Post closed tickets in your report. | Note: This channel is for our weekly status meetings (Tuesdays at 19:30 UTC); you probably want #parrot instead. | irclog: irclog.perlgeek.de/
Set by moderator on 23 July 2014.
07:19 basiliscos joined 08:07 lizmat_ joined 14:41 Chirag joined 17:29 basiliscos joined
Util I have been called out of town, and will not be able to attend today's #ps meeting live. 17:43
Please conduct the #parrotsketch meeting as scheduled.
I will read the backscroll when I return, and contribute if relevant.
rurban Ok, will do 17:57
19:23 Chirag joined 19:28 rurban joined
rurban Ok, time to #ps. Chirag are you here? 19:30
Chirag Hey! yes
#Done: 19:31
* Tested my tasks against parent commits to confirm speed gain values.
Results on my machine seem unreliable:
gist.github.com/ZYROz/499d1e19623eff085619
#ToDo:
rurban Util is away today. I haven't found any errors with pcc-gh1083 yet. Ready to merge from my side. 19:32
But I still wanted to do some 32bit testing (small machines)
Next release is Aug 19, so we have time until Sat. 19:33
Chirag a ran the bench again today and got 0.5% faster .. have lost faith in my machine
rurban No, such small variations are to be expected.
esp. on slow machines 19:34
I got problems with my VM's and test machine. One got hacked
Chirag oh.. but I could not confirm any gain even on the first task
hacked!
rurban it was sending some spam, yes. did a re-setup 19:35
Chirag oh
rurban but it was running the buildfarm service
so I have no good result on the builds anymore for a while
Chirag hmm.. 19:36
rurban So I have to test it manually
but my VM's are now running again. /dev/net/tun was gone
So it looks pretty good, so far. The biggest problem is the perl6/nqp bug with parrot callcontext reuse. 19:37
There's not even a ticket for it yet. so far fixed in perl6 by disabling some optimizations on parrot 19:38
Chirag oh.. I was sick so I didnt follow.. not fixed yet
?
I see 19:39
rurban fixed for now, but we still have to find the real error somewhen
I tested the pcc-gh1083 branch fine on perl6
Chirag do you suspect it is due to a recent commit?
oh thats good
rurban the relevant commit was reverted yes. 19:40
Chirag I was wondering if I could close GH#1083?
rurban cf544ab7ed0cc4f02de3ce2a83f989ee in rakudo was the problem 19:41
I will close GH#1083 as soon as it is merged
Chirag sure 19:42
rurban It doesn't do any harm, and helps a bit with fast machines
Chirag yes.. ~1.5% .. 19:43
rurban I pushed the rakudo branch parrot-ccreuse to repro the perl6 problem with parrot 19:44
to my fork
Chirag Also, I was wondering if I am supposed to upload git diff for GSoC evaluations? (maybe between the commits I have used to run tests... but would also overlap with others' work) 19:48
rurban Added now github.com/perl6/nqp/issues/176 19:51
your work is easily checkable by git log --author=ZYROz 19:52
Chirag even I will try to play around and see if I hit something
rurban Write a nice blog entry about the 3rd part 19:53
You can refer to whiteknights blog post about the slowdown, and that the PMC'ification by bacek wasnt't that big a culprit. More likely the autoboxing as whiteknight already suspected 19:55
Chirag my last blog covers the 3rd part completely... for evaluations do I write one big-blog to summarize everything?
rurban I see. yes, would be nice to have an overall
Chirag oh.. alright maybe I will write a blog for this week tomorrow.. and on Friday write one complete final blog 19:56
rurban great
Chirag and when submitting for evaluations, I just submit the output of git log --author=ZYROz? I mean no code required to be submitted separately right? 19:58
rurban I forgot. 19:59
Chirag ?
rurban since it was already merged, there's no real need to submit code 20:00
is my opinion
Chirag hmm.. even the guidelines say I can upload what I please
rurban lemme check. do you have a link?
Chirag let me find it.. it was a mail 20:01
www.google-melange.com/gsoc/documen...guidelines
rurban I see, they need required code samples to Google
Then just add a -p 20:02
git log -p --author=ZYROz
Chirag oh great! That helps.. 20:03
rurban and links to the 3 tickets
I would even include the 3 tickets as full textfiles 20:04
Chirag yes, I can make one text file and include these links along with links to all my blogs
rurban ah, yes
Chirag will finish everything on friday 20:06
Also, I could start working on the autobox problem after 22
rurban $ g l --author=ZYROz -p | wc -l
81767
80k lines!
Chirag :D 20:07
most of them must be from task 1 20:08
rurban ops, there are some stupid patches in it. with src/pmc/callcontext.i and more generated sources e.g.
Chirag oh 20:09
When do you move to Germany?
rurban and I'll have to assign d815693c4648 to you, that was really your work, not mine
Chirag thats fine I think.. there is a lot of work you did but is attributed to me :) 20:11
rurban yes, but I should be fair
Chirag thanks! and a bigger thanks for being more of a co-worker :)) 20:12
rurban sure
Chirag so you moving to Germany soon? 20:13
rurban ok, that should be it. I started a few smokers on my small machines now
I'm moving end of september, in 2 months
lot of moving stress
Chirag which city?
rurban Dresden 20:14
if you submit codesamples, please remove the slack from the patch. like the .i file and the generated sources
Chirag ok.. umm, is there any way to filter that out automatically? 20:15
not much .. I can remove it manually 20:17
rurban g l --author=ZYROz -p --exclude=src/pmc/callcontext.i does not help 20:18
manually should be easier
Chirag yes, I will do that
rurban ok, seems that we are clear
Chirag yes 20:19
rurban ok, done for today. continue at #parrot
Chirag thanks! 20:21
20:43 basiliscos 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!