Parrot 3.4.0 released | parrot.org | Log: irclog.perlgeek.de/parrot/today
Set by moderator on 17 May 2011.
00:01 plobsing joined 00:03 theory joined 00:07 theory left 00:12 plobsing left 00:15 plobsing joined
whiteknight nqp doesn't appear to generate 'file' annotations 00:24
I only see 'line' annotations
pbc_merge seems to kill annotations 00:27
cotto_work lta 00:29
whiteknight if I compile all my winxed files down to pbc individually, and pbc_merge them together, the resulting bytecode files don't contain annotation information
but if I compile down to pir, concat together the pir, and compile that to bytecode, it does have annotation information
plobsing ~~ 00:32
mikehh ok I have a weird one - rakudo test.pir fails in the build with --optimize in both i386 and amd64 BUT builds without --optimize
cotto_work whiteknight: it doesn't load annotations. 00:33
nopaste "mikehh" at 192.168.1.3 pasted "rakudo test.pir fails to builkd in --optimize builds" (27 lines) at nopaste.snit.ch/50775 00:39
mikehh it failed --optimized builds with gcc and g++ on Ubuntu 11.04 i386 but build with g++ without --optimize on both i386 and amd64 00:41
cotto_work mikehh: thanks for finding that. It looks like it's my fault and related to something fperrad++ found in lua too. 00:48
mikehh cotto_work: sorear said that jnthn__ thought that it was a missing write barrier 00:54
cotto_work mikehh: it works fine when I revert my change
looks like there's a strong correlation. 00:56
mikehh you could say that :-}
cotto_work I'll get my reverting hat. 00:57
01:03 davidfetter left
cotto_work oh. there it is 01:03
whiteknight NotFound:
NotFound: ping
sorear mikehh: for me, it manifested as a segfault
dalek rrot: a680d04 | cotto++ | src/call/pcc.c:
Revert "add experimental fix for a bug found by jnthn++"

This reverts commit ee17ddde9f878e9d90a896cefcf871bceea27824.		 01:04
cotto_work sorear: are you on x86 or x86_64? 01:09
whiteknight NotFound: unping
sorear cotto_work: yes 01:12
(x86 specificaly)
cotto_work sorear: ok. I haven't seen that bug on x86_64, which makes me suspicious. 01:13
sorear I haven't tested in over a week. Might not be the same issue at all
mikehh cotto_work: that worked, rakudo builds ok now, running tests 01:27
dalek sella: f04d754 | Whiteknight++ | s (2 files):
the bootstrapped setup.winxed fails after recent winxed updates. Add some functions for getting a friendlier backtrace using annotations
 01:28
sella: 2ec8d8d | Whiteknight++ | src/test/Builder.winxed:
Change Test.Builder to use the new backtrace functions so test failure reports are more pretty
sella: a8d979a | Whiteknight++ | setup.winxed:
pbc_merge doesn't keep annotations, so cut it out of the build. concat winxed files together, then compile down to pir and pbc. This preserves annotations in the generated .pbc files
sella: 4d922ca | Whiteknight++ | s (9 files):
move include_const declarations, where necessary, into 'header' files to prevent duplicates
sella: 177308a | Whiteknight++ | s (2 files):
fix problem with bootstrap and unnecessary 'using static'
sella: e090f6c | Whiteknight++ | / (2 files):
Add docs for Builder
whiteknight msg lucian I just pushed improvements to Rosella to print test failure backtraces using annotations. Should be nicer for you
aloha OK. I'll deliver the message.
01:29 whiteknight left
dalek p: 335b8f6 | pmichaud++ | / (2 files):
Add support for "config.default" file to hold default arguments to Configure.pl.
 01:43
p: 57c3ba2 | pmichaud++ | Configure.pl:
Update --help text to tell about config.default.
 01:46
rrot/m0-prototype: 7c6c1ce | cotto++ | src/m0/perl5/m0_assembler.pl:
remove some obsolete ops from the assembler
 01:47
rrot/m0-prototype: 62f35ce | cotto++ | / (2 files):
possibly make the calling conventions test work

It seems to output what I expect it to, but the PC jumps around suspiciously.
mikehh looks like there may be segfaukts with rakudo, 4 tests fail - 3 with bad plan (did not complete) 01:51
just 1 reported sub-test failure 01:52
01:56 mtk left 02:02 mtk joined
dalek rrot: 9f6220f | jkeenan++ | t/ (2 files):
Applying patch for both TT #2116 and TT #2118, in part based on patch by ligne++.
 02:32
soh_cah_toa cotto: ping 02:39
actually, i guess anybody could help me 02:44
02:44 hudnix left
soh_cah_toa question about bitmasks/flags: when you "and" a bitmask w/ a value, what does that do? does that set the mask to the value? for instance, "foo & bar". does that set the bar flag in the foo mask? 02:44
sorear no
quux = foo & bar # quux now contains all flags that are in BOTH foo AND bar 02:45
& doesn't modify either argument
02:45 jsut_ joined
sorear are you comfortable with binary numbers? 02:46
soh_cah_toa yeah
i know my bitwise operators. though i've never actually applied them to anything other than school :)
so if i did "if (quux)" then that would check that the bar flag was set in foo? 02:48
02:50 jsut left 02:52 woosley joined
soh_cah_toa "or" would set a flag and if i wanted to un-set a flag i would "and" the bitmask w/ the inverse of the flag (e.g. foo &= ~bar) 02:53
03:30 jsut joined 03:35 jsut_ left 03:57 eternaleye_ joined
cotto ~ 03:59
03:59 eternaleye left
cotto I have a spam from "Sofia Yak". 04:01
On an unrelated note, I have a release name.
04:14 soh_cah_toa left
plobsing shaved spam? 04:14
cotto It's such a great name. 04:15
05:02 eternaleye_ is now known as eternaleye
cotto msg dukeleto If you have the tuits, the m0 assembler needs to be able to support labels in set_imm. poke_caller is significantly harder to grok than it needs to be and that'd help make it more approachable. 05:33
aloha OK. I'll deliver the message.
cotto sleeps
dalek rrot: 8e0711f | pmichaud++ | / (2 files):
[nci]: Fix NCI::Utils::ncifunc -- don't free up arguments to NCI until we've established the return value.
 05:35
06:45 mikehh left 07:01 mikehh joined 07:10 dod joined 07:20 mj41 joined 08:29 janus left 09:06 ligne joined 09:07 benabik joined 09:13 AzureStone left 09:22 ambs joined 09:26 jsut_ joined 09:31 jsut left 09:33 AzureStone joined 09:35 woosley left
ambs moritz: ping 09:43
moritz ambs: pong
09:44 mtk left
ambs moritz: quick question: you know if chromatic hangs on any irc channel? 09:44
moritz seen chromatic
aloha chromatic was last seen in #parrot 117 days 15 hours ago saying "All of the core tests did pass for me, FWIW.".
ambs ok :P 09:45
:D
fair enough
moritz irclog.perlgeek.de/search.pl?channe...tic&q= seems he attends #parrotsketch more often recently
ambs: but I think it's easiest to reach him by email 09:46
ambs @wgz, right?
snet 09:49
tr/ne/en/
moritz at wgz dot org 09:50
09:51 mtk joined 10:16 AzureStone left 10:20 AzureStone joined
dalek sella: 978a6b0 | Whiteknight++ | src/winxed/repl/View.winxed:
update Repl.View to use the new backtrace functions. Also, clean up the printing code
 11:17
sella: 74231ab | Whiteknight++ | t/proxy/builder/Immutable.t:
+stub test file for Proxy.Builder.Immutable
11:25 contingencyplan left 12:15 ambs_ joined, ambs left, ambs_ is now known as ambs 12:31 whiteknight joined
whiteknight good morning, #parrot 12:34
13:30 hudnix joined 13:48 plobsing left, ambs left 14:00 hercynium joined
dukeleto ~~ 14:28
msg cotto can you show me example code of what you want set_imm to look like when it understands labels? 14:29
aloha OK. I'll deliver the message.
14:37 PacoLinux left, PacoLinux joined 15:02 whiteknight left 15:05 JimmyZ joined
dalek nxed: r1049 | NotFound++ | trunk/winxed_installed.winxed:
use write to output feature in installed driver
 15:06
nxed: r1050 | NotFound++ | trunk/ (3 files):
update installable files
 15:11
rrot/with_winxed: a5ab390 | NotFound++ | ext/winxed/ (2 files):
update winxed snapshot to r1050
 15:13
15:19 klavs joined 15:29 klavs left
cotto dukeleto, something like "set_imm I0, foo", where foo is a label and gets translated the same way as it would for goto. 15:47
dukeleto cotto: a full code sample would be best 15:48
cotto dukeleto, ok. I think the best way to do it will be to separate out label, op and argument parsing. I don't relish the idea of adding another special case to m0b_bytecode_seg 15:49
It's ok if the assembler allows things that don't make sense, as long as it allows the ones that do. 15:50
filtering out things like add_i I0, foo_label wouldn't be a bad idea though. 15:51
actually, I see a problem with this approach 15:55
15:58 davidfetter joined
cotto dukeleto, gist.github.com/1017046 is what I'm thinking of, though that example wouldn't work because the goto is used incorrectly. I'll update it later. 16:00
dukeleto, if the set_imm part works, that's all I think I'll need.
16:04 dmalcolm joined 16:14 ligne left
cotto_work ~~ 16:24
16:24 dodathome joined 16:25 ambs joined 16:40 klavs joined
cotto_work aloha: clock? 16:40
aloha cotto_work: LAX: Thu, 09:40 PDT / CHI: Thu, 11:40 CDT / NYC: Thu, 12:40 EDT / UTC: Thu, 16:40 UTC / LON: Thu, 17:40 BST / BER: Thu, 18:40 CEST / TOK: Fri, 01:40 JST / SYD: Fri, 02:40 EST
cotto_work good evening, klavs
klavs eoog evening, cotto 16:41
mikehh All tests PASS (pre/post-config, make corevm/make coretest, make world/make test, fulltest) at 3_4_0-182-g8e0711f
Ubuntu 11.04 i386 (g++)
16:44 bluescreen joined 17:11 mj41 left
dalek rrot/m0-spec: 37d00ea | cotto++ | docs/pdds/draft/pdd32_m0.pod:
remove a couple of superceded ops from the m0 spec
 17:18
p: 547c431 | pmichaud++ | / (2 files):
Update --makefile-timing flag from nom (based on mberends++ patch in master).
 17:31
17:55 kurahaupo joined 17:57 kurahaupo left 17:58 mj41 joined
ambs moritz: \\o- 18:24
moritz: \\o/
atrodo cotto_work> ping 18:26
18:27 JimmyZ_ joined 18:31 JimmyZ left, JimmyZ_ is now known as JimmyZ
cotto_work atrodo: pong (might be distracted) 18:40
atrodo cotto_work> I just noticed the commit on taking out free and alloc 18:41
cotto_work> I haven't had time lately to follow m0 and just wondered what spawned that
cotto_work atrodo: they're still there, just as sys_alloc, gc_alloc and sys_free
and those ones are permanent
atrodo cotto_work> Ah, I see it now 18:42
18:43 JimmyZ left 19:02 whiteknight joined
dalek p: b02ce6f | pmichaud++ | tools/build/gen-cat.pl:
Make tools/build/gen-cat.pl utf8 aware.
 19:02
dukeleto cotto_work: can you explain to me in mere-mortal english words what "set_imm I3, do_print" in M0 is supposed to do?
whiteknight dukeleto: I assume set a pointer to "do_print" to the register I3 19:03
rohit_nsit03: ping
cotto_work dukeleto: sure. assume do_print is a label at a certain bytecode offset. Take the number of that offset, convert it to a pair of immediate values ($a*256 + $b) and use $a and $b as the arguments to the op. 19:04
PerlJam cotto_work: what is the "m" in set_imm mnemonic for? 19:05
or is it just "imm" == immediate?
cotto_work dukeleto: so if do_print is a label at offset 42, the arguments would be "set_imm I3, 0, 42"
PerlJam: right
PerlJam adjusts his mental context appropriately 19:06
cotto_work PerlJam: what did you think? I'd like to minimize the possibility for confusion. 19:07
s/think/think it was/ 19:08
PerlJam cotto_work: I was just thinking it was like in PASM where you have op_i, op_s, op_p, etc. and the i, s, and p, were mnemonic for integer, string, pmc 19:09
cotto_work PerlJam: ok. 19:10
PerlJam (that a label morphs into two values is non-obvious from what little I've seen so far though)
dukeleto is not impressed with the clarity of "set_imm I3, label" 19:16
cotto_work dukeleto: that's the first thing that came to mind. I'm open to improvements. 19:17
19:24 soh_cah_toa joined
soh_cah_toa ~~ 19:25
cotto_work dukeleto: I do want some way to use labels as data. Setting the PC of a call frame gets clunky and hard to follow if all that's available is relative offsets.
I'll think about what alternatives exist. 19:30
dalek p: 8328ee1 | pmichaud++ | src/HLL/Compiler.pm:
Fix --combine flag to HLL::Compiler.
 19:44
p: b4ae484 | pmichaud++ | tools/build/gen-version.pl:
Fix typo in tools/build/gen-version.pl .
rrot/with_winxed_2: 3bb9dd5 | NotFound++ | / (8 files):
Merge branch 'with_winxed' into with_winxed_2
 19:47
rrot: a5ab390 | NotFound++ | ext/winxed/ (2 files):
update winxed snapshot to r1050
 19:51
rrot: 3bb9dd5 | NotFound++ | / (8 files):
Merge branch 'with_winxed' into with_winxed_2
NotFound winxed snapshot is now on master. 19:53
whiteknight NotFound++
19:57 klavs left 19:58 contingencyplan joined
soh_cah_toa NotFound++ indeed. very nice 19:58
i think winxed has a very bright future ahead of it. it has a lot of potential 20:00
especially now that it's included on master 20:01
cotto_work warning: good times ahead
NotFound World domination is the goal.
soh_cah_toa ha! destroy all humans 20:02
NotFound All your servers are belong to me!
20:08 jevin left
dalek rrot: d9e105e | NotFound++ | NEWS:
add a note about winxed snapshot to NEWS
 20:13
soh_cah_toa could somebody help clarify what the Parrot_set_executable_name() function is used for? the perldoc says that it sets the name of the executable launching parrot. why would one need to call this? when is it useful? it seems kinda useless to me 20:21
NotFound soh_cah_toa: to make it available inside parrot, without needing to access global argv or something like that. 20:26
soh_cah_toa alright 20:27
20:30 whiteknight left 20:31 dmalcolm left
soh_cah_toa i'm noticing a lot of (sadly undocumented) mem_gc_allocate_* macros. i'm guessing that these family of macros allocate memory that can be garbage collected? is that right? 20:31
dukeleto where is it documented how and when Winxed-in-Parrot will be synced with Winxed trunk? 20:33
NotFound dukeleto: nowhere yet. 20:34
I'm going to add a README to ext/winxed/ 20:35
20:37 dodathome left
NotFound About how. The "when" part will be informal for a now. Standard deprecation policy should apply. 20:37
cotto_work NotFound: is there any reason to keep the with_winxed branches around? 20:56
aloha: clock? 20:57
aloha cotto_work: LAX: Thu, 13:57 PDT / CHI: Thu, 15:57 CDT / NYC: Thu, 16:57 EDT / UTC: Thu, 20:57 UTC / LON: Thu, 21:57 BST / BER: Thu, 22:57 CEST / TOK: Fri, 05:57 JST / SYD: Fri, 06:57 EST
NotFound cotto_work: No, it will be easier to create new ones if something fails. I'll delete them.
cotto_work NotFound: thanks 21:00
dalek rrot: ec21896 | NotFound++ | ext/winxed/README:
Document how to update the winxed snapshot
 21:05
p: eeb63e9 | pmichaud++ | tools/lib/NQP/Configure.pm:
Update NQP::Configure to handle Win32 oddities.
 21:06
NotFound Deleted. 21:08
cotto_work NotFound++ 21:10
21:11 bluescreen left 21:23 ambs left 21:31 hercynium left 21:41 Psyche^ joined 21:42 Patterner left, Psyche^ is now known as Patterner, mj41 left
cotto_work dukeleto: ping 23:00
23:11 bubaflub left 23:28 bubaflub joined 23:40 bubaflub left
dalek rrot/soh-cah-toa/hbdb: 153d55c | soh_cah_toa++ | / (3 files):
Added a few comments on typedefs and structs
 23:48
rrot/soh-cah-toa/hbdb: ac5eae7 | soh_cah_toa++ | / (3 files):
Began building list of commands
rrot/soh-cah-toa/hbdb: 6dcf033 | soh_cah_toa++ | include/parrot/hbdb.h:
Defined a minimal set of status flags
rrot/soh-cah-toa/hbdb: 10bbc15 | soh_cah_toa++ | src/hbdb.c:
Made some further revisions to the perldoc in src/hbdb.c
rrot/soh-cah-toa/hbdb: 4f048d4 | soh_cah_toa++ | frontend/hbdb/main.c:
Added a call to Parrot_api_set_executable_name()
23:53 whiteknight 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!