Parrot 6.1.0 "Black-collared Lovebird" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 18 February 2014.
00:24 rurban1 joined
rurban1 Found the fix for #1041, the next -O2 bug. 3 remaining -O2 test fails 00:24
dalek rrot/rurban/imccO2: 5ee237a | rurban++ | compilers/imcc/ (2 files):
[imcc] Fix one more -O2 bug #1041 get_results write

correctly detect get_results writes to a reg. skip constant propagation then.
There was an old wrong in instruction_writes for exception handling before get_results. exceptions are now proper PMCs.
Also add more debug prints for -d41 in constant propagation.
Fixes all testO2 testcases but 3: t/op/exit.t (Wstat: 256 Tests: 9 Failed: 1)
   Failed test: 6
   Non-zero exit status: 1
t/op/jitn.t (Wstat: 256 Tests: 14 Failed: 1)
   Failed test: 13
   Non-zero exit status: 1
t/pmc/exception-old.t (Wstat: 256 Tests: 22 Failed: 1)
   Failed test: 19
   Non-zero exit status: 1		 00:32
rrot: 5ee237a | rurban++ | compilers/imcc/ (2 files):
[imcc] Fix one more -O2 bug #1041 get_results write

correctly detect get_results writes to a reg. skip constant propagation then.
There was an old wrong in instruction_writes for exception handling before get_results. exceptions are now proper PMCs.
Also add more debug prints for -d41 in constant propagation.
Fixes all testO2 testcases but 3: t/op/exit.t (Wstat: 256 Tests: 9 Failed: 1)
   Failed test: 6
   Non-zero exit status: 1
t/op/jitn.t (Wstat: 256 Tests: 14 Failed: 1)
   Failed test: 13
   Non-zero exit status: 1
t/pmc/exception-old.t (Wstat: 256 Tests: 22 Failed: 1)
   Failed test: 19
   Non-zero exit status: 1		 00:48
rrot: 132ad1a | rurban++ | ChangeLog:
ChangeLog: add GH #1041 summary
rurban1 Looks like I can add the jit also soon, so we have a bit of an advantage over moar for some time. async io would be nice. I added that to p2 easily. 00:53
kid51 runs smolder tests for first time in ... a long time 00:56
01:02 travis-ci joined
travis-ci [travis-ci] parrot/parrot#958 (master - 132ad1a : Reini Urban): The build is still failing. 01:02
[travis-ci] Change view : github.com/parrot/parrot/compare/d...2ad1af078c
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/19773310
01:02 travis-ci left
rurban1 kid51: I always run fulltest before every commit to master. smolder needs some warmup time, doesn't react to the first few uploads. 01:04
see #1040
02:01 FROGGS_ joined 03:11 kid51_ joined 04:49 rurban1 joined
dalek rrot: 8c9f827 | rurban++ | t/ (4 files):
[test] TODO the 3 remaining failing -O2 tests

Each one with a ticket. Still trying to fix them though. Looks easy enough. Add -O2 to t/fullharness. Save to run now, to detect further regressions.		 05:44
rrot: 03345be | rurban++ | / (3 files):
[imcc] fix -O2 check types in constant propagation #1043

set reg, const is only a valid const setter ifregx and const have the same type, otherwise the type are coerced, like set N0, 1 or set I1, 1.5. Do not propagate such values. Fixes t/op/jitn_13.pir [GH #1043]
05:53 Chirag_ joined 05:58 travis-ci joined
travis-ci [travis-ci] parrot/parrot#959 (master - 03345be : Reini Urban): The build is still failing. 05:58
[travis-ci] Change view : github.com/parrot/parrot/compare/1...345bee87e8
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/19782396
05:58 travis-ci left 06:38 davidfetter joined 07:07 rurban1 joined 07:45 rurban1 joined 08:32 FROGGS joined 08:46 rurban1 joined 08:48 TonyC joined 09:10 denis_boyun joined 09:46 rurban1 joined 10:48 rurban1 joined 11:49 rurban1 joined 12:32 Chirag_ joined 12:41 bluescreen joined 12:44 mtk joined 12:51 rurban1 joined 13:25 denisboyun joined 13:49 FROGGS joined 13:52 rurban1 joined 14:42 benabik joined 14:56 rurban1 joined
dalek rrot: 414d7f8 | rurban++ | / (3 files):
[imcc] fix -O2 empty ins->next in used_once [GH #1042]

If the deleted op leads to the last op, which has no op->next break the used_once loop.
Fixes GH #1042 t/op/exit_6.pasm
   set I0, 0 # used_once => deleted
   end		 15:30
rrot: f495335 | rurban++ | ChangeLog:
ChangeLog: Add -O1 and -O2 to fulltest
rrot/rurban/imccO2: 8c9f827 | rurban++ | t/ (4 files):
[test] TODO the 3 remaining failing -O2 tests

Each one with a ticket. Still trying to fix them though. Looks easy enough. Add -O2 to t/fullharness. Save to run now, to detect further regressions.
rrot/rurban/imccO2: 03345be | rurban++ | / (3 files):
[imcc] fix -O2 check types in constant propagation #1043

set reg, const is only a valid const setter ifregx and const have the same type, otherwise the type are coerced, like set N0, 1 or set I1, 1.5. Do not propagate such values. Fixes t/op/jitn_13.pir [GH #1043]
rrot/rurban/imccO2: 414d7f8 | rurban++ | / (3 files):
[imcc] fix -O2 empty ins->next in used_once [GH #1042]

If the deleted op leads to the last op, which has no op->next break the used_once loop.
Fixes GH #1042 t/op/exit_6.pasm
   set I0, 0 # used_once => deleted
   end
15:45 travis-ci joined
travis-ci [travis-ci] parrot/parrot#960 (master - f495335 : Reini Urban): The build is still failing. 15:45
[travis-ci] Change view : github.com/parrot/parrot/compare/0...95335c309b
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/19811397
15:45 travis-ci left 16:04 mtk joined
rurban Fixed now all -O2 tests 16:11
dalek rrot/rurban/imccO2: cf85d51 | rurban++ | / (3 files):
[imcc] Fixed last remaining -O2 issue #1044 t/pmc/exception-old_19.pir

Stop constant propagation at yield or invokecc for possible push_eh/pop_eh non-local effects. This can undo the set of the constant value.
See t/pmc/exception-old_19.pir: cont = test() contains a push_eh, and the subsequent pop_eh at cont() changes the value of the constant redux from 0 back to 1. So constant propagation for redux=0 cannot delete "if redux goto done" and the done branch.		 16:38
rurban tested ok on linux and darwin. nothing arch specific in it anyway 16:45
dalek rrot: cf85d51 | rurban++ | / (3 files):
[imcc] Fixed last remaining -O2 issue #1044 t/pmc/exception-old_19.pir

Stop constant propagation at yield or invokecc for possible push_eh/pop_eh non-local effects. This can undo the set of the constant value.
See t/pmc/exception-old_19.pir: cont = test() contains a push_eh, and the subsequent pop_eh at cont() changes the value of the constant redux from 0 back to 1. So constant propagation for redux=0 cannot delete "if redux goto done" and the done branch.		 16:46
rrot: 1d972e3 | rurban++ | .gitignore:
[cage] add .gdbinit and .lldbinit to .gitignore
17:01 travis-ci joined
travis-ci [travis-ci] parrot/parrot#961 (master - 1d972e3 : Reini Urban): The build is still failing. 17:01
[travis-ci] Change view : github.com/parrot/parrot/compare/f...972e3d4a36
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/19817028
17:01 travis-ci left
Coke rurban: So if our users had been using O1 or O2 before this, there was a chance that their programs would be buggy? 17:09
rurban -O1 and -O2 was disabled 17:39
I just enabled it again this week
I'm out this weekend. 17:40
Later I want to benchmark rakudo-p -O2
I got only 3-5% 17:41
Coke ah. so I assume it was known buggy and disabled. Thanks for cleaning this up. 17:44
dalek rrot: 40badac | rurban++ | MANIFEST.SKIP:
[cage] regenerate MANIFEST.SKIP
 17:51
rurban yes, exactly. similar with our old jit.
and the exec runcore and maybe more such things
18:07 travis-ci joined
travis-ci [travis-ci] parrot/parrot#962 (master - 40badac : Reini Urban): The build is still failing. 18:07
[travis-ci] Change view : github.com/parrot/parrot/compare/1...badacb1357
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/19821465
18:07 travis-ci left, mtk joined 18:39 mtk joined 19:11 Chirag__ joined 19:40 TonyC joined 19:53 TonyC joined 20:11 mtk joined 20:16 rurban1 joined 21:35 jsut joined 21:44 rurban1 joined 21:59 rurban1 joined 22:00 denis_boyun joined 22:03 rurban1 joined 22:07 denis_boyun joined 22:09 TonyC joined 22:35 denis_boyun joined 22:36 rurban1 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!