Perl 6 - the future is here, just unevenly distributed

IRC log for #parrot, 2014-07-08

Parrot | source cross referenced

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:16 kid51 The tarball generated from 'make smoke' on Linode appears to be sound -- or at least as sound as that generated on Darwin/PPC.
00:16 kid51 So I don't really have any idea why smolder.parrot.org can't extract data from it.
00:57 kid51 Util: check your mail.  It appears to be a Smolder problem, but not really an OSUOSL problem.
00:57 kid51 It appears that Smolder is now being maintained by Maddingue:  https://github.com/Smolder/smolder/blob/master/README.md
00:58 kid51 Perhaps we could file an issue there -- but that's it for me tonight.
00:58 Coke smolder used to need to be cleaned out on a regular basis, and we didn't have the keys to do it, IIRC.
00:59 kid51 Coke: How's life?!
00:59 Coke sokay. you?
01:00 kid51 Good.
01:00 kid51 I had lunch 2x with particle at YAPC in Orlando, as well as dinner with Util and spouse
01:00 kid51 IIRC -- it was two weeks ago and my memory is already hazy
01:45 FROGGS_ joined #parrot
02:34 Tene joined #parrot
03:53 Util kid51: I will prod it in the morning.
03:54 Util 26 of the last 30 smokes are blanks. kid51: are all 26 yours?
04:45 davidfetter joined #parrot
07:08 he joined #parrot
07:09 Maddingue kiI more of less try to maintain it
07:10 Maddingue s/ki/kid51: /  but, he's not here >_>
07:11 Maddingue the important thing is that that merged all the fixes that several people had made over the years in RT or Github, and merged with the PSGI version
07:27 basiliscos joined #parrot
07:29 FROGGS_ joined #parrot
09:13 basiliscos joined #parrot
09:36 basiliscos joined #parrot
09:42 basiliscos joined #parrot
09:47 Psyche^ joined #parrot
09:54 basiliscos joined #parrot
10:33 basiliscos joined #parrot
10:34 bighugedog joined #parrot
10:40 basiliscos joined #parrot
10:42 basiliscos joined #parrot
10:52 basiliscos joined #parrot
10:54 kid51 joined #parrot
13:43 rurban1 joined #parrot
13:55 basiliscos joined #parrot
13:55 tadzik joined #parrot
13:56 tadzik left #parrot
14:05 rurban kid51: see https://github.com/parrot/parrot/issues/1078
14:07 rurban Util: most of the blank smolder reports are from my automatic linux smokers
14:07 rurban Util: however only linux fails to upload. darwin, windows, cygwin, solaris works fine
14:07 rurban I checked out a local smolder version but it was too hairy for now
14:12 bluescreen joined #parrot
14:13 Coke All the smolder issues make we want to get back to work on my replacement. Will see how far I can get this week.
14:15 rurban good. I switched to python's buildbot locally
14:15 rurban But buildbot is super slow with only one server instance and ~8 clients
14:27 Chirag joined #parrot
14:40 FROGGS joined #parrot
14:58 Chirag rurban: I have got perl-5.14.2 and perl-base 5.14... but now git wants 5.18... Can you please suggest a way out?
14:59 rurban cannot you have both?
15:00 rurban if not, then I guess it's time to compile a proper perl5.14 into /usr/local/ (default), not via perlbrew
15:01 rurban ./Configure -des && make -s -j4 && sudo make install
15:01 Chirag i am not using perlbrew anymore.. I used .deb for that
15:02 rurban but the 5.14 deb deleted your 5.18 perl?
15:02 Chirag yes
15:02 rurban and git is broken now?
15:03 Chirag yes
15:03 Chirag maybe I can now download 5.18 and give it another destination
15:03 rurban git is more improtant. re-install ubuntu perl proper (5.18), find the old perlbrew source dir and re-compile there
15:03 rurban ./Configure -des && make -s -j4 && sudo make install
15:04 rurban should be <10 min
15:08 rurban most of the work is sweat (recompile + testing), not thinking or debugging
15:10 Chirag hmm.. not much to show this week..
15:14 rurban well, you can describe your search to find the next problem
15:14 rurban and the various obstacles in its path
15:19 Chirag sure..
15:21 Chirag done.. now I guess I need to set the PATH for perl-5.14.4 so that parrot doesn't pick up the new one
15:29 rurban no PATH changes, just install it into /usr/local/bin
15:29 rurban which is the default
15:30 rurban perlbrew messes with the PATH and PERL5LIB env, which complicates things
15:30 rurban I install everything with App::perlall which install
15:30 rurban s multiple perls side-by-side without any perlbrew problem
15:30 rurban But this is overkill for you
16:41 basiliscos joined #parrot
17:00 Chirag finally .. its working.. running bench.sh on all commits..
17:01 rurban :)
17:01 rurban and now try to keep the error rate low, < 1%
17:01 rurban i.e. no video streaming, kvm or such
17:02 Chirag ok...
17:03 Chirag "Invalid option "without-libffi". See "perl Configure.pl --help" for valid options" --- Is this error normal? showing for all commits...
17:06 rurban yes, this is normal. libffi-3.1 is broken with older parrot's, so I disabled it, but very old parrots didn't understand this flag.
17:06 rurban that's why I try several Configure options
17:07 Chirag good to know..
17:07 Chirag also, ../bench.sh: 71: ../bench.sh: perf: not found ?
17:09 rurban perf is required. package linux-tools
17:10 rurban sudo apt-get install linux-tools
17:17 Chirag oh.. so I will re-run bench.sh
17:19 Chirag last one.. "Use of qw(...) as parentheses is deprecated at config/init/defaults.pm" ?
17:21 rurban yes, ignore, just a warning
17:22 rurban runb a single ../bench.sh first, to see if it works at all. it wipes out all private files in your parrot src dir
17:37 Chirag this is what i get for a commit... http://pastebin.com/mZjn9Se9
17:40 rurban Chirag: looks god, but:
17:40 rurban loadavg  3.53 is too high
17:41 rurban +-  0.14% is good
17:41 rurban 34.246651062 seconds time elapsed, looks like a very slow machine :)
17:41 rurban I would also remove the installed /usr/local/lib/libparrot.so
17:42 rurban but a slow machine is better to see the differences. my machine is already too fast
17:43 Chirag slow machine undoubtedly ... what is loadavg exactly?
17:45 rurban a number how busy your cpu is. ~ number of waiting processes
17:45 rurban https://en.wikipedia.org/wiki/Loadavg
17:45 Chirag oh.. got it
17:45 rurban should be < 1
17:46 rurban how many CPU cores do you have?
17:47 Chirag 4
17:47 rurban ok, then it's better
17:47 rurban you just need to observe that loadavg will not deviate too wildly, or you'll get too high error rates
17:48 Chirag hmm.. I will run it overnight and keep my system untouched
17:48 rurban in my tests I almost cannot use the numbers per commits, too busy machine.
17:49 rurban so I have to check ranges which commits exactly did slow down the tests
17:49 rurban so far I'm still at the guessing stage
17:50 Chirag ok..
17:50 rurban => https://github.com/parrot/parrot-bench/blob/master/log.bench-3.0.slowdown.numbers
17:52 Chirag so that 8-10s range corresponds to my 34s?
17:52 rurban yes :)
17:52 Chirag :D
17:52 rurban more to 8s
17:53 rurban or your benchmarks threw an exception
17:53 rurban maybe run it manually, without throwing away stdout and stderr
17:53 rurban we really should check the exit codes also
17:55 Chirag manually as in - one commit at a time?
18:02 rurban in run-bench.sh
18:05 Chirag removing ">/dev/null" might help?
18:18 Util My RL meeting may go long, so I may be late to #parrotsketch. If so, please meet as normal, and I will catch up.
19:13 bluescreen_ joined #parrot
20:07 rurban Chirag: yes, both redirects
22:44 dalek parrot: 868c410 | rurban++ | / (3 files):
22:44 dalek parrot: namealias: small pod changes
22:44 dalek parrot: review: https://github.com/parrot/parrot/commit/868c410383
22:57 travis-ci joined #parrot
22:57 travis-ci [travis-ci] parrot/parrot#1060 (master - 868c410 : Reini Urban): The build passed.
22:57 travis-ci [travis-ci] Change view : https://github.com/parrot/parrot/compare/b1b4b875888b...868c41038310
22:57 travis-ci [travis-ci] Build details : http://travis-ci.org/parrot/parrot/builds/29459602
22:57 travis-ci left #parrot
23:01 kid51 joined #parrot

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

Parrot | source cross referenced