Perl 6 - the future is here, just unevenly distributed

IRC log for #perl11, 2016-07-27

| Channels | #perl11 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
02:53 willthechill joined #perl11
05:41 travis-ci RPerl build passed. Will Braswell says 'Learning RPerl, Constant Data, Part 1'
05:41 travis-ci https://travis-ci.org/wbraswell/rperl/builds/147666441 https://github.com/wbraswell/rperl/compare/b28e573f2cfd...8858a20970b4
05:41 willthechill yay
05:41 willthechill left #perl11
05:41 willthechill joined #perl11
06:44 rurban joined #perl11
07:06 willthechill joined #perl11
07:25 rurban1 joined #perl11
07:28 rurban2 joined #perl11
07:35 travis-ci perl11/cperl#1349 (smoke/various - 6f8d196 : bulk88): The build was canceled. https://travis-ci.org/perl11/cperl/builds/147685737
07:36 punter joined #perl11
07:44 willthechill punter: howdy!  any progress over the last day or two?  :)
07:44 punter no...
07:44 punter been busy with other stuff
07:44 willthechill okie dokie, I will be awake for while longer
07:45 willthechill waiting for tests to finish on travis now
07:45 willthechill then an early dinner  :-)
07:47 willthechill will be publishing an updated version of Learning RPerl in the next few days, so that should be helpful!
07:47 willthechill (yes it will obviously include the v1.7 -> v2.0 fix, and tons of other more critical info)
07:50 punter nice
07:53 sten rurban2 hi
07:53 rurban2 hi
07:53 sten is unexec will be replacement for B::C?
07:54 rurban2 no. B::C can do lot of memory optimizations. unexec is just very fast
07:54 sten ok, do you time to fix memory leak issue?
07:54 rurban2 But I hope unexec will be more stable, and is much simplier
07:54 rurban2 nope
07:55 rurban2 I didn't see a leak yet. So far all is expected perl behaviour
07:55 sten I mean different leak, wait
07:56 sten https://github.com/rurban/perl-compiler/issues/368
07:56 rurban2 https://github.com/rurban/perl-compiler/issues/368 the bc-leak2
07:57 rurban2 this seems to be critical for event loops, yes
07:57 sten yes, in real project it consume about 8gb per day
07:58 rurban2 I think I already fixed the CV refcnt down to 1. Maybe the padlists have a +1. the pads itself cannot
07:58 sten and zero with no compiled version
08:00 rurban2 hmm
08:00 sten all ref counted object can be seen from arena map? if yes then its not related
08:00 travis-ci RPerl build passed. Will Braswell says 'Learning RPerl, Constant Data, Part 2'
08:00 travis-ci https://travis-ci.org/wbraswell/rperl/builds/147684303 https://github.com/wbraswell/rperl/compare/8858a20970b4...5df7f15f7f78
08:00 rurban2 it can only be a pad
08:01 rurban2 that's in a seperate array
08:03 rurban2 unexec will not have this problem btw
08:05 rurban2 unexec already works on linux with criu btw. but it's in no distro yet. need to compile it by your own, and run a daemon. see self-dump with a python example
08:06 rurban2 on solaris it is also stable
08:09 sten what about windows?
08:10 sten http://pastebin.com/YnLmr8hQ valgrind report for bc-leak2
08:11 rurban2 windows seems to be the hardest.
08:12 rurban2 yes, it's definitely the pad. a better test would be to close over an object with a DESTROY method in this closure. and see if the DESTROY is timely called.
08:17 travis-ci perl11/cperl#1350 (maint-5.24c - cb0aa3c : bulk88): The build passed. https://travis-ci.org/perl11/cperl/builds/147686172
08:28 sten its called timely
08:29 sten I've added test to issue
08:30 punter joined #perl11
08:37 punter joined #perl11
08:48 rurban2 this is now interesting. wonder why this doesn't happen with core then
08:50 travis-ci perl11/cperl#1351 (master - 6f8d196 : bulk88): The build passed. https://travis-ci.org/perl11/cperl/builds/147686507
09:29 rurban joined #perl11
10:24 rurban1 joined #perl11
11:02 mako joined #perl11
11:03 punter joined #perl11
11:40 travis-ci perl11/cperl#1352 (smoke/various - 3fdd5ff : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/147726420
12:32 punter joined #perl11
12:42 mako joined #perl11
13:41 travis-ci perl11/cperl#1353 (maint-5.24c - 51e74b8 : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/147748985
13:57 punter joined #perl11
14:11 rurban joined #perl11
14:44 punter joined #perl11
15:11 rurban joined #perl11
15:59 punter joined #perl11
17:27 rurban joined #perl11
18:13 willthechill joined #perl11
18:15 punter joined #perl11
18:25 bulk88 http://paste.scsys.co.uk/528612
18:25 bulk88 willthechill ^
18:27 travis-ci perl11/cperl#1354 (smoke/various - 7429257 : Reini Urban): The build was broken. https://travis-ci.org/perl11/cperl/builds/147822800
18:34 willthechill bulk88: checking
18:35 willthechill bulk88: okay let's debug it, when I run this command in Linux... lib/RPerl/Test/Operator10NamedUnaryChdir/program_00_good.pl
18:36 willthechill I get this output:    have $foo = 1
18:36 willthechill have $bar = 0
18:36 willthechill and I get no errors when I run this command:    rperl -V -D -nop -t lib/RPerl/Test/Operator10NamedUnaryChdir/program_00_good.pl
18:36 willthechill so go ahead and try those and see what your results are, please
18:36 willthechill presumably you will find an error somewhere
18:45 rurban1 joined #perl11
18:51 mako bulk88: Do this first on Win: 'set HOME=%HOMEPATH%'
18:52 mako Than 'lib/RPerl/Test/Operator10NamedUnaryChdir/program_00_good.pl' works. 'have $foo = 1' and 'have $bar = 0' as expected.
18:54 rurban joined #perl11
18:54 travis-ci perl11/cperl#1354 (smoke/various - 7429257 : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/147822800
18:57 bulk88 http://paste.scsys.co.uk/528617
18:59 bulk88 http://paste.scsys.co.uk/528619
19:16 punter joined #perl11
19:37 willthechill checking
19:38 willthechill okay the first paste is fine where you run script/rperl
19:38 willthechill it only failed at the very end
19:38 willthechill the rest was okay
19:38 willthechill the second paste is where our values differ
19:41 willthechill bulk88: from http://perldoc.perl.org/functions/chdir.html
19:41 willthechill ...changes to the directory specified by $ENV{HOME} , if set; if not, changes to the directory specified by $ENV{LOGDIR} .... If neither is set, chdir does nothing and fails.
19:42 willthechill so apparently mako is correct
19:42 willthechill :-)
19:42 willthechill how should we handle this?
19:42 willthechill oops gotta go offline, will check IRC log when I get back online
19:43 willthechill bulk88: the current question for you is... when somebody (like you) doesn't have HOME or LOGDIR set, what do you think we should do?
19:43 willthechill okay bye!
19:59 punter joined #perl11
20:01 mako bulk88: If on Windows what about setting '$ENV{HOME} = $ENV{HOMEPATH}' or '$ENV{HOME} = $ENV{USERPROFILE}' in 'lib/RPerl/Test/Operator10NamedUnaryChdir/program_00_good.pl'?
20:02 mako Maybe the last one is slightly better suited?
20:26 mako left #perl11
21:23 travis-ci perl11/cperl#1355 (smoke/various - f4d716c : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/147867260

| Channels | #perl11 index | Today | | Search | Google Search | Plain-Text | summary