ | config/ (2 files):
Remove now-redundant .mak variable.
mikehh Coke: have you any thoughts on generating the header/index pages in html_cleanup
Coke: we have the docs/index/*.jsom files, just now need to generate the pages from them - we have the rest of the .html files now 10:33
Coke mikehh: haven't given it any thought. Doesn't matter because no one ever runs "make html" anyway. ;) 10:34
there, checkdepend_pmcs is done.
mikehh Coke: just me and for release docs 10:36
Coke mikehh: yes, no one runs that, see the mailing list. ;)
mikehh Coke: been following that
Coke must sleep for an hour or 2. later. 10:37
msg dukeleto I'm not sure Karen had anything in particular to do with your grant getting awarded, FYI. 10:38
aloha OK. I'll deliver the message.
mikehh got to go out for a bit - bbl 10:43
tadzik wklej.org/id/445971/ -- what can possibly break stuff in lex_lookup()? 11:41
well, the lex_lookup is ok, but checking the result fails with NPE 11:42
11:46 redicaps joined 12:33 bluescreen joined 13:09 lidden joined 13:17 fbrito joined 13:21 redicaps left 13:35 rurban_ joined 13:37 rurban left 13:38 rurban_ is now known as rurban 13:49 whiteknight joined, smash joined
smash hello everyone 13:49
whiteknight good morning smash 13:52
14:01 mtk joined 14:03 mtk left 14:04 mtk joined, mtk left 14:05 mtk joined
fbrito hm, how should I test EventHandler PMC set_pmc VTABLE? argument is a ParrotInterpreter... 14:29
whiteknight fbrito: let me look 14:30
fbrito whiteknight: ok, thank you very much :)
whiteknight fbrito: do you need to test that VTABLE to complete your task? 14:33
fbrito not really :P. I already completed it 14:34
I was working on this task: www.google-melange.com/gci/task/sho...9228651258 14:35
and now only invoke and this VTABLE are not covered 14:36
but according to the EventHandler test file, invoke is broken (seg fault) 14:37
whiteknight EventHandler.invoke is broken? Invoke is like the only thing that damn PMC type does 14:44
EventHandler serves no purpose if it can' tbe invoked
fbrito github.com/parrot/parrot/blob/mast...thandler.t 14:45
whiteknight: line 88 14:46
whiteknight damnit
fbrito I haven't look the problem well, but if you uncomment the test it will seg fault
whiteknight I think we should deprecate that type. Can you open a ticket on Trac suggesting we deprecate and remove EventHandler PMC? 14:47
fbrito (weird... some times, after reading what I have just wrote, I feel that my english is getting worse and worse :P) 14:48
whiteknight: ok, sure
trac.parrot.org/parrot/ticket/1894 14:50
whiteknight fbrito
: your english is very good
I think we can fix EventHandler.invoke for now, if you want 14:51
in the invoke VTABLE, try to get e->interp. if that's null, use INTERP instead 14:52
I think that should fix the segfault
fbrito I will give it a try... just a moment 14:53
(compiling) 15:02
no success 15:07
dalek rrot: b3af8c9 | Whiteknight++ | / (3 files):
fix main.c so that we check the is_error flag and only show an error message and backtrace if we have an error. This fixes TT #1907. Add two tests for the test cause provided by Coke++ to show that it does indeed have an exit code of 3, but that it produces no error output
fbrito I guess the opportunity to learn gdb has finally come! :D 15:08
whiteknight gdb --args ./parrot t/pmc/eventhandler.t 15:09
fbrito: can you nopaste your attempted fix?
fbrito whiteknight: sure 15:10
Coke whiteknight: thanks. as time permits, I'll rebuild parrot & rakudo and see what happens. 15:11
fbrito whiteknight: pastie.org/1411201 15:13
(I still have to get used to the pastie script in tools folder :S) 15:14
whiteknight Coke++
Coke: I'm very sorry that it took so long to bisect. If I had seen the error report earlier I likely could have narrowed down the cause much more quickly
15:17 Matt_ joined
whiteknight fbrito: change "if (e->interp)" to "if (!PMC_IS_NULL(e->interp))" 15:18
actually, don't bother
that whole thing is so screwed up
fbrito seg fault again 15:23
I have also played with the tests (initialized in another way, changed the invoked sub, and so on) but got no success 15:25
Matt_ whiteknight: How should I go about testing functions like Parrot_api_load_language(Parrot_PMC interp_pmc, Parrot_String lang) ? I wanna finish this task up today 15:26
whiteknight Matt_: comment that function out. It was a prototype 15:27
(sorry there are so many of those!)
Matt_ whiteknight: How about Parrot_api_add_dynext_search_path . Seems to be called from frontend/parrot/main.c 15:28
fbrito ok, I have to go now. When you guys have some time, please accept my claim request on this task: www.google-melange.com/gci/task/sho...9228651258 15:29
Matt_ I dont see anyway of getting the set paths from the Parrot_api_* functions
whiteknight fbrito: I am testing a fix for EventHandler.invoke now. If it works I will push it for you
fbrito ok :D 15:30
I have already increased the code coverage in 17.8% anyway
whiteknight ok 15:34
fbrito if the task is to increase code coverage in 10% and I do 20%, is it ok to ask to you to create another task? 15:35
whiteknight I'm testing now to make sure I didn't break anything. If that works well I'll push it for you to experiment with
fbrito: submit the work you have now for the first task, and I'll open a new task, keeping track of the extra work you've already done
fbrito I saw some other gci students asking for that, but I was always too embarrassed to ask :(. I have even made that some tasks ago (increasing code coverage more than 2x) 15:36
whiteknight I can't edit the task in-place, or I would simply bump it up to 20% and 4 points 15:37
fbrito: if you drop that task, I'll edit it and you can re-accept it at a higher difficulty
fbrito done. thank you :). 15:38
whiteknight 4 points is the same as 2 2-point tasks
Matt_ whiteknight: Any ideas on how to test functions like: Parrot_api_set_warnings(Parrot_PMC interp_pmc, Parrot_Int flags) ? 15:40
Maybe getters should be implemented first in a separate task?
whiteknight Matt_: Search the code for a place that throws a warning message. Write a test that tries to create that condition. Do a c_error_output_like test with the necessary set_warnings code and the expected warning message 15:43
grep for Parrot_warn (src/warnings.c) 15:44
Matt_ thanks!
whiteknight I hope that helps
actually, looking at Parrot_warn I have to ask why it returns 2 on failure, 1 on success. That seems...stupid
Matt_: in your greppings, if you can figure out a reason why it returns those values, I would be very grateful 15:45
fbrito whiteknight: I have to lunch right now... will be back in 40 minutes or so 15:48
I am starting to get afraid of someone claiming the task :P, so I will claim it and if I manage to increase more than 20% you can create another task, ok? 15:49
thank you very much for your attention :D 15:51
Matt_: good luck with your task :)
15:51 fbrito left
dalek rrot: b3de476 | Whiteknight++ | src/pmc/eventhandler.pmc:
attempted fix for EventHandler.invoke, since it obviously hadn't been updated since before the big PCC refactors several months ago. This VTABLE is untested, though tests are in development
 16:00
16:03 theory joined 16:09 JimmyZ joined 16:15 hercynium joined 16:30 JimmyZ left 16:36 dngor left
Coke whiteknight: oddly, while it fixes my PIR snippet, "exit 3" in rakudo still fails. 16:52
16:54 fbrito joined
dalek Some 'nwellnhof' person just gave me a 67602 byte commit packet. They need to be more careful in the future. 16:56
whiteknight Coke: that's extremely weird. do you have the error message text it prints out? 16:59
Coke: as basic sanity, that Rakudo is running against the new Parrot version? 17:00
er, the version I pushed this morning?
nopaste "coke" at 192.168.1.3 pasted "latest rakudo/parrot" (5 lines) at nopaste.snit.ch/27518 17:01
Coke whiteknight: yup.
whiteknight okay, I had to ask
17:05 dngor joined
whiteknight Coke: I'm looking at the source for &exit in rakudo now. It's a very thin wrapper around PIR exit_i. 17:06
Coke: Can you try it with "parrot perl6.pbc ..." 17:07
?
instead of the fakecutable? 17:08
Coke ah, good call. that's it. 17:09
whiteknight okay. my last commit updated main.c, but not pbc_to_exe to match it
lemme fix it
dalek nxed: r723 | NotFound++ | trunk/ (2 files):
fix for parrot deprecated out usages
 17:10
nxed: r724 | NotFound++ | trunk/winxedst1.winxed:
fix string indexing, Issue 15, plobsing++
 17:15
fbrito whiteknight: hello :). I am still getting seg fault error even after your commit :( 17:18
and I didn't understand what that "real_next" variable does :s 17:20
dukeleto ~~ 17:25
fbrito: how goes it?
fbrito dukeleto: hello :) 17:26
dukeleto fbrito: how is GCI treating you? 17:29
fbrito: do you have stuff to work on?
fbrito dukeleto: I am working on this task: www.google-melange.com/gci/task/sho...9228651258
I have already improved it by 17.8% 17:30
dukeleto fbrito: nice work!
fbrito whiteknight agreed that if I improve it by 20%, he will create another similar task
but the most important VTABLE (invoke) is broken 17:31
he tried to fix it (see last commit on master) but no success
dukeleto fbrito: interesting!
fbrito yeah! Increased it by 19.8% :) 17:37
davidfetter fbrito++ :)
fbrito but invoke is still broken, and as whiteknight, it makes no sense to have a EventHandler that can't invoke 17:38
as whiteknight said*
he was even talking about deprecation 17:39
dalek rrot: 264c681 | NotFound++ | t/pmc/hashiterator.t:
test HashIterator with a Hash with non default key type
 17:40
dukeleto fbrito: well, you can still write a TODO test for invoke, which will cover it, even if it doesn't work 17:42
fbrito: did you create a trac ticket for that problem?
fbrito dukeleto: there is already one covering that: trac.parrot.org/parrot/ticket/1894 17:44
dukeleto fbrito: ok, sounds good 17:46
fbrito I guess I will wait for whiteknight to see if he will try to fix it again 17:47
dukeleto fbrito: don't wait on him, work on something else :)
fbrito yeah, but now I will have to work for my father, ahhaha 17:48
17:50 kennym left
whiteknight fbrito: what happens when you use the new invoke? 17:50
fbrito whiteknight: same thing as before. segmentation fault error
whiteknight fbrito: nopaste your test again? I want to try it myself and see what is going wrong 17:51
fbrito ah, and I wrote a test to catch the exception you wrote (e->code null). It is working fine
whiteknight: I have just uncommented what was already written in "t/pmc/eventhandler.t" 17:52
whiteknight oh,okay
17:53 nwellnhof joined
tadzik ~~ 17:54
fbrito pastie.org/1411572
I also tried to initialize the EventHandler PMC with a Hash (like the first test on this same file), but still got the same error
whiteknight ok
dalek rrot: 3562ee6 | Whiteknight++ | / (2 files):
fix pbc_to_exe for Coke++
 17:55
rrot: 8effd61 | Whiteknight++ | t/pmc/hashiterator.t:
Merge branch 'master' of github.com:parrot/parrot
17:55 kennym joined
Matt_ whiteknight: I tried using Parrot_warn but the warning message never gets printed out 17:57
pastie.org/private/nnpykbcef0gpxprnne2p0a 17:58
Seems like Parrot_api_set_warnings has no effect (something to do with the interp?)
whiteknight Matt_: in the call to Parrot_warn, try using PARROT_WARNINGS_EXPERIMENTAL_FLAG instead of PARROT_WARNINGS_ALL_FLAG 18:01
Matt_ whiteknight: same issue (which I why I tried *_ALL) 18:02
whiteknight Matt_: ah, I think you have found a bug 18:05
Okay, write the test how you think it should work and mark it TODO
and then can you open a ticket on Trac for this? 18:06
Matt_ yup 18:08
18:09 fbrito1 joined 18:10 M_o_C joined
Matt_ whiteknight: Should I leave the test in there when submitting the task as it fails? 18:11
whiteknight yes. Mark it TODO
dukeleto Matt_: call todo() instead of ok() in PIR to mark a TODO test 18:13
Matt_ dukeleto: This is a C-based test
whiteknight Matt_: grep for "$TODO", that should give you some examples
18:14 fbrito left
dukeleto Matt_: if your test is in Perl, then you need to do : local $TODO = "a reason"; 18:17
whiteknight msg fbrito: I think your eventhandler test has discovered a nasty GC bug. Awesome
aloha OK. I'll deliver the message.
cotto_work ~~
fbrito++ then 18:18
Matt_ dukeleto: Yep - I looked up. Is there an equivalent for C-based tests (that use c_output_is)
whiteknight Matt_: it's the same, since the test file is written in Perl
the test is C, but the file is perl
fbrito1 :D 18:19
18:19 fbrito1 is now known as fbrito
dukeleto Matt_: you are writing a Perl test that executes C 18:20
fbrito whiteknight: I think I am done with the task: www.google-melange.com/gci/task/sho...9228651258 18:22
whiteknight fbrito: will look in a bit
fbrito ok :D thank you 18:23
tadzik hmm, any Rubyists out there?
bah, nvm
wklej.org/id/445971/ -- what can possibly break stuff in lex_lookup()? 18:24
well, the lex_lookup is ok, but checking the result fails with NPE
dukeleto tadzik: i see you are fighting the good fight 18:27
tadzik dukeleto: to the last commit 18:29
Kapace_ whiteknight: eventhandler test still fails: dpaste.com/291601/ 18:40
fbrito: did you get invoke to work on eventhandler?
fbrito no. according to whiteknight, it may be a "nasty GC bug" 18:41
whiteknight Kapace_: yeah, that event handler test exposes some weird kind of PCC bug
I don't think it's a GC bug anymore
fbrito ah
at the moment, only invoke and set_pmc are not covered by tests 18:42
Kapace_ set_pmc isn't going to be 10% i don't think 18:43
fbrito and 1 line somewhere else
Kapace_ github.com/parrot/parrot/pull/52 :D
whiteknight actually, now it's looking like a stupid EventHandler bug 18:44
this code sucks
18:45 simcop2387 left
Kapace_ fbrito: I had already made tests for mark and get_string_null... 18:47
fbrito oh :o 18:48
where?
Kapace_ fbrito: I just left my tests in create_an_event_and_set_attributes
see "sweep 1" ?
I had tested bad arguments too 18:49
I just couldn't finish it with invoke being broken
18:50 rfw joined
fbrito oh, I really didn't know that you was working on this task 18:50
Kapace_ Tene said that invoking an eventhandler didn't make sense, so I just made the test and commented it
tadzik dukeleto: any ideas? I don't even know where to seek
fbrito Kapace_ so you have also claimed a task to improve EventHandler test coverage in 10%? 18:52
Kapace_ fbrito: yes, I closed one with 17%, but I could not get to the next 10% :( 18:53
dukeleto tadzik: not sure. Tene is a good person to ask, or parrot-dev
tadzik Tene: you around?
fbrito ahh, I see 18:54
Kapace_ I could really use the points, if you know what I mean :P 18:55
Coke whiteknight: that fixed it. 18:56
whiteknight Coke: awesome. Sorry about the bug 18:57
Kapace_, fbrito: the EventHandler PMC really is garbage. I think we should remove it, not improve it's coverage
but that's a personal thing
Kapace_ whiteknight: yeah, Tene was saying that too 18:58
that nobody uses it etc
bluescreen nwellnhof: ping
nwellnhof bluescreen: pong 18:59
dukeleto whiteknight: we need to improve coverage to know how broken it is 19:00
fbrito Kapace_ I was looking an old coverage report
bluescreen nwellnhof: finally found the issue with TT#1765, but it is not easy to fix it... Basically we're creating interpreters and interp's pmc wraps 2 things attributes and the bare interp
fbrito Kapace_ that's why I didn't see the improvements you made
dukeleto whiteknight: and i agree with you, but i still think that GCI students can learn from improving the test coverage of it, and we learn how broken it is
bluescreen the destroy method of the parrotinterpreter.pmc gets rid of one structure not the other 19:01
Kapace_ fbrito: ah, that make sense
fbrito Kapace_ ehhe, my bad. the code I wrote is duplicated and useless 19:02
Kapace_ nah, it better organized! 19:03
bluescreen the problem is that multiple interp's pmc point to the same raw interp... so if we destroy the raw interpreter in the parrotinterpreter.pmc we will get double free corruptions as there are 2 interp pmcs sharing same interp
nwellnhof bluescreen: yes, i tried similar things. it isn't easy to fix. 19:04
Matt_ nwellnhof: I'm once again having trouble running `make cover` with GCC 4.5.1 after the recent changes were merged in
nwellnhof: no source src/src/dynext.c found for src/dynext.c.gcov
nwellnhof Matt: can you try 'make quickcover_new'?
bluescreen nwellnhof: tough i don't understand why we share the interp.
dalek rrot: 76a4452 | Whiteknight++ | src/pmc/eventhandler.pmc:
fix EventHandler invoke. 'Fix' is a strong word considering that this PMC is still old garbage
 19:05
whiteknight Kapace_, fbrito: That last commit fixes invoke and gets that one test to run on my machine
bluescreen I'm guessing that in order to fix that we need some sort of redesign
nwellnhof bluescreen: i don't know that code very well. that's why i gave up after a while...
bluescreen who should be the expert there?
so I can bug him/her
nwellnhof whiteknight? 19:06
fbrito whiteknight: yes, but the invoke test was written by Kapace_. If the task is still available, I think he should claim it 19:07
Kapace_ thanks fbrito, whiteknight what do you think?
dalek rrot/encapsulate-main: 9e74275 | plobsing++ | frontend/pbc_merge/main.c:
fixup constants after merging bytecode
rrot/encapsulate-main: 3bac222 | plobsing++ | t/compilers/tge/basic.t:
falling off the end of a :main sub executes an end op

this exits the test prematurely
bluescreen whiteknight: are you the "to-go" person when it comes to interp and interp_pmc ? 19:08
whiteknight bluescreen: No. I don't think there is a to-go person for that stuff 19:09
I can take a look at it though
I've been through that code many times
bluescreen ok.. then I'll prepare an email for parrot-dev explaining what the issue is with the interp pmc 19:10
whiteknight okay, that works too
bluescreen it has a memory leak, but the design makes it hard to fix
whiteknight where's the leak?
bluescreen the pmc has 2 things the attributes + raw interp 19:11
the destroy method of parrot interpreter pmc only gets rid of the first part
but the trick is that there may be several interp pmcs using the same raw interp
so the fundamental question is freeing the raw interpreter is responsibly of whom? 19:12
Coke original parent interp?
whiteknight bluescreen: Parrot_destroy 19:13
bluescreen Parrot_destroy calls the GC and that calls destroy in the interp PMC.. at that point if destroy does destroy the raw interp, you get double free corruption 19:14
whiteknight bluescreen: Right, the PMC should not free the interp
bluescreen Coke: the parent interp has no way to know which are its children interps 19:15
whiteknight Parrot_destroy kills the GC and destroys all PMCs. Then Parrot_destroy frees the interp
bluescreen for example: (PIR) test_interp = new 'ParrotInterpreter' 19:16
test_interp is only destroyed by the destroy method of the PMC, Parrot_destroy is never called for that interp
the "on_exit" handler is only called for the interp calling exit 19:17
not for the other interps around
whiteknight ah, okay. I see 19:18
fun
as a temp fix, maybe ParrotInterpreter PMC needs a "should I destroy?" flag
bluescreen child interp doesn't register the on_exit on its parents
nwellnhof the GC destroys every PMC with the --leak-test option. the problem is probably the order of destruction. 19:20
Tene Kapace_, dukeleto, tadzik, whiteknight: I can't think of any meaningful or useful interpretation of invoke vtable on an EventHandler PMC, and beyond that, the current impl just crashes. That's not something we want tests for. 19:22
whiteknight ParrotInterpreter.destroy does not call Parrot_destroy
Tene: I fixed the current implementation. It hadn't been updated since the PCC refactors months ago
Tene whiteknight: Ah.
Kapace_ Tene: see whiteknight recent commit, he kinda "fixed" it, but yeah, I agree with ^
whiteknight Tene: I can't think of a meaningful use of the EventHandler PMC
I think we need to deprecate that
Kapace_ whiteknight: should I still take the task? 19:23
Tene whiteknight: I'm sorry? I think a task scheduling system is something we do want, as shown by other work on similar projects: threading, async io, etc. 19:24
whiteknight Kapace_: If you think you can complete it. It will be good practice, but ultimately I think we are getting rid of that type
Tene: yes, ultimately. This implementation is not something that we want. That whole system needs a redesign
Tene whiteknight: Ah, I perhaps misunderstood you. 19:25
Yes, I agree with redesign.
Kapace_ whiteknight: I already wrote the tests, but I couldn't complete it because invoke was broken, now its fixed, but we think EventHandler shouldn't even exist..
Tene IMO threading is rather *more* important, so I'd be surprised to see it addressed before threading is.
Kapace_: It's always better to have tests, so we can *know* when we break things. 19:26
I see tt757 still hasn't been touched in over a year :( 19:27
19:28 Patterner left
whiteknight Kapace_: Yes, complete the task. Get the points. We'll create something more important for you next 19:28
Kapace_ whiteknight: awesome
www.google-melange.com/gci/task/sho...9228651258
whiteknight Tene: I'm waiting on more Lorito developments before working on threading 19:29
hopefully threading will be a central concern for the Lorito design
Tene whiteknight: Sure. I don't think that anyone's priorities are unreasonable here.
whiteknight Kapace_: accepted 19:30
Tene: I see Tasks and Threads being basically the same thing
a good system for one should include the other
Tene whiteknight: Agreed. 19:32
Kapace_ whiteknight: how can I make a pull request only for my last commit?
or should I just make the pull request regardless? 19:33
whiteknight Kapace_: is there stuff in your fork that you don't want to have included?
Kapace_ whiteknight: well theres stuff from other tasks..
whiteknight Kapace_: it would be easier for me to just uncomment that test here instead of trying to pull that commit 19:35
Kapace_ ok cool, just submit a patch then?
rfw did anyone get round to pulling my changes, lol 19:36
dalek rrot: 74a5256 | Whiteknight++ | t/pmc/eventhandler.t:
uncomment out test from Kapace++
nwellnhof whiteknight: i just found an embed_api regeression. when you call exit with a non-zero status, a backtrace is printed to stderr. this causes a Rakudo spectest failure.
whiteknight Kapace_: Pulled. Done 19:37
rfw: let me do that now
rfw yay \\o/
Kapace_ thanks whiteknight, when your not busy, www.google-melange.com/gci/task/sho...9228651258 19:38
19:38 Psyche^ joined, Psyche^ is now known as Patterner
whiteknight Kapace_: done 19:38
Kapace_ whiteknight: awesome, thanks again 19:39
19:39 estrabd joined
dalek rrot/gci_rfw_coverage: dbab727 | (Tony Young)++ | t/pmc/role.t:
Added a test in role.t to cover make_namespace_autobase.
 19:39
rrot/gci_rfw_coverage: a17b5b9 | (Tony Young)++ | / (144 files):
Merge branch 'master' of git://github.com/parrot/parrot
rrot/gci_rfw_coverage: 91fe6d6 | (Tony Young)++ | t/pmc/continuation.t:
Increased coverage of Continuation PMC to 86.60%.
 19:40
rrot/gci_rfw_coverage: b874a65 | (Tony Young)++ | t/src/misc.t:
Adding src/misc.c tests.
rrot/gci_rfw_coverage: 04b407a | (Tony Young)++ | / (4 files):
Merge branch 'nwellnhof/make_cover' of git://github.com/parrot/parrot
rrot/gci_rfw_coverage: 458de73 | (Tony Young)++ | t/src/misc.t:
100% coverage for src/misc, whoo!
rfw i would preferred you rebased my commits together but too late :D 19:44
prefer*
whiteknight rfw: I like to keep the raw history as-is 19:45
rfw heh
actually
i swear that first commit
dbab727
was already pulled in
nwellnhof whiteknight: forget my last message. that was caused by my own changes. 19:47
rfw do i get an entry in CREDITS yet :D 19:48
whiteknight nwellnhof: no, that was a real error. Coke reported it this morning and I fixed it shortly thereafter
nwellnhof ah, ok. 19:49
cotto_work mikehh: ping
seen mikehh
aloha mikehh was last seen in #parrot 9 hours 12 mins ago saying "Coke: been following that".
20:06 fbrito1 joined 20:09 fbrito left 20:19 kid51 joined
mikehh cotto_work: pong 20:19
20:20 allison joined
cotto_work mikehh: how close to merge-ready is html_cleanup? 20:22
mikehh cotto_work: got to generate the index pages otherwise ok 20:24
see my #parrotsketch report (submitting in a minute)
cotto_work ok
kid51 GCI participants: You are welcome to join our weekly on-line developers meeting in #parrotsketch 20:28
ps in 1 20:29
ps START 20:33
plumage? 20:41
aloha, plumage?
aloha kid51: plumage is supposed to be that (I think). but it is mostly built on top of distutils.pbc and setup.pir
tadzik aloha: forget plumage 20:42
aloha tadzik: I forgot about plumage.
tadzik plumage is a package manager for Parrot 20:43
aloha: plumage?
aloha tadzik: plumage is a package manager for Parrot
tadzik I like you, aloha
dalek rrot: d34e58a | Whiteknight++ | / (3 files):
small fixes to help the g++ build. still does not complete
 20:57
rrot/encapsulate-main: ac00c83 | plobsing++ | compilers/imcc/ (2 files):
eliminate seen_main field
 21:01
21:03 TonyC left 21:09 nopaste left 21:16 hercynium left 21:19 hercynium joined 21:21 cognominal left 21:24 [hercynium] joined 21:28 whiteknight left 21:29 hercynium left, [hercynium] is now known as hercynium 21:31 mtk left 21:32 mtk joined 21:34 bluescreen left 21:36 rurban_ joined 21:37 fperrad left 21:38 rurban left 21:39 rurban_ is now known as rurban 21:40 simcop2387 joined 21:47 Yuki`N joined
Yuki`N aloha, coverage? 21:47
aloha Yuki`N: coverage is cv.perl6.cz or tapir2.ro.vutbr.cz/cover/cover-results/
tadzik aloha: trac? 21:48
aloha tadzik: trac is a bug tracker that just happens to bundle a smallish wiki for your convenience.
tadzik I remember that having "or is slow, or is really slow, or..." 21:49
Yuki`N :P
21:49 mtk left
atrodo tadzik> purl had that in it's definition. 21:51
tadzik ah, good ol' purl 21:52
21:55 silug joined
Yuki`N karma N 21:57
aloha N has karma of 15.
Yuki`N lol
tadzik karma C 21:58
aloha C has karma of 143.
Yuki`N The ` in my name breaks the karma system.
karma Yuki`N
aloha Yuki`N has karma of 0.
tadzik I see what you did there, C++
mikehh karma g
aloha g has karma of 113.
Yuki`N :P
g++ file.c 21:59
tadzik yes, that's how it works :)
Yuki`N msg dukeleto can you take care of the 'silent configure' task on GCI? It's distracting and unfortunately invalid. :( 22:00
aloha OK. I'll deliver the message.
cotto_work kid51: ping
mikehh make corevm/make coretest works with g++ now, but frontend/parrot_debugger/main.c fails with make (world) 22:01
kid51 cotto_work: pong 22:04
cotto_work kid51: the gci task to rename makefile targets seems *really* trivial. Am I missing something?
kid51 It may seem trivial ... but we've been having so many problems with those two targets that I suspect the task-doer will think it non-trivial. 22:05
However, feel free to make adjustments
I'll defer to others' judgment re GCI tasks, because I haven't yet been mentor. 22:06
I just care about the corresponding Trac tickets.
moritz there's nothing wrong with trivial tasks, IMHO
cotto_work kid51: I don't see how it could be non-trivial. It's just editing two lines in the makefile template. Any other problems the student would run into wouldn't be the student's concern.
dalek TT #1908 created by jkeenan++: Create new testing target: 'make test' + more
TT #1908: trac.parrot.org/parrot/ticket/1908
kid51 So, please modify the task (including hours needed, if needed) 22:07
Or, if you think it's extremely trivial, delete it.
cotto_work ok
I'm just wondering if you see something I don't. 22:08
kid51 If you delete, just msg me about it so that I know I'll handle the TT myself.
22:08 dduncan joined
kid51 away 22:08
22:08 kid51 left
Yuki`N For 'creating a script creates a trac wiki page', the script doesn't have to actually do the HTTP post and all that, right? 22:10
cotto_work Yuki`N: no
spitting out the markup is fine
Yuki`N Yeah. 22:11
That makes it easier.
Well that was rather easy. 22:27
Thank you Perl!
cotto_work perl ftw 22:29
Yuki`N I'm writing up some POD for good measure.
cotto_work Yuki`N: good. I'd ask for some if I didn't see any. 22:31
nobody'll use if if they don't know what it does
It'd also be good to start a wiki page with the output of the script.
Yuki`N Is there a wiki page there? 22:32
Oh well I'll make one I guess.
cotto_work That's the spirit. 22:33
make sure to document the name of the wiki page in the script
nwellnhof Yuki`N: making a HTTP post would be trivial in Perl. managing the credentials is more of a problem. 22:34
Yuki`N Ah right.
I'm used to full-fledged programming languages
22:34 simcop2387 left
Yuki`N i.e. anything with sockets is usually a pain. 22:34
nwellnhof Yuki`N: have you ever used LWP::UserAgent? 22:35
Yuki`N Nope.
How do I start a new wiki page?
Oh wow.
nwellnhof Yuki`N: give it a try. it will make you appreciate Perl5. 22:36
Yuki`N Hm I cannot edit the wiki apparently. 22:37
nwellnhof Yuki`N: someone has to give you edit permissions. i can't, unfortunately. 22:38
Yuki`N Oh.
cotto_work Coke: ping
Yuki`N: what's your trac username?
Yuki`N Yuki`N
cotto_work If coke pongs, tell him you need a wiki editing bit. 22:39
22:39 M_o_C left
Yuki`N Ok. 22:39
Thinking of ways to make my output nicer.
22:41 hercynium left
cotto_work Yuki`N: the goal is to make it easy to figure out which comments are worth pursuing. Anything that makes that as easy as possible will help. 22:50
Yuki`N I can try perhaps bolding comments with XXX in them. 22:52
Right now they're already sorted by directory, so it's easy to distinguish between say examples, tests, and actual source.
cotto_work Yuki`N: does the script handle multi-line comments correctly? 22:57
Yuki`N No, unfortunately.
It's purely line-based.
It's alleviated somewhat by an automatically generated github link, but I'm not quite sure how I can try to parse adjacent lines from a file given only perhaps the end of a multi-line comment. 23:00
The perl script itself is quite simple, maybe not even 6 lines of code.
cotto_work Don't feel like you need to make it perfect before sending a pull request. 23:01
Yuki`N Well I want to at least test it on trac and perhaps make some purely asthetic changes. 23:02
I just wish there was a quick and easy way to get lines xxx-yyy of a text file. 23:03
cotto_work split on a newline and index into the array?
Yuki`N Maybe. 23:05
Sounds slow though.
cotto_work The first version can be slow. You can profile and optimize later. 23:06
dalek rrot/encapsulate-main: 0ee91bb | plobsing++ | compilers/tge/TGE/Compiler.pir:
tag *every* tge sub :main. emulates old broken pick-first-sub behaviour
 23:07
23:09 dduncan left 23:10 simcop2387 joined 23:11 simcop2387 left 23:13 simcop2387 joined
cotto_work plobsing: how backwards-compatible of you 23:17
nwellnhof year end summary: 23:23
Coke cotto_work: pong 23:24
nwellnhof $ git diff --shortstat -M RELEASE_2_0_0..master -- src
263 files changed, 76955 insertions(+), 48332 deletions(-)
Coke (wiki edit bit) what's the UID?
nwellnhof $ git grep ^ src |wc -l
173422
(skewed by generated files) 23:26
23:27 vm joined 23:32 vm left
Yuki`N Coke, account name is Yuki`N, idk uid. 23:33
Coke The user Yuki`N has been granted the permission WIKI_MODIFY. 23:35
Yuki`N Thanks.
cotto, thanks for your suggestions, multi-line comments work now. :D 23:38
Coke perhaps a C programmer can fix the build for g++ with TT#1909 23:39
Yuki`N Coke, does WIKI_MODIFY also include new page creation? 23:40
Coke no. 23:41
cotto_work Coke: Could I get a trac admin bit so I can make these kinds of changes myself?
Yuki`N Can you grant that one to me as well? I need to create a new page for my script output.
Coke The user Yuki`N has been granted the permission WIKI_CREATE. 23:42
Yuki`N Thanks again.
cotto_work I have a partial bit.
Coke The user cotto has been added to the group admin.
cotto_work phenomenal cosmic power 23:43
thanks
Yuki`N In a git merge conflict, the incoming changes are second, yes?
dalek tracwiki: v180 | Yuki`N++ | WikiStart 23:52
tracwiki: Add hack list wiki page link.
tracwiki: trac.parrot.org/parrot/wiki/WikiSta...ction=diff
TT #1909 created by coke++: g++ build broken
TT #1909: trac.parrot.org/parrot/ticket/1909
Yuki`N Coke, How do you force usage of g++? 23:53
NotFound Yuki`N: Configure.pl --cc=g++ --cxx=g++ --link=g++ --ld=g++ 23:54
Yuki`N trac.parrot.org/parrot/wiki/ParrotHackList how does this look?
Coke NotFound: aren't you our g++ guy? 23:56
Yuki`N g++ build is pretty broken lol.
Coke (i'm trying to bisect where the build went sour for g++ now.)
I'm assuming it was the merge back of embed_api (grumble)
(mainly because I don't know enough C or C++ to fix the damn problem.) 23:57
NotFound Coke: sometimes
cotto_work Yuki`N: not bad. It'd be nice if it highlighted "hack" and didn't catch things like "hackers"
It's certainly enough to mark the task complete.
Coke ooh, looks like it was borked in 2_11_0
btw, the skeleton needs to be updated to remove the Id$ 23:58
(skeleton.vim) 23:59
23:59 whiteknight joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!