parrot.org/ - clean up those smolders for the release!
Set by moderator on 20 October 2008.
00:09 AndyA joined
allison pmichaud: yup, very helpful, I've crystallized my collection of options down to one I currently prefer, and am writing up the details for Andrew 00:13
dalek r32120 | allison++ | pdd22io: 00:43
: Creating branch for I/O PDD implementation.
diff: www.parrotvm.org/svn/parrot/revision?rev=32120
00:49 petdance joined, hiroyuki_y joined 01:11 Ontolog joined 01:25 petdance joined
dalek r32121 | jkeenan++ | trunk: 01:48
: To avoid this warning:
: Use of uninitialized value $ENV{"TEST_PROG_ARGS"} in pattern match (m//) at
: t/op/bitwise.t line 510.
: test for definedness of $ENV{"TEST_PROG_ARGS"} before testing pattern match.
diff: www.parrotvm.org/svn/parrot/revision?rev=32121
r32122 | pmichaud++ | lex: 02:19
: create a branch to re-work lexical handling
diff: www.parrotvm.org/svn/parrot/revision?rev=32122
02:40 Psyche^ joined 02:42 stockwellb joined
stockwellb is it reasonable for a non perl programer to learn Perl 6 as their first go at Perl? 03:03
Tene stockwellb: depends on whether you're asking if it would be useful or if it would be feasible.
It's certainly feasible. It's probably not useful yet. 03:04
Hinrik it would probably be more fun to learn Perl 5 because you can do much more with it
stockwellb could you elaborate on why it wouldn't be useful?
Hinrik and if you know Perl 5 then Perl 6 isn't that hard to learn 03:05
Tene The only problem is that some parts of Perl 6 are defined as "the same as in Perl 5", so there are some omissions from the specification.
Sure. There's not much useful you can do with Perl 6 right now without a lot of work.
Except work on Parrot, mostly.
Hinrik stockwellb: because there are not complete Perl 6 implementations
and Perl 6 the language is still being designed 03:06
as an example, last month these changes were made to the language: perlgeek.de/blog-en/perl-6/tidings-2008-10.html 03:07
and probably others
stockwellb so the migration path for new commers is still via Perl 5. You don't think there will be too much to unlearn. Is it worth waiting? 03:08
Tene stockwellb: Depends on what your goals are. If you want to do productive things with it right now, 5 is better to learn.
Hinrik not too much
Perl 6 will still be Perl
Tene 6 isn't too much of a change. Pretty easy migration path. 03:09
stockwellb I'm not worried about productivity. Most of my paying work is in c#, ruby and python. This really is strickly for balls out fun.
Hinrik then learn both :P 03:10
but I suggest starting with Perl 5
Tene 6 is a fun language, 5 has lots of exciting modules on cpan
those are the bright features for me.
stockwellb I'm not sure my brain will take well to learning 5 and 6 and PIR at the same time.
Hinrik 5 is fun too, 6 is just fun+1 03:11
you wouldn't have to learn PIR unless you want to get into implementing languages on parrot
Tene Which he does, iirc.
stockwellb Actually I really like PIR. Learning it may be a waste of time professionally, but I really like it. There is something primal about it. 03:12
Hinrik heh 03:13
they way I see it: the easiest way to learn a language is by using it, and Perl 5 is infinitely more useful than Perl 6 at the moment, and failing that, most of the material on learning Perl 6 assumes that the reader knows Perl 5 anyway... 03:17
stockwellb It's PIR that I'm really after, It's just obviously such a Perl environment that I felt It might be smart to learn how the locals speak. 03:22
Tene sleeps.
stockwellb If only at a very basic level. 03:23
Hinrik a copy of Programming Perl would probably prove useful 03:24
stockwellb I was looking at that in Barnes and Noble the other day.
chromatic It's current as of 5.6.0, unfortunately. 03:33
cotto are there plans for a new version? 03:34
TimToady tchrist is working on one
cotto tchrist++ then 03:35
chromatic www.perl.com/pub/a/2008/04/23/a-beg...l-510.html 03:38
stockwellb heh... swiss army chainsaw. I like that. 03:40
Hinrik the current version is remarkably relevant for its age, though 03:41
stockwellb ubuntu 7.10 comes with Perl 5.8.8. I should get 5.10 and compile it? 03:42
Hinrik ubuntu 8.10 comes with Perl 5.10 03:43
stockwellb subtle hint to upgrade?!
Hinrik I don't use any of the 5.10 features though, not enough to warrant a dependency on it anyway :P 03:44
but say() is convenient for some one-liners if you're into golfing
stockwellb It's getting late on the east coast. Nite everyone one! 03:46
chromatic It's pretty easy to compile and install 5.10 alongside an existing 5.8. 03:47
stockwellb I'll look into that. thanks chromatic. I'm off to bed! 03:48
04:23 Bzek joined 04:56 particle1 joined 05:17 Theory joined
pmichaud I'm trying to add a new opcode in the 'lex' branch, but after doing 'make' I keep getting a "skipped opcode is also in ..." error at the tools/build/ops2pm.pl step. What have I forgotten? 05:26
particle1 i think there's a util to add an op...
05:26 petdance joined
pmichaud there's the make 'opsrenumber' target 05:26
particle1 maybe
pmichaud i.e., I added my new op to src/ops/core.ops, then did "make opsrenumber' 05:27
that regenerated src/ops/ops.num
particle1 it's possible that this used to work but was changed and nobody tested
but, i'm not willing to give up yet 05:28
pmichaud it has changed, yes -- previously there was a special makefile that was run.
particle1 i just added an op to experimental. rebuilding 05:30
inline op erk(out INT) {
$1 = 1;
}
looks like it's building fine. i'll move it to core.ops and rebuild 05:31
erk_i FATAL: not in ops.num nor ops.skip 05:32
ok running make renumber
pmichaud try "make opsrenumber"
particle1 hole in ops.num before #1 at lib/Parrot/Ops2pm.pm line 122, <$op> line 38.
is this what you get? 05:33
pmichaud not exactly
particle1 oh, nm
skipped opcode is also in src/ops/ops.num at lib/Parrot/Ops2pm.pm line 145, <$op> line 11. 05:34
sorry, i tried to add the op before 'end' which *must* be opcode 0
now i put the op at the end of core.ops and got that error
pmichaud I'm putting mine right before the newclosure op
skipped opcode is also in src/ops/ops.num at lib/Parrot/Ops2pm.pm line 145, <$op> line 11.
purl i already had it that way, pmichaud.
particle1 ok, looks like a bug. *sigh* 05:35
i bet the test coverage is high on this file
*sigh*
pmichaud shall I file a ticket or message to dev list?
particle1 ticket, for sure 05:36
i'll investigate a bit now, though i'm wiped out so bear with me
pmichaud I'm wiped out also -- was ready to simply add an opcode to the branch, commit, and then to bed, but don't have a lot of energy for troubleshooting much further :-(
particle1 hrmm... 05:37
i wonder
are there any ops in both ops.skip and ops.num 05:38
pmichaud good guess 05:39
particle1 yes, there are
pmichaud try reverting, then do "make opsrenumber" and then "make"
we get the same error.
particle1 aye
it's from the mmd merge
the special skipped ops were never removed from ops.skip 05:40
pmichaud *of course it is* B-|
particle1 yeah, this was a failed merge. we have lessons to learn with this one
pmichaud shall I still file a ticket? 05:41
particle1 if you have th estreagth, yes
erk
*the strength
pmichaud okay, writing ticket 05:42
05:44 baest joined
pmichaud ticket filed. 05:49
time for sleep.
any word on booking travel?
particle1 i was hoping to get a response from allison, but nada
go ahead, i'll book mine tomorrow as well
we can talk more tomorrow about timing for sunday evening, if you wish 05:50
pmichaud I may go ahead and make the reservation (to hold the price), I think I'll have 24 or 48 hours to purchase.
particle1 ok
pmichaud I've already decided I'll book a very late flight on Sunday
(9:30p from SJC)
it goes through LAX instead of direct to DFW, but then there's no question about missing anything important at the summit, and I'm still back home on Monday morning 05:51
particle1 it seems the answer is to eliminate all entries in ops.skip 05:57
yeah, that's a late flight alright 05:58
nopaste "particle" at 98.232.28.49 pasted "src/ops/seen.pl" (25 lines) at nopaste.snit.ch/14377
particle1 that's the hack i used to determine it 05:59
pmichaud oops, I was wrong, 9:10p from SJC :-) 06:01
anyway, I have the reservation held until Friday 23:59p, so have a good fare locked in until then
particle1 -responded to you ticket 06:06
i'm bedward &
dalek r32123 | chromatic++ | trunk: 06:13
: [src] Tidied PIC file to make the code somewhat easier to follow; no functional
: changes.
diff: www.parrotvm.org/svn/parrot/revision?rev=32123
r32124 | chromatic++ | trunk: 06:31
: [src] Made fetch_arg_op() always create a COW STRING from constant STRINGs
: stored in the bytecode so that callees never modify constant strings
: unilaterally. This is a more generic fix for RT #60030 and supersedes r32076.
diff: www.parrotvm.org/svn/parrot/revision?rev=32124
06:40 uniejo joined 07:20 cosimo joined 07:34 jq joined 07:37 rene joined
rene hi 07:37
anyone interested in listing the T2 SDE package of parrot on the download page? 07:38
www.t2-project.org/packages/parrot.html
www.parrot.org/download
chromatic What's T2 SDE? 07:39
rene an open source System Development Environment
kind of a source distribution, but with cross build support and tiny c libraries, like uclibc and dietlibc
chromatic Does Parrot work with tiny libcs?
rene used by some folks like gentoo on workstations, and by others to cross build products ... 07:40
at least it passes tests with glibc and eglibc
do not know uclibc off hand
not all t2 packages do neccessarily have to support the tinier uclibc or dietlibc ones
chromatic I'm curious to see the shared library size when linked against one of those. Interesting.
rene I can spin a quick uclibc test build if you like 07:41
chromatic Sure; I'm going to sleep now, but I'll backlog the results.
rene of course a whole lot influences the size, including e.g. -Os vs. -O2 etc... 07:43
chromatic Not as much as you might think, at least with my tests. 07:45
moritz rene: added to www.parrot.org/download
chromatic Anyhow, thanks for reporting the new package.
rene moritz: thanks, too :-) 07:46
dalek r32125 | fperrad++ | trunk: 07:47
: [Lua] bytecode translation
: - now generate .tailcall
diff: www.parrotvm.org/svn/parrot/revision?rev=32125
08:24 barney joined 08:31 iblechbot joined 08:58 cognominal joined 09:14 bacek joined 09:49 allison joined
dalek r32126 | bernhard++ | trunk: 09:49
: [docs] Report success with gcc 4.2.4
diff: www.parrotvm.org/svn/parrot/revision?rev=32126
09:51 cognominal joined 09:53 kj joined 10:18 bacek joined 10:39 tomyan joined 10:41 bacek joined
moritz purl, karma C++ 11:00
purl c++ has karma of -80
szbalint how did that happen? :) 11:12
moritz nobody ever writes C++++ but C++-- somtimes is appropriate ;) 11:13
11:13 bacek_ joined
bacek evening everyone 11:13
hi purl
moritz mahlzeit ;)
purl hola, bacek.
bacek :)
11:14 cognominal joined
bacek moritz: gutten tag 11:14
(sorry for my bad language ;)
moritz s/tt/t/, apart from that it's fine ;) 11:15
dalek r32127 | bernhard++ | trunk: 11:16
: [Pipp] add a TODO test for attribute access
diff: www.parrotvm.org/svn/parrot/revision?rev=32127
bacek moritz: did you see my message about "todo" in Str.pir? 11:17
moritz bacek: yes, but had no tuits so far 11:18
bacek going to find some tuits for moritz ;) 11:19
moritz: any outstanding bugs for ex-developer? 11:20
barney found a tuit from YAPC::EU
moritz bacek: dunno, I have no overview whatsoever atm 11:21
bacek barney: throw it to moritz!
moritz somebody write my diploma thesis, then I'll hack on parrot
(and on the test suite)
barney What's it about ? 11:22
purl it's always about the BOOBIES
moritz barney: spin transport phaenomena on mesoscopic scales
(spin hall effect and so on)
actually I'm just about to begin with it, and already have no time ;( 11:23
barney experimental or theoretical ? 11:25
moritz theoretical
purl i think theoretical is 1.54Mbps
szbalint C++++
12:41 Bzek joined 13:11 stockwellb joined
dalek r32128 | bernhard++ | trunk: 13:14
: [Pipp] add rule 'member' with a dummy action
diff: www.parrotvm.org/svn/parrot/revision?rev=32128
barney learned the hard way that rebuilding pipp.pbc doesn't really rebuild the fakecutable pipp
stockwellb I upgraded to ubuntu 8.04 LTS and now I'm getting make errors. Any ideas?
nopaste "stockwellb" at 70.160.222.145 pasted "make error" (4 lines) at nopaste.snit.ch/14378
particle stockwellb: did you make realclean?
or is this a fresh co? 13:15
stockwellb fresh co
particle did you make -j?
stockwellb line 189 is a parrot include
barney Kubuntu 8.04 works for me
stockwellb make -j?
purl it has been said that make -j is good
particle make -j is parallel make 13:16
we've occasionally had problems with that
stockwellb like parallel universe :0
I'll try it right now.
particle can you make sure runtime/parrot/include/interpinfo.pasm exists? 13:17
stockwellb interpinfo.pasm does exist. 13:18
I checked that first thing.
make -j gave me all sorts of No such file or directory errors 13:19
side note, what irc client do you folks use. This ircii is very boring looking.
barney Do you have an installed parrot? What does 'which parrot' say ? 13:20
stockwellb which parrot says /usr/local/bin/parrot which is a symlink to my build dir.
particle i don't understand why that's happening. can you try make realclean && perl Configure.pl && make 13:23
stockwellb I'm doing that as we speak.
particle: same error. 13:26
particle i irc from windows, and use pidgin, btw 13:27
many folks love irssi and customize it wildly
szbalint irssi++ 13:28
karma irssi
purl irssi has karma of 208
PerlJam irssi++ indeed
szbalint that many flies can't be wrong...
particle stockwellb: try running that command (./parrot -o ...) manually 13:29
stockwellb i was using bitchx but that package isn't available now. I'm installing irssi right now
okay, I'm loggin out and I'll be right back. 13:30
13:30 stockwellb joined, masak joined
stockwellb irssi is much better! 13:30
szbalint a clean, dry experience 13:31
stockwellb particle: what next? 13:32
particle stockwellb: try running that command (./parrot -o ...) manually
stockwellb I did already. It bombs. I'll try again just to make sure.
particle ok, i didn't see that response 13:33
stockwellb yea it's like parrot doens't know where it's include path is.
particle right
13:33 ruoso joined
particle make tags && vim -t parrot_init_library_paths 13:35
...if you're a vim user
the library paths are defined in src/library.c, line 142
why they're not found... 13:36
stockwellb I'm a young vim user. Haven't done tags. I run the command as you've typed it in vim?
particle vim -t foo searches your tags file for a tag named 'foo' and brings you to the location 13:37
vim src/library +142 will open to line 142
so, you can do either 13:38
stockwellb I get no tags found
particle you need "make tags" first 13:39
or maybe i misspelled the function...
stockwellb no rule 'tags'
particle hey, what happened!?!? 13:40
purl We don't know what happened, so tell everyone nothing happened.
particle where is the tags make target?
uggh.
maybe it's been renamed 13:41
PerlJam make help
purl make help is my friend
stockwellb I feel like I've been renamed... from bruce to whiner.
PerlJam looks like it's tags-vi
particle tags-vi apparently
13:42 nopaste joined, cognominal joined
nopaste "stockwellb" at 70.160.222.145 pasted "make error" (13 lines) at nopaste.snit.ch/14379 13:42
stockwellb errors!
particle sudo apt-get install exuberant-ctags 13:43
perl Configure.pl && make tags-vi
13:43 gryphon joined
stockwellb ok, I'm at the function in library.c 13:45
particle all this was just to show you that the paths are defined
stockwellb I see them 13:46
aaaaand I've learned something.
particle it doesn't get us closer to fixing it
good :)
i have ubuntu 8.10 (i think i skipped 8.04) and everything just works 13:47
but i don't think this is an ubuntu problem
actually, i have no idea why it's a problem at all
stockwellb when I was 7.10 I did a reallyinstall. 13:48
could that be causing problems
particle possibly, but the command is specifically running ./parrot
stockwellb true
PerlJam if parrot was compiled shared and installed, even ./parrot could be using the wrong lib 13:49
(unless LD_LIBRARY_PATH is set correctly)
particle well then, it can be a reallyinstall problem 13:50
PerlJam stockwellb: in general, don't install parrot. It's bad mojo (still)
(unless your specifically going to fix that bad mojo :-)
stockwellb I know that now!
particle i don't know how to uninstall it. locate libparrot i guess 13:51
libparrot.so that is
stockwellb its in /usr/local/lib 13:52
particle yeah, that's a problem
hopefully, it's *the* problem
stockwellb and in /usr/lib 13:53
should I kill them
particle yes please!
stockwellb both?
purl both is country *and* western or salt and MSG
particle yes
they belong only in the build dir
...and you haven't built them there yet. 13:54
stockwellb libparrot.a libparrot.so libparrot.so.0.4.13
safe to kill all three
particle ooh, that's old. yes
stockwellb ok all six files have been deleted. clean configure and make right? 13:56
particle yep
13:57 apeiron joined
stockwellb same error. 14:01
PerlJam stockwellb: what was the errro? 14:02
er, error
particle the error isn't instructive
nopaste "stockwellb" at 70.160.222.145 pasted "make error" (6 lines) at nopaste.snit.ch/14381
PerlJam stockwellb: is your parrot from svn or the release tarball? 14:04
stockwellb from svn
PerlJam Have your tried nuking your WC and checking it out again?
s/your/you
stockwellb yes I have. 14:05
particle the problem is outside the wc 14:06
it's definitely due to a previous reallyinstall
i just don't know how to clean that up
stockwellb yet it worked with ubuntu 7.10?!
Tene Just write a cron job to make reallyinstall every 20 minutes 14:07
stockwellb I don't get the humor in that?
dalek r32129 | bernhard++ | trunk:
: [Pipp] Add directory for test scripts written in PHP
diff: www.parrotvm.org/svn/parrot/revision?rev=32129
r32130 | bernhard++ | trunk:
: [codingstd] Update MANIFEST, as a new file has been added
diff: www.parrotvm.org/svn/parrot/revision?rev=32130
Tene stockwellb: most people don't get much of my humor. 14:08
stockwellb I'm probably not learned enough to appreciate it.
Tene No, you just haven't learned to ignore me yet. :)
dalek r32131 | bernhard++ | trunk:
: [codingstd] add SVN props for new file
diff: www.parrotvm.org/svn/parrot/revision?rev=32131
14:08 NotFound joined
stockwellb Ok, I'm really really sorry I ran reallyinstall! 14:09
Tene Nah, it's fine. I did it several times before I learned not to.
stockwellb Several? Perhaps there is hope for me. 14:10
barney Last time I had the 'make reallyinstall' problem I did a "find /usr/local '*parrot*'" and deleted all files manually
14:10 bpphillips joined
Tene could work 14:10
PerlJam stockwellb: barney's solution sounds reasonable to me too
stockwellb Ok, there are definately a grunch of files in there. 14:11
a bunch in /usr/local/runtime 14:12
14:12 mj41 joined 14:14 jhorwitz joined
stockwellb should I kill /usr/local/runtime? it's empty now. 14:15
Tene doesn't hurt either way 14:16
stockwellb so it's definately a parrot dir
running make... 14:17
uhggg!! 14:19
same error. 14:20
find /usr/local *parrot* returns no files now.
pmichaud try find /usr/local -name '*parrot*' 14:21
stockwellb ah yes, more file thingies. 14:22
pmichaud (the asterisks need escaping from the shell, thus the quotes.)
particle this needs to go into the faq 14:24
pmichaud or we need a 'make reallyuninstall' :-P
particle are you sure you want to uninstall? it may fix your system. if so, type make reallyuninstall. 14:25
Tene make maybeinstall
14:26 cognominal joined
stockwellb all the pbc* binarys are parrots right? 14:28
particle *.pbc probably 14:29
stockwellb no in /usr/local/bin
pbc_disassemble
pbc_info
particle ah, pbc_* yes
stockwellb pdump?
purl pdump is, like, just that well named ;)
particle yes
pdb 14:30
Tene We should disable the reallyinstall target for now, to prevent things like this from happening?
stockwellb Great, I'm the poster child for stupid parrot tricks. 14:31
particle no
we should improve the message for 'make install'
Tene No, it's just that nobody has had enough incentive yet to actually FIX installing.
pmichaud the whole point of 'reallyinstall' was to make people think twice about it
particle it's unclear what the ramifications of installing are
rurban has been working on fixing install
Tene We really really do need to get it fixed before 1.0, but it hasn't been important enough yet with 1.0 still a decent ways off.
rurban++
NotFound I tryied recently to install in a non privileged directory, and using LD_RUNTIME_PATH to locate the dynamic library all seems to work as expected. 14:33
stockwellb Damn. Still the same error
would installing and uninstalling the 0.4 package help? 14:35
via apt-get
particle huh, what? you installed a package?
definitely
NotFound stockwellb: The ctags error you pasted before?
particle sudo apt-get remove parrot-0.4.13 or whatever
stockwellb way back I installed 0.4 then unistalled it when I decided to get the svn src. 14:36
I'm just shooting in the dark.
NotFound: what do you want from me? 14:38
NotFound stockwellb: I don't understand what do you care about make tags-vi 14:39
stockwellb NotFound: particle was showing me something. 14:40
dalek r32132 | bernhard++ | trunk: 14:41
: [Pipp] Try to simplify t/harness and add support for tests
: written in PHP and executed in the non-default variants of Pipp.
diff: www.parrotvm.org/svn/parrot/revision?rev=32132
14:41 jq joined 14:42 hercynium joined 14:43 Andy joined
stockwellb libparrot-0.4.13 was still showing as an installed package. I removed it, make realclean, Configure.pl, make. Still same error. 14:45
particle nuke the site from orbit 14:46
stockwellb aye captain!
nuke deployed... rebuilding from scratch! 14:47
7 1/2 minutes to midnight. 14:48
PerlJam Isn't that an Iron Maiden song? ;) 14:52
stockwellb I'm not sure. I was just referencing the doomsday clock.
Still the same error!! 14:53
particle have you considered running windows? ;) 14:54
stockwellb Ouch!
So we're at wits end? 14:55
particle i r
i wish rurban was around
stockwellb frack!
Perhaps he'll be around this evening? 14:56
particle i haven't installed parrot since 0.0.6 btw :)
and that was 5 machines ago
stockwellb apparently installing parrot is not such a good idea.
particle i really don't know why what you've done to uninstall isn't enough 14:57
PerlJam stockwellb: can you give me access to the machine? I may have a little time to look at it.
stockwellb I really don't know how to do that?
remote?
purl remote is the HTTP connection
NotFound stockwellb: Have you deleted any symbolic link you created for the install?
particle parrot -v 14:58
stockwellb I kept the symlink in /usr/local that points to build dir.
particle hopefully there's no 'parrot' left
stockwellb parrot --version says 0.8.0-devel 14:59
NotFound stockwellb: delete it
particle it == the symlink?
NotFound Yes 15:00
pmichaud particle: think I should commit chromatic's patch in RT #60070? It seems to fix 'make opsrenumber' for me
stockwellb how do you rm a symlink
particle pmichaud: i reviewed it but didn't test. if it works, go for it
PerlJam stockwellb: with rm :)
stockwellb damn it was a typo that's why it didn't work. 15:01
symlink gone.
no more whereis parrot
PerlJam: same error. 15:06
The only parrot that exists now is in the build dir.
NotFound stockwellb: make a realclean and reconfigure and rebuild, to be sure. 15:07
PerlJam stockwellb: Seriously, I was trying to think of an appropriate response if you said "same error" and now you have. Format the disk. Wipe it with magnets. Reinstall the OS. etc.
dalek r32133 | bernhard++ | trunk:
: [Pipp] add test file for working on a simple testing lib in PHP
diff: www.parrotvm.org/svn/parrot/revision?rev=32133
stockwellb NotFound: I did just that. 15:08
PerlJam: I've actually got paying work on this partition, so I'm trying to be careful. That's why I was still running ubuntu 7.10. I was just being cautious. 15:10
I should have stayed cautious.
Parrot worked fine until I upgraded. 15:11
pmichaud I'm running 8.04 with no problems. I tried installing 8.10 beta on my notebook last week and it failed to install, so I'm waiting for next week's release before trying 8.10 again. 15:12
NotFound stockwellb: I upgraded ubuntu on several machines with parrot, and never have such problems. 15:13
stockwellb why me!!!!
PerlJam stockwellb: so ... in my copy of parrot, line 189 of parrotlib.pir is including interpinfo.pasm. do you have interpinfo.pasm?
stockwellb yes I have that file. 15:14
moritz in any case, after an OS upgrade you should 'make realclean' and rebuild it from scratch, because some tools might have changed
particle moritz: done and done
PerlJam stockwellb: where is it located?
NotFound General rule: after any problem, make realclean ;)
PerlJam stockwellb: And, perhaps more importantly, where is parrot looking for it?
NotFound Or even before the problem
particle the problem is: ubuntu 7.04 -> make reallyinstall -> upgrade ubuntu -> kablooey
Infinoid has "make realclean; svn update; perl Configure.pl; make -j8 test" in muscle memory 15:15
stockwellb it's in the runtime/parrot/include dir
particle Infinoid: why even bother with a number after -j?
Infinoid habit. the linux kernel runs you out of system RAM when you do that.
thousands of gcc processes executing at once = bad 15:16
PerlJam stockwellb: it's like the version of parrot you're running isn't looking in the right place. That would imply that you've still got some remnants of the previous parrot installation hanging around.
stockwellb: are there any parrot-related environment variables in your environment?
stockwellb and a good way to check that would be?
moritz env|grep -i parrot
env | grep -i ld_ 15:17
NotFound stockwellb: find / -name "*parrot*"
PerlJam I usually env | sort | less and scan manually just in case (the list is usually small enough)
Infinoid if make reallyinstall put some stuff in /usr/local, you can probably do a find to get those files
stockwellb gasp!! PARROT_RUNTIME=/usr/lib !!!
PerlJam ding! 15:18
stockwellb I know how to set it, how do I kill it
PerlJam unset PARROT_RUNTIME
(is that set in your profile or something?) 15:19
NotFound stockwellb: look at /etc/profile and family, maybe the package you had installed set it, and don't unset at uninstall
But first grep PARROT_RUNTIME .* in your home directory 15:20
stockwellb nothing in /etc/profile
grep PARROT_RUNTIME .* shows bash_history only 15:21
NotFound grep -R PARROT_RUTIME /etc/ 15:22
stockwellb nothing
NotFound N
stockwellb N?
purl N is, like, load runner with a physics engine, robots, and ninjas or for Neville who died of ennui
NotFound grep -R PARROT_RUNTIME /etc/ 15:23
Don't copy my typo ;)
PerlJam stockwellb: btw, you might want to install ack if you haven't already :)
Andy ALL THE COOL KIDS RUN IT
stockwellb PARROT_RUNTIME was in /etc/environment 15:24
I removed it.
I'm typing like a one legged man in an ass kicking contest. I've got no time for installing ack!! 15:25
Andy yes you do 15:26
NotFound stockwellb: close your session to completely get rid of it
Andy sudo cpan App::Ack
stockwellb by session you mean my terminal right?
PerlJam stockwellb: yes
NotFound stockwellb: the X session
stockwellb can I compile now? 15:27
NotFound stockwellb: recheck the environment
stockwellb still there. 15:29
particle you're typing too slowly. 15:30
NotFound stockwellb: close the X session. If not, any terminal you open still inherits it from the main process
particle how do you restart x?
NotFound You don't need to restart the server, just the user session
particle ah
nuke the site from orbit 15:31
purl it's the only way to be sure.
particle ...that's all i know how to do sometimes... :)
stockwellb alright I'm loggin off of here. logging off session... be back soon.
NotFound particle: BTW, control-alt-backspace nukes the X server in Linux PC 15:32
15:33 stockwellb joined
Infinoid we're so geeky, we have multiple three-finger salutes. 15:33
stockwellb Ok, PARROT_RUNTIME environmental varible is no longer set!
particle NotFound: thanks, that's good to know, and works in my ubuntu vm
NotFound And kdm has a menu option to restart the server 15:34
particle Infinoid: i thought -j without # took the # of cores + 1 by default
Infinoid nope, it just spawns as many processes as it can, as quickly as it can
NotFound stockwellb: realclean again to be sure, and all must go well now 15:35
Infinoid (or as many as the dependency tree will allow.)
particle aha
stockwellb in process right now.
make went well. Running tests... 15:40
PerlJam even with the human delay factor, it looks like you need a faster computer :) 15:42
stockwellb $ound$ like a good idea plea$e $end help. 15:43
Tests are done! Yea!!!!
particle group hug! 15:44
purl Cannonball!
stockwellb Thank you NotFound, PerlJam and Particle et al!!!
So 220 wallclock seconds is slow?
testing that is. 15:45
moritz only needs 117, hehe ;)
stockwellb dang.
moritz but that's cheating, I use parallel testing
on two cores at once
stockwellb My poor Latitude D820
I have intel Core2 Duo. How do I cheat? 15:46
particle moritz: is that simply TEST_JOBS=9 make test ??
moritz particle: TEST_JOBS=2
particle i'll try 5 (quad-core) 15:47
particle tries TEST_JOBS=5 make -j 5 make world test
i wonder how that'll interact 15:48
15:49 sjansen joined
stockwellb TEST_JOBS=2 make test took 189 secs. Compared to 220 without. 15:51
moritz one CPU?
stockwellb Yea duo core
I'm definitely *NOT* the cool kid on the block. 15:52
particle make -j 5 world #fails 15:53
stockwellb Once again, thank you everyone for helping me get parrot back up and running. I learned a bunch of other shit too! I'm going to lunch. Later! 15:54
NotFound particle: I think there are several non explicit dependencies in the sources that makes any parallel build risky. 15:56
particle NotFound: so it seems. we keep trying to remove them, but they keep showing up 15:57
NotFound particle: is hard to fix and to keep it fixed without some automated way.
particle the automated way should be smokers 15:58
dalek r32134 | pmichaud++ | trunk: 16:06
: [core]: Fix some issues with opcode renumbering (RT #60070)
: * NOTE: "make realclean" required
: * Patch to lib/Parrot/OpsRenumber.pm from chromatic++
: * Bring ops.num up to date
: * Update PBC_COMPAT, fix instructions in PBC_COMPAT
: * Give 'pdump' an explicit path in tools/dev/mk_native_pbc
diff: www.parrotvm.org/svn/parrot/revision?rev=32134
r32135 | pmichaud++ | lex:
: Remove 'lex' branch to re-sync from trunk (opsrenumber fixes)
diff: www.parrotvm.org/svn/parrot/revision?rev=32135
r32136 | pmichaud++ | lex: 16:07
: Create new lex branch from trunk for lexicals work.
diff: www.parrotvm.org/svn/parrot/revision?rev=32136
16:13 pjcj joined 16:27 Ademan_ joined 16:35 Theory joined
NotFound Someone can take a look at src/inter_call.c:2669 ? 16:36
It gives a warning, and looks like an error.
Tene you mean this line: 16:39
}
?
NotFound if (interp->run_core = PARROT_CGP_CORE) 16:40
Ups, the line number has changed because I was testing headerizing 16:41
Tene 2665
There was a recent commit about CGP 16:42
Chromatic did something with it in r32094
NotFound Looks like a typo for == 16:46
Tene ... oh, right.
Yeah, I agree.
Yeah, that was added with the commit I referenced. 16:47
Definitely a bug. Good catch.
NotFound++
NotFound -Wall++ 16:48
dalek r32137 | fperrad++ | trunk: 16:54
: [PCT]
: - now generate .tailcall
diff: www.parrotvm.org/svn/parrot/revision?rev=32137 16:55
Tene NotFound: are you going to commit the fix? 16:56
NotFound Tene: I'm testing before
pass
Tene Wait, you can run tests before you commit? Interesting idea. 16:57
I'll have to consider doing that in the future.
Maybe.
dalek r32138 | julianalbo++ | trunk:
: fix a typo from r32094
diff: www.parrotvm.org/svn/parrot/revision?rev=32138
NotFound Tene: only thuesday afternoons. 16:58
moritz I get a failure in t/stm/runtime.t 17:01
and then a pass on running it again. 17:02
dalek r32139 | julianalbo++ | trunk: 17:07
: add some PARROT_..._RETURN_NULL to make headerizer happier and update headerizing
diff: www.parrotvm.org/svn/parrot/revision?rev=32139
NotFound Uh, oh...
pmichaud iirc, some of the stm tests have race conditions on some platforms 17:10
17:12 Zaba joined
NotFound Don't kill me yet, I'm fixing it. 17:14
17:20 chromatic joined 17:21 ruoso joined
dalek r32140 | julianalbo++ | trunk: 17:22
: forward declare PackFile_ByteCode in parrot.h to avoid headerizing and #include problems
diff: www.parrotvm.org/svn/parrot/revision?rev=32140
NotFound Sorry for the break, I forgot I was testing with c++ 17:23
Coke attempts to catchup from yesterday, Oy. 17:49
particle pmichaud: would you agree that src/classes/Module.pir should become Package.pir, and Module isa Package? 17:55
Coke "what's this? Now." 17:56
pmichaud particle: I don't know what's in Module.pir at the moment -- I guess jonathan++ created that 17:57
I agree that Module isa Package
particle it's just onload and WHAT
pmichaud is it even used?
particle i'll instead make Package derived from NameSpace and Any, and make Module a subclass
pmichaud make it so. 17:58
particle aye
pmichaud ultimately we're likely to end up with a Perl6NameSpace PMC
particle is there a NameSpace class in perl 6? i haven't seen one in the spec 17:59
i'll have to ack, of course
pmichaud obviously Perl 6 has namespaces, but no, there hasn't been mention of a namespace class
TimToady and I had a discussion about this a couple of weeks ago on #perl6 18:00
particle ok, because i'm thinking i'll put the EXPORT/EXPORTALL methods in Package
irc log?
purl it has been said that irc log is irclog.perlgeek.de/parrot/
pmichaud I was asking what was the type of object that received the .EXPORTALL invocation
particle irclog.perlgeek.de/search.pl?channe...=NameSpace 18:01
ah, ok i'll look for EXPORTALL
pmichaud irclog.perlgeek.de/perl6/2008-10-11#i_617225
also, there's nothing that says that Package has to be isa NameSpace -- it just needs to know how to forward to a NameSpace when apropriate 18:05
so I think I'd prefer Package to not be a subclass of NameSpace 18:09
chromatic Ditto.
pmichaud unless there's a big reason it has to be that way
particle i don't see a reason, that's confused me from the start 18:12
however, parrot NameSpaces should probably be accepted as rakudo Packages, so p6meta.'register'('NameSpace') seems appropriate 18:13
pmichaud until we actually need that, I'd rather not implement it
particle ok, fine, won't do that either 18:14
pmichaud i.e., I'd prefer to see the use case
particle so Package isa Any
so all i need is .new_class, don't need .register at all
chromatic don't need .register, don't need .stack 18:16
If you've got the VM honey, I've got the time.
particle heck, i'm not sure i even need WHAT
pmichaud right
I'm not sure why Module was implemented in the first place :-|
particle so, is there a way i can declare Package in perl 6 code?
lemme rip it out and see what falls down
this is wacky 18:17
of course, i need to put a EXPORT method on Package still, don't i?
Tene look at r29922
pmichaud it needs to go somewhere, yes.
Tene "[rakudo] Make .WHAT called on a NameSpace return a Module proto-object. Added to unblock mod_perl; probably will want a re-visit when we look at modules more comprehe...
particle shakes his fist at jhorwitz 18:18
pmichaud it's okay with me if .WHAT on a NameSpace returns a Package protoobject
if mod_perl needs it
Tene particle: it was added by Jonathan.
particle yes, but added for jhorwitz
chromatic Never trust anyone under 30, or with a J in his name.
particle can i declare Package in nqp instead of pir now? 18:19
pmichaud no 18:20
particle ah, well
pmichaud (I don't know what you mean by "declare Package in nqp", but I'm certain the answer is no. :-)
particle class Package { ... }
heck, in perl 6 i mean.
pmichaud that wouldn't make it a subclass of Any
(in nqp) 18:21
particle "isa Any" is easy to add
ok, so we need Package. can i get rid of Module? 18:22
pmichaud I don't know if mod_perl expects Module
you'd probably have to ask jhorwitz about all of this.
particle yep, he's ignoring me 18:23
dalek r32141 | pmichaud++ | trunk:
: [rakudo]: spectest-progress.csv update: 206 files, 4429 passing tests
diff: www.parrotvm.org/svn/parrot/revision?rev=32141
particle there's roughly a 1/3 chance mccain will be dead before jhorwitz returns my call
actuarial statistics are amazing things 18:24
chromatic Hey, the stats went up again!
pmichaud I'm so glad to see that the members of the Parrot Foundation board communicate so well :-) :-)
Tene Communicate like a fox!
pmichaud or maybe that's just officers :-)
dalek r32142 | pmichaud++ | lex: 18:26
: Initial implementation of capture_lex .
diff: www.parrotvm.org/svn/parrot/revision?rev=32142
Tene pmichaud: do you have any work for me tonight? 18:31
jhorwitz awakens
Tene flees in terror.
pmichaud Tene: what are we blocking on for .hll support, if anything?
Tene pmichaud: HLLCompiler needs to know the .HLL of the language it's compiling for, or we need to store namespaces or protoobjects or something in the attributes 18:33
pmichaud protoobjects
purl protoobjects are constrained to be undefined
jhorwitz particle & pmichaud: do what you will with module & package. mod_perl6 will comply. 18:34
Tene All languages that will use HLLCompiler will always use P6object?
Okay, I can do that.
pmichaud tene: there's more, just a sec
Coke communicate? With those JOKERS?
Coke comes in late to the party.
pmichaud jhorwitz: we need to know what mod_perl6 needs
i.e., when is it using .WHAT on namespaces now?
jhorwitz mod_perl6 is very needy. :) 18:35
Tene gone for teaching, bbiab
pmichaud tene: at present, the reason why I'm using strings instead of protoobjects is that at the time the HLLCompiler object is being created, the action grammar and grammar haven't been created yet
tene: it may be as simple as moving the .include 'src/gen_grammar.pir' and .include 'src/gen_actions.pir' to the top of the language.pir file instead of at the bottom 18:36
then the protoobjects should be available at the time the onload sub is executed
jhorwitz pmichaud: i was going to use .WHAT to determine if a handler name was a namespace or a subroutine
it's not currently implemented
pmichaud oh 18:37
use .isa
jhorwitz sure -- if that works now, i'll gladly use it.
pmichaud although it might not work now for NameSpace PMCs
jhorwitz ah. hm.
pmichaud anyway, using .WHAT for deciding class membership is almost always wrong 18:38
(yes, there's still a lot of it in actions.pm)
particle well, shoot, i can clean that up
particle roots around for his meme eraser... it's here somewhere, i swear! 18:39
pmichaud particle: I think some places even have .HOW().isa()
I'd like those to just become .isa
chromatic I believe that's called "Not being an adolescent".
pmichaud we may need to add .isa to PCT::Node to make that happen, but that would be fine.
particle ok, so i'll leave Module.pir alone for now 18:40
change .WHAT and .HOW.isa to .isa, implementing PCT::Node.isa if necessary, and then try removing Module.pir
purl particle: that doesn't look right
pmichaud that works for me
we may also need to have Rakudo define a .isa method for the NameSpace PMC 18:41
(for mod_perl6)
particle ok, but that's not something mod_perl6 needs today, correct jhorwitz?
jhorwitz i need it to implement method handlers, but it's not urgent, no. 18:42
pmichaud yes, if mod_perl6 doesn't need that today, I'm in favor of simply removing it for now.
jhorwitz so go ahead and remove it
pmichaud yay.
particle i wonder how much this will affect chromatic's favorite benchmark 18:43
(ripping out code)++ 18:44
chromatic Fixing isa/does in PMCs helps greatly too. 18:45
Rough guess: 20% improvement by reducing GC pressure.
jhorwitz just reconfirmed that .isa gives an exception when called on a namespace. 18:46
dalek r32143 | pmichaud++ | lex:
: [imcc]: Fix IMCC so it doesn't generate Closure PMCs
diff: www.parrotvm.org/svn/parrot/revision?rev=32143
Coke (reducing GC pressure) tcl needs that; badly. 18:47
crap, now I reaaaaly need to get tcl working with svn head. 18:48
chromatic We need to get PIR profiling working.
If I could tell you where you spend most of your time, we could optimize there.
jhorwitz mmmmmm
Coke chromatic: yes. 18:49
I took a look at the callgrind output, and it was a bigger than a one day project, so I didn't pursue it.
chromatic It's not that bad. 18:50
You can get by with the simple output.
Coke bigger than one day for me. ^_^
chromatic The trouble is following Parrot's control flow.
nopaste "particle" at 98.232.28.49 pasted "PCT::Node.isa + sample changes in actions.pm." (47 lines) at nopaste.snit.ch/14383 18:56
particle let me know if that looks correct before i go changing everything the wrong way
jhorwitz looks like i can use try blocks and some crafty logic to implement method handlers. i don't think i'll be blocking on NameSpace.isa.
Coke ponders creating a branch for the (*#@$ tcl/tge update. 18:58
(in tcl, not parrot)
pmichaud PCT::Node.isa needs to do the equivalent of .HOW.isa
particle ah, ok 18:59
and i can't use [PAST;Val] in nqp
i need PAST::Val i suppose
pmichaud right
particle or 'PAST::Val'
pmichaud it's PAST::Val
no quotes
purl no quotes are ever needed on multi lines, because of indenting
chromatic It would be fairly easy to make IMCC support literal Key creation.
pmichaud purl, forget no quotes
purl pmichaud: I forgot no quotes
pmichaud in Perl 6, we would write $x.isa(Foo::Bar) so that's what NQP expects to do. 19:00
and Foo::Bar would be a protoobject.
moritz (actually we'd write $x ~~ Foo::Bar ;-)
pmichaud (that's possible also, but .isa works.)
moritz right 19:01
particle ok so $P0 = self.'HOW'() ; $I0 = isa $P0, type 19:03
pmichaud no
$P0 = self.'HOW'() $I0 = $P0.'isa'(type)
particle ok, i'll stop being lazy and investigate further then
pmichaud because <type> isn't necessarily a Parrot class 19:04
particle ah, right
pmichaud (in fact, it's almost certainly _not_ a Parrot class)
particle can i .tailcall $P0.'isa'(type) ?
pmichaud should be okay
since it's not a vtable function 19:05
(no, don't be tempted to make it one.)
particle don't worry, i'm not.
19:06 ruz joined
particle i still have trouble thinking the rakudo way... i so often try things the parrot way first 19:06
pmichaud the key is remembering that perl 6 (and the parrot compiler toolkit) almost never use Parrot classes directly. 19:07
s/perl 6/rakudo/
so opcodes like 'isa', 'does', 'new', etc. need special treatment (or, more likely, should be delegating to something else that already does the special treatment) 19:08
afk for a while (kid pickup)
Coke thinks he's got tcl's tge issues worked out. 19:35
branches++
particle1 coke++
Coke runs make test. 19:36
aw, the closure update was in the lex branch. 19:39
am I right in assuming that's the one that is supposed to reduce GC pressure?
woot. make test in partcl's tge_update passes, along with the expected passing TODO tests. 19:40
jhorwitz coke++
19:45 particle2 joined
Coke meh. is that karma worthy? Thanks, though. 19:47
particle2 there really is no way to make you less cranky, is there. 19:48
moritz feels that half of his karma isn't deserved either. It's just a general noise level... 19:49
Coke particle2: not that I've noticed.
particle2: if you figure it out, be sure to tell my wife. =-) 19:50
Coke realizes that he can't rely on getting karma in here from partcl commits any more, and becomes more effusive. 19:53
Coke does yet another test against release 0.8.0 19:54
moritz Coke: well, you just need a bot that pastes partcl commit messages here. 19:55
Coke There; code.google.com/p/partcl/wiki/PartclSource 20:03
I am tempted to try out that closure patch and see if tcl's spec test runs any better. 20:04
NotFound Interesting: if I add a methot can_handle that returns true to a subclass of ExceptionHandler, it handles the exception.
chromatic Makes sense to me. 20:05
NotFound So it looks like the problem is that can_handle in ExceptionHandler does not do the right thing by default.
Maybe just another symptom of the generic problems of inheriting from pmc. 20:06
20:07 apeiron joined
pmichaud coke: the 'lex' branch is where I'm working on a re-implementation of lexicals 20:09
GC pressure isn't at all a driving force (afaik)
NotFound Will be the correct way to add a new test with the can_handle method, and let the test without the method as todo? 20:10
pmichaud currently the lex branch breaks lots of tests (as expected) until I can get all of the pieces in place
NotFound Did we have some test called free? 20:13
I want to break free!
chromatic Free's failing.
particle2 make or break
Coke what was the GC pressure comment from before directed at, chromatic? 20:16
particle2 about isa, iirc
Coke ah. 20:17
particle2 like, caching the result so the lookup is done once
Coke hurm. I can't have a slurpy macro, can i.
chromatic Not always doing stringy comparisons to test isa.
particle2 i forget if macros support varargs. don't think so
NotFound particle2: C macros? 20:18
purl C macros are the pits
Coke no, PIR
chromatic Pretty sure they don't slurp either. 20:19
dalek r32144 | julianalbo++ | trunk: 20:23
: add a new test for ExceptionHandler subclassing
diff: www.parrotvm.org/svn/parrot/revision?rev=32144
Coke so much for my clever idea to write a macro for setting up an exception handler. 20:28
(with an arbitrary # of types.)
(ooh, shiny webdev work) 20:29
chromatic Subclass EH and override its can_handle method.
Coke that's a lot of work for syntactic sugar.
I'd consider it if I had more uniform handler args, though. 20:30
chromatic What kind of uniform, matching slacks and a blazer?
pmichaud well, as I noted yesterday -- internally Parrot's isa is still doing stringy comparisons :-(
anyway, I'm sure that'll get fixed. 20:31
chromatic It's on my list. 20:33
20:41 gmansi_ joined
Coke chromatic: I wonder if it's time to re-attack the integer type ids branch. 20:52
(mainly by deleting it and starting over.)
20:52 bacek joined
particle2 are there still blockers to removing type ids? 20:52
chromatic Why delete it? 20:53
Coke because I don't want to try to do the merge.
perhaps that's the wrong kind of lazy 20:54
chromatic Yeah, deleting branches and throwing away work seems kind of silly.
Of course, if it's more work trying to merge than it is to recreate the work, that's another thing. 20:55
Hard to say without more evidence.
Coke it was, the last time I tried.
There's a reason it's been idle for so long.
chromatic I say do a diff against the point where you branched, save that as a patch, and see what applies to trunk. 20:56
We can redeem part of that perhaps.
Coke That was my plan; try to get a patch of things that didn't say "merge from trunk" and use that as a the starting point of a fresh attempt.
particle2: the problem was, as i recall, that a LOT of code wants those type ids. 20:57
and I don't think we had a plan for how to deal with some of those bits.
(have to check the various tickets for particulars)
-> 21:02
21:06 Zaba_ joined 21:07 Theory joined
particle2 www.parrot.org/wiki/parrot-develope...8-arrivals 21:10
jhorwitz should probably book his flight 21:11
pmichaud is there an rss feed for the wiki? 21:35
21:36 Zaba joined, allison joined
jhorwitz hm, doesn't appear so. 21:37
PacoLinux kidd #parrot, del server irc://irc.perl.org/parrot 21:41
sorry
recruiting some people .. 21:42
21:47 gmansi joined 21:50 rgrau joined
Tene particle2: any word on funding for PDS? 21:54
particle2 PacoLinux++ 22:00
PacoLinux :) 22:02
22:28 cotto joined
Tene pmichaud, particle: what are the times of your return flights? 22:56
Are the arrivals on that wiki page Nov 14 or Nov 15? 22:57
23:02 davidfetter joined 23:08 ruoso joined
pmichaud Tene: Nov 14 23:13
my return flight is late Nov 16
Tene Okay, thanks. 23:15
23:15 TiMBuS joined
cotto chromatic, ping 23:22
chromatic ponig 23:24
pong
23:25 bacek_ joined
nopaste "cotto" at 96.26.202.243 pasted "Complex PMC segfault" (52 lines) at nopaste.snit.ch/14387 23:25
cotto I don't know if that's minimal enough, but it reliably segfaults for me.
23:26 pjcj joined
chromatic Looks easy enough to reproduce. 23:29
23:35 tewk joined
particle2 tene: i fixed up the arrivals page to include date 23:41
also, i created a departures page
chromatic Also you should win stuff for attending.
Tene thanks
23:45 apeiron joined
particle2 we all win just because you're attending, chromatic 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!