Camelia, the Perl 6 bug

IRC log for #parrot, 2011-08-31

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 soh_cah_toa kid51: ok
00:00 Coke joined #parrot
00:01 plobsing kid51: I have no idea why that would be effected by the branch
00:02 plobsing can you check what HAS_GMP says about what configure found?
00:03 kid51 Since I'm building master right now, I no longer have the Config/Generated.pm that I had in that branch.  However, I know that auto::gmp did say that GMP had been located.
00:04 kid51 In case it helps, here's link to Darwin's libtool manpage:
00:04 kid51 http://developer.apple.com/library/mac/#docume​ntation/Darwin/Reference/ManPages/man1/libtool.1.html
00:05 kid51 $ grep -n HAS_GMP lib/Parrot/Config/Generated.pm
00:05 kid51 46:  'HAS_GMP' => 1,
00:05 kid51 384:#define PARROT_HAS_GMP 1
00:05 plobsing do you actually have gmp installed?
00:05 kid51 And since branch entails few changes in configuration system, I'm confident of that.
00:07 kid51 Let's put it this way.  If I *don't* have gmp installed, then config step auto::gmp has been wrong ever since it was introduced.
00:07 plobsing ok, just covering my bases
00:08 Coke joined #parrot
00:08 kid51 auto::gmp -           Does your platform support GMP......................yes.
00:09 plobsing so then it is just a matter of the linking being wrong?
00:10 kid51 I guess so ... but since I've never had to directly use 'libtool' before, I can only guess.
00:12 plobsing it isn't directly using libtool, g++ calls it
00:13 soh_cah_toa_ joined #parrot
00:13 plobsing I don't think we can get much further if you aren't on branch, but I think I understand the problem better now
00:14 Coke_ joined #parrot
00:19 Coke joined #parrot
00:25 Coke_ joined #parrot
00:28 plobsing kid51: where is libgmp installed on your system? has it moved recently?
00:29 plobsing can you show me what the '/usr/bin/g++ .* -o blib/lib/libparrot.3.7.0.*" command looks like on master?
00:31 Coke joined #parrot
00:34 dalek parrot/all-hll-test: 9aea1cf | cotto++ | tools/dev/all_hll_test.pl:
00:34 dalek parrot/all-hll-test: factor out build commonalities, start using System::Command
00:34 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/9aea1cffe8
00:34 dalek parrot/all-hll-test: 7b29f1e | cotto++ | tools/dev/all_hll_test.pl:
00:34 dalek parrot/all-hll-test: make code less repetitive
00:34 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/7b29f1e07d
00:34 dalek parrot/all-hll-test: eb92eb7 | cotto++ | tools/dev/all_hll_test.pl:
00:34 dalek parrot/all-hll-test: start using System::Command for all commands
00:34 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/eb92eb7440
00:34 dalek parrot/all-hll-test: 29f2f6a | cotto++ | tools/dev/all_hll_test.pl:
00:34 dalek parrot/all-hll-test: install parrot, get builds working in a loggable fashion
00:34 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/29f2f6a43b
00:34 dalek parrot/all-hll-test: 6566c2e | cotto++ | tools/dev/all_hll_test.pl:
00:34 dalek parrot/all-hll-test: add partcl-nqp
00:34 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/6566c2e95c
00:34 dalek parrot/all-hll-test: d5e596e | cotto++ | tools/dev/all_hll_test.pl:
00:34 dalek parrot/all-hll-test: log output from the build
00:34 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/d5e596e8ce
00:36 cotto_work eyeballs on that are appreciated
00:36 Coke_ joined #parrot
00:37 nopaste "kid51" at 192.168.1.3 pasted "A log of a recent, successful build on Darwin" (872 lines) at http://nopaste.snit.ch/76141
00:37 kid51 That's from about 6 days ago.
00:38 kid51 t/src/checkdepend.t has a failure in master on linux; will examine later
00:40 dalek parrot/all-hll-test: 68400d6 | cotto++ | tools/dev/all_hll_test.pl:
00:40 dalek parrot/all-hll-test: add a description
00:40 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/68400d6f23
00:42 Coke joined #parrot
00:45 plobsing kid51_at_dinner: running a diff on the good vs bad commands, I see that we aren't using the flags q[-L"/Users/jimk/gitwork/parrot/blib/lib" -L/sw/lib] in the bad variant
00:46 plobsing those seem like odd things for configure to add to your link flags automatically
00:47 Coke_ joined #parrot
00:53 Coke joined #parrot
00:54 plobsing cotto_work: does Ωη;)XD count as an HLL?
00:55 plobsing it has tests (and currently passes them)
00:56 plobsing also, would a shorthand for plumage be useful? I see some duplication between it and all_hll_test.pl
00:56 nbrown joined #parrot
00:59 dukeleto nbrown++ # awesome m0 pull request
01:00 Coke joined #parrot
01:01 woosley joined #parrot
01:04 nbrown dukeleto: thanks. Glad I could help
01:07 Coke joined #parrot
01:10 dalek jsop: 28821ad | Whiteknight++ | / (2 files):
01:10 dalek jsop: Flesh out the JSObject class and runtime. Compiles, but is untested
01:10 dalek jsop: review: https://github.com/Whitekni​ght/jsop/commit/28821ad9be
01:13 Coke_ joined #parrot
01:14 herat joined #parrot
01:17 nbrown dukeleto: I got your comment about the tests, but I'm pretty sure the POD is correct. I am still using the perl assembler and that is the only perl reference I saw
01:18 nbrown dukeleto: if I missed something, feel free to point me in the right direction and I'll be glad to fix it
01:19 kid51 plobsing:  AAMOF, I was thinking about those flags while poking at that TT ticket I mentioned during #ps today.
01:19 plobsing kid51: what were you thinking about them?
01:20 Coke joined #parrot
01:20 kid51 I was studying how they come about.
01:20 kid51 We start with the fact that I compile Perl 5 from source whenever a new major version comes out.
01:20 kid51 And then, since we're on Darwin, we probe for fink and macports.
01:20 kid51 That's *roughly* what's happening.
01:21 kid51 I'd have to use Parrot::Configure::Trace to speak more precisely.
01:23 dalek jsop: f1abb49 | Whiteknight++ | stage0/runtime/jsobject.winxed:
01:23 dalek jsop: + a few more builtin methods
01:23 dalek jsop: review: https://github.com/Whitekni​ght/jsop/commit/f1abb49c98
01:23 kid51 I am now trying to compare the Generated.pm between master and branch on both Linux and Darwin.
01:24 plobsing I looked at the diff between master and branch, nothing stuck out as odd
01:25 kid51 On Linux, the only diff I get is in embed-ldflags ... which is a result of the cleanup/deduplication of that KVP I did last night.
01:27 Coke joined #parrot
01:27 kid51 Darwin is slower.  Will report in a few minutes.
01:32 Coke_ joined #parrot
01:33 kid51 plobsing: On Darwin tonight, I get no diff in Generated.pm between master and branch.
01:33 kid51 I will now try to build again.
01:35 kid51 Curiously, the only difference I observe between the two versions of Generated.pm is ... the amount of whitespace indent for all the KVPs!
01:38 bubaflub joined #parrot
01:38 Coke joined #parrot
01:44 Coke_ joined #parrot
01:45 cotto ~~
01:46 cotto plobsing, your crazy language thingy definitely counts.  I'm not using plumage to minimize dependencies.
01:48 kid51 Hmm, looks like I'm getting much farther than yesterday with branch build on Darwin.
01:49 cotto I'd love to test plumage itself.
01:49 * kid51 crosses fingers
01:49 cotto It's a valuable part of our infrastructure that hasn't been getting the attention we need to give it.
01:51 kid51 Alright, I have a hypothesis as to what happened yesterday.
01:51 kid51 In order to work on TT #1744, I had to modify the shell script I also use to wrap around Configure.pl on Darwin.
01:51 plobsing so... the branch *does* work?
01:52 kid51 That modification -- since reverted -- had an explicit CLI for --ldflags
01:52 Coke joined #parrot
01:52 kid51 And that explicit setting may have excluded GMP, which, on Darwin, I probably installed years ago via either Fink or macports
01:53 kid51 Which plausibly could have prevented the compiler from locating GMP
01:53 dalek parrot/soh-cah-toa/podds: 0d8a9af | soh_cah_toa++ | / (3 files):
01:53 dalek parrot/soh-cah-toa/podds: Replaced all occurences of 'Odius' with 'PODDS' including the branch name. Also added POD to pack functions.
01:53 dalek parrot/soh-cah-toa/podds: review: https://github.com/parrot/parrot/commit/0d8a9af860
01:54 kid51 That seems right.  'make' just completed successfully on Darwin/PPC in frontend_parrot2 branch.
01:54 kid51 Now running 'make test'.
01:54 plobsing \o/ # we're on a roll
01:54 kid51 Sorry for the confusion.   That's what happens when I try to work on 3 branches simultaneously.
01:54 soh_cah_toa whiteknight: ping
01:59 Coke joined #parrot
01:59 herat joined #parrot
02:00 rfwazoo joined #parrot
02:06 Coke joined #parrot
02:09 kid51 soh_cah_toa: Do you expect to conduct any activity in the hbdb branch anytime soon?
02:10 kid51 (I want to know if I should delete my local branch.)
02:12 Coke_ joined #parrot
02:12 soh_cah_toa kid51: i'm probably not gonna do much until the new debug segment is complete (being worked on now)
02:12 kid51 thx
02:15 RobertLJ joined #parrot
02:17 Coke joined #parrot
02:23 Coke_ joined #parrot
02:29 Coke joined #parrot
02:31 kid51 frontend_parrot2 branch: darwin/ppc: make test passes except for the previously reported failure in t/dynpmc/select.t
02:33 plobsing msg whiteknight kid51 confirms frontend_parrot2 passing on darwin
02:33 aloha OK. I'll deliver the message.
02:34 Coke_ joined #parrot
03:06 RobertLJ left #parrot
03:06 baest_ joined #parrot
03:06 Drossel joined #parrot
03:06 NotFound_b joined #parrot
03:25 jevin_ joined #parrot
03:25 jlaire_ joined #parrot
03:25 Koven joined #parrot
03:25 jsut joined #parrot
03:29 jsut joined #parrot
03:29 Koven joined #parrot
03:29 jlaire_ joined #parrot
03:29 jevin_ joined #parrot
03:29 NotFound_b joined #parrot
03:29 Drossel joined #parrot
03:29 baest_ joined #parrot
03:29 rfwazoo joined #parrot
03:29 bubaflub joined #parrot
03:29 woosley joined #parrot
03:29 nbrown joined #parrot
03:29 soh_cah_toa joined #parrot
03:29 plobsing joined #parrot
03:29 benabik joined #parrot
03:29 contingencyplan joined #parrot
03:29 cosimo joined #parrot
03:29 mls joined #parrot
03:29 preflex joined #parrot
03:29 awwaiid_ joined #parrot
03:29 TiMBuS joined #parrot
03:29 particle joined #parrot
03:29 mtk joined #parrot
03:29 bacek_at_work joined #parrot
03:29 aloha joined #parrot
03:29 snarkyboojum joined #parrot
03:29 Util joined #parrot
03:29 nopaste joined #parrot
03:29 jkitazawa joined #parrot
03:29 AzureSto_ joined #parrot
03:29 tty234 joined #parrot
03:29 perlite joined #parrot
03:29 cotto joined #parrot
03:29 silug joined #parrot
03:29 autark joined #parrot
03:29 TonyC joined #parrot
03:29 Tene joined #parrot
03:29 eternaleye joined #parrot
03:29 COKE-WORK joined #parrot
03:29 dngor joined #parrot
03:29 simcop2387 joined #parrot
03:29 PerlJam joined #parrot
03:29 Hunger joined #parrot
03:29 ttbot joined #parrot
03:29 tewk__ joined #parrot
03:29 atrodo joined #parrot
03:29 slavorgn joined #parrot
03:29 cotto_work joined #parrot
03:29 ewilhelm joined #parrot
03:29 hernan joined #parrot
03:29 dalek joined #parrot
03:29 wagle joined #parrot
03:29 knewt joined #parrot
03:29 slavorg joined #parrot
03:29 he joined #parrot
03:29 athomason joined #parrot
03:29 nine joined #parrot
03:29 sri joined #parrot
03:29 moritz joined #parrot
03:29 pjcj joined #parrot
03:29 Felipe joined #parrot
03:29 tadzik joined #parrot
03:29 ingy joined #parrot
03:29 elmex joined #parrot
03:29 TimToady joined #parrot
03:29 rblackwe joined #parrot
03:29 jjore joined #parrot
03:29 GeJ joined #parrot
03:29 kthakore joined #parrot
03:29 arnsholt joined #parrot
03:29 ascent_ joined #parrot
03:29 tcurtis joined #parrot
03:29 sorear joined #parrot
03:29 eqhmcow_ joined #parrot
03:29 jnthn__ joined #parrot
03:29 Infinoid joined #parrot
03:29 janus joined #parrot
03:29 allison joined #parrot
03:29 jtpalmer joined #parrot
03:29 szbalint joined #parrot
03:29 Maddingue joined #parrot
03:29 cxreg joined #parrot
03:33 p6eval joined #parrot
03:49 dalek ohm-eta-wink-kzd: fbae3a1 | plobsing++ | src/ometac.winxed:
03:49 dalek ohm-eta-wink-kzd: add grammar debug feature to Ωη compiler
03:49 dalek ohm-eta-wink-kzd: review: https://github.com/plobsing/ohm​-eta-wink-kzd/commit/fbae3a15e0
03:49 dalek ohm-eta-wink-kzd: 5a8a4ac | plobsing++ | / (2 files):
03:49 dalek ohm-eta-wink-kzd: avoid double function-wrapping #And grammar nodes
03:49 dalek ohm-eta-wink-kzd:
03:49 dalek ohm-eta-wink-kzd: This improves performance slightly (1 or 2 percent) while reducing noise in the generated code.
03:49 dalek ohm-eta-wink-kzd:
03:49 dalek ohm-eta-wink-kzd: Example:
03:49 dalek ohm-eta-wink-kzd:     [#Or, [#And ...], [#And ...]]
03:49 dalek ohm-eta-wink-kzd: now becomes
03:49 dalek ohm-eta-wink-kzd:     self._or((function (var self) { ... }),
03:50 dalek ohm-eta-wink-kzd:      (function (var self) { ... }));
03:50 dalek ohm-eta-wink-kzd: rather than
03:50 dalek ohm-eta-wink-kzd:     self._or((function (var self) { return (function (var self) { ... })(self); }),
03:50 dalek ohm-eta-wink-kzd:      (function (var self) { return (function (var self) { ... })(self); }));
03:50 dalek ohm-eta-wink-kzd: review: https://github.com/plobsing/ohm​-eta-wink-kzd/commit/5a8a4ac00f
04:03 soh_cah_toa msg whiteknight i took care of the structs and what not. also renamed branch to soh-cah-toa/podds. what's next?
04:03 aloha OK. I'll deliver the message.
04:43 preflex_ joined #parrot
04:44 logie joined #parrot
04:52 benabik left #parrot
04:52 benabik joined #parrot
05:28 SHODAN joined #parrot
05:34 fperrad joined #parrot
05:44 AzureStone joined #parrot
06:39 cotto plobsing, does Ωη;)XD depend on anything that's not installed by the default Parrot install?
06:49 dalek parrot/all-hll-test: eff654d | cotto++ | tools/dev/all_hll_test.pl:
06:49 dalek parrot/all-hll-test: add more projects, delete some debugging output
06:49 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/eff654dbfc
07:29 mj41 joined #parrot
07:56 lucian joined #parrot
08:13 dalek parrot/all-hll-test: bc1305a | cotto++ | tools/dev/all_hll_test.pl:
08:13 dalek parrot/all-hll-test: start to (brokenly) keep track of the state of all the builds
08:13 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/bc1305a322
09:42 ambs joined #parrot
10:55 woosley left #parrot
12:10 mtk joined #parrot
12:30 JimmyZ joined #parrot
12:34 plobsing cotto: it depends only on winxed
12:36 whiteknight joined #parrot
12:37 whiteknight good morning, #parrot
12:39 bluescreen joined #parrot
12:47 JimmyZ_ joined #parrot
12:52 lucian joined #parrot
13:12 mtk joined #parrot
13:25 redicaps joined #parrot
13:40 bluescreen joined #parrot
13:41 jsut_ joined #parrot
13:58 jsut joined #parrot
14:11 benabik o/
14:12 tadzik o/
14:18 whiteknight hello benabik, tadzik
14:18 benabik 'lo whiteknight.  What's up?
14:20 contingencyplan joined #parrot
14:23 whiteknight not much, benabik. you?
14:23 JimmyZ whiteknight: jnthn++ is back
14:23 benabik Not much here either.  Kinda just killing time until school starts.
14:23 JimmyZ whiteknight: FYI
14:23 whiteknight JimmyZ: yes, I saw him yesterday
14:32 plobsing whiteknight: after kid51 gave us the all-clear, are we go for merge on frontend_parrot2?
14:32 whiteknight plobsing: I opened a pull request on github to try to get some input
14:32 whiteknight none yet
14:37 moritz should I test rakudo on top of frontend_parrot2?
14:38 plobsing moritz: please do
14:38 * moritz does
14:39 whiteknight it shouldn't make any difference for programs using fakecutables like the perl6 binary
14:39 whiteknight only when run with the parrot binary
14:41 moritz well, we do that in the compilation step once or twice
14:41 plobsing if you run rakudo as 'parrot perl6.pbc @ARGS', you may see a difference
14:48 dmalcolm joined #parrot
14:49 whiteknight startup time should be a little faster, but probably not noticable
14:51 moritz frontend/parrot2/main.o: In function `main':
14:51 moritz main.c:(.text+0xa8c): undefined reference to `get_program_code_size'
14:51 moritz main.c:(.text+0xa94): undefined reference to `get_program_code'
14:51 moritz should I have done a realclean before switching the branch? :-)
14:52 plobsing possibly
14:53 * moritz swings the 'git clean -xdf' cudgel
14:55 ambs joined #parrot
14:56 atrodo moritz> that's my favorite cudgel
14:56 moritz I still get the same error
14:56 atrodo i also, enjoy using the word cudgel when i can.  Will now work that into all my conversation for the rest of the day
14:56 moritz thi is on gcc 4.4.5
14:56 moritz *this
14:59 logie joined #parrot
15:04 dukeleto moritz: https://github.com/leto/Util/commit/5b​ba56b023c25a135088bbdf2732ad478b97233d
15:04 tadzik moritz: realclean nqp too
15:05 tadzik well, parrot, then nqp, then rakudo
15:05 tadzik that helped me
15:06 * plobsing hasn't defined a 'make clean' rule for Ωη since 'git clean' works great
15:07 moritz huh? it's a parrot compilation error, not nqp related at all
15:07 plobsing hmmm
15:08 NotFound moritz: delete that file
15:09 * moritz hopes he doesn't have to delete files manually in the build process in future
15:09 NotFound Uh... no, delete frontend/parrot2/prt0.pir.c
15:09 dodathome joined #parrot
15:10 moritz too late :-)
15:10 moritz NotFound: that file doesn't exist
15:10 NotFound Then no need to delete ;)
15:10 NotFound Have you pulled the branch?
15:11 moritz it's on commit RELEASE_3_7_0-123-g25edcb4
15:11 moritz which seems to be up to date
15:12 moritz make: Nothing to be done for `all'.
15:12 moritz make install -> same error as before
15:12 moritz so... 'make install' tries to compile more files?
15:12 moritz that seems broken
15:13 plobsing I haven't tried installing it yet. I'll try to reproduce that.
15:13 * moritz declares the branch not yet ready for prime time
15:15 plobsing make install wfm
15:15 plobsing did you run make before make install?
15:15 moritz yes
15:16 moritz it says: make: Nothing to be done for `all'.
15:16 moritz (after I deleted frontend/parrot2/main.o)
15:16 plobsing as to why make install compiles more files, we need to set the paths differently for the installable executables
15:17 moritz and why can't we do that during 'make'?
15:17 moritz the problem is that 'make install' often runs as root, while 'make' not
15:17 plobsing I don't know. seems like it could be either or.
15:18 plobsing patches welcome ;)
15:18 moritz so if 'make install' generates more files, it leaves behind some files that are owned by root
15:18 moritz which means that subsequent working in that directory becomes rather nasty
15:19 plobsing that makes some sense
15:19 moritz fwiw I configured with custom --prefix= and --optimize
15:19 plobsing I've never had that issue because I never install parrot to locations that require root
15:20 moritz I usually don't either, but our distributors usually do :-)
15:20 plobsing I thought they ran it under fakeroot
15:20 moritz I know that Debian does
15:21 moritz at least by default
15:21 plobsing run make install as root?
15:21 plobsing why would they do that?
15:21 moritz but some distributions are more source centric
15:21 moritz no, debian uses fakeroot
15:21 moritz LFS (Linux From Scratch) doesn't
15:22 moritz afk &
15:37 plobsing moritz: can you run 'nm frontend/parrot2/{main,prt0.pbc}.o' and nopaste the output? I suspect it is an issue of symbols not being exported
15:38 nopaste "moritz" at 192.168.1.3 pasted "nm output for plobsing++" (77 lines) at http://nopaste.snit.ch/76293
15:41 cosimo joined #parrot
15:44 alester joined #parrot
15:52 plobsing moritz: which compilation step is it that gives you the missing symbol errors?
15:55 moritz plobsing: cc -o installable_parrot \ frontend/parrot2/main.o src/install_config.o src/longopt.o \ -Wl,-rpath="/home/moritz/tmp/rakudo/install/lib" -L"/home/moritz/tmp/raudo/parrot/blib/lib" -lparrot -ldl -lm   -L/usr/lib  -licuuc -licudata  -ldl -l    -lnsl -ldl -lm -lcrypt -lutil -lpthread -lrt -lgmp -lreadline  -lffi   -fstck-protector -L/usr/local/lib -Wl,-E
15:56 plobsing yeah, that needs to include prt0.pbc.o somehow...
15:56 alester joined #parrot
15:59 alester_ joined #parrot
16:02 plobsing Looking at callgrind benchmarks, I really wish I could tell the C compiler about complex invariants (such as "Parrot_gc_fixed_size_allocator_free will not alter with pmc->vtable->attr_size")
16:07 dalek parrot/all-hll-test: a884e3e | cotto++ | tools/dev/all_hll_test.pl:
16:07 dalek parrot/all-hll-test: (probably) fix project/stage exit status tracking
16:07 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/a884e3e5b8
16:11 plobsing moritz: should be fixed now
16:11 plobsing well, after my last push
16:12 dalek parrot/whiteknight/frontend_parrot2: ede0f47 | plobsing++ | config/gen/makefiles/root.in:
16:12 dalek parrot/whiteknight/frontend_parrot2: add prt0 to installable_parrot build
16:12 dalek parrot/whiteknight/frontend_parrot2: review: https://github.com/parrot/parrot/commit/ede0f47349
16:16 dalek parrot/whiteknight/frontend_parrot2: 0d7014c | plobsing++ | frontend/parrot2/build.pir:
16:16 dalek parrot/whiteknight/frontend_parrot2: static up encapsulated vars
16:16 dalek parrot/whiteknight/frontend_parrot2: review: https://github.com/parrot/parrot/commit/0d7014c44d
16:23 plobsing moritz: can you confirm that install works for you now?
16:23 mtk joined #parrot
16:44 mtk joined #parrot
16:53 mj41 joined #parrot
17:00 jsut_ joined #parrot
17:05 dalek parrot: b929480 | plobsing++ | src/gc/gc_gms.c:
17:05 dalek parrot: manually enforce invariant
17:05 dalek parrot:
17:05 dalek parrot: Parrot_gc_fixed_allocator_free() will not modify &interp->gc_sys or &pmc->vtable->attr_size
17:05 dalek parrot: Using local const variables informs the compiler of this invariant which allows it to avoid
17:05 dalek parrot: redundant loads, leading to a 12.5% reduction in cost (0.05% of Ωη total).
17:05 dalek parrot: review: https://github.com/parrot/parrot/commit/b9294804f1
17:22 dukeleto plobsing++ # nicely explained optimization
17:41 cotto_work ~~
17:49 plobsing dukeleto: that was mostly a PoC optimization (it improved the benchmark totaly by insignificant amounts)
17:49 plobsing what it does show is that we can get significant improvements if we can do this sort of thing accross the codebase
17:49 plobsing I'm going to look into automated theorem provers for C to automatically find these kinds of invariants
17:51 moritz plobsing: yes, install works now. Testing rakudo...
18:17 dukeleto plobsing: yes, it sounds like a promising thing to look into. Perhaps we can make macros that make writing faster code easier than writing slower code
18:26 whiteknight that edit is probably not amenable to macroization
18:36 plobsing agree
18:36 plobsing it requires deep knowledge of remote areas of the codebase
18:40 dalek parrot/all-hll-test: eb7a585 | cotto++ | tools/dev/all_hll_test.pl:
18:40 dalek parrot/all-hll-test: better status tracking, less verbose output
18:40 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/eb7a585d7a
18:41 dukeleto whiteknight: even if that edit can't macroized, some others probably can. Anything we can get.
18:43 plobsing dukeleto: macros tend to act locally, which is when C compiler optimizers are strongest
18:44 Eclesia joined #parrot
18:45 Eclesia hi
18:45 plobsing dukeleto, cotto_work: it may be a good idea to consider how Lorito will manage invariants such as this.
18:50 cotto_work dukeleto: could you eyeball tools/dev/all_hll_test.pl in the all-hll-test branch?
18:52 cotto_work dukeleto: my goal is to show that an HLL's tests pass on at least one platform, so I don't care too much about making it portable (for now).
18:58 Eclesia question : when I have an exception, how do I get the instruction line and file which raise it ?
18:59 plobsing Eclesia: ex.backtrace or ex.backtrace_strings
19:00 Eclesia plobsing: I use this but there is one line missing
19:00 plobsing oh? I've run into this problem elsewhere but wasn't able to golf it down. do you have a relatively toy example?
19:01 Eclesia I can try to make one
19:11 dalek jsop: 1365216 | Whiteknight++ | / (5 files):
19:11 dalek jsop: Fixed x['foo'] syntax to generate similar winxed syntax.
19:11 dalek jsop: review: https://github.com/Whitekni​ght/jsop/commit/1365216799
19:11 dalek jsop: a98377c | Whiteknight++ | stage0/js/nodes.js:
19:11 dalek jsop: Fix binary operators so they generate correct winxed code
19:11 dalek jsop: review: https://github.com/Whitekni​ght/jsop/commit/a98377cfbb
19:11 dalek jsop: 3b092bb | Whiteknight++ | stage0/js/ (2 files):
19:11 dalek jsop: Get unary operators working
19:11 dalek jsop: review: https://github.com/Whitekni​ght/jsop/commit/3b092bbadf
19:11 dalek jsop: 3453778 | Whiteknight++ | stage0/js/ (2 files):
19:11 dalek jsop: Implement while() statements and StatementBlock. Get postfix ++ and -- working. Misc fixes
19:11 dalek jsop: review: https://github.com/Whitekni​ght/jsop/commit/345377898b
19:11 dalek jsop: 99b9067 | Whiteknight++ | stage0/js/ (2 files):
19:11 dalek jsop: if, for, and for/in
19:11 dalek jsop: review: https://github.com/Whitekni​ght/jsop/commit/99b9067cc3
19:11 dalek jsop: 1be0aea | Whiteknight++ | stage0/js/ (2 files):
19:11 dalek jsop: do/while, ternary-?:, return. Break generated __main__ up to separate error-handling from the rest, and make indenting code more consistent
19:11 dalek jsop: review: https://github.com/Whitekni​ght/jsop/commit/1be0aeab69
19:11 dalek jsop: 38df42f | Whiteknight++ | stage0/js/wast.js:
19:11 dalek jsop: small formatting changes
19:11 dalek jsop: review: https://github.com/Whitekni​ght/jsop/commit/38df42f611
19:11 dalek jsop: b6bf29c | Whiteknight++ | stage0/js/nodes.js:
19:11 dalek jsop: Fix VarDecl when we have an IdPatt instead of an InitPatt
19:11 dalek jsop: review: https://github.com/Whitekni​ght/jsop/commit/b6bf29cd60
19:26 dukeleto cotto_work: i like it. Small comments are: add $$ to $tmp_dir so you can run multiple copies (perhaps with different compilers) at the same time. Also, adding in parrot-gmp would be nice, since it's build infrastructure requires winxed
19:26 dukeleto cotto_work: otherwise, awesome work!
19:27 cotto_work dukeleto: wfm
19:28 cotto_work any reason not to merge?  It's not complete, but I feel like people can start playing with it.
19:38 Eclesia plobsing: it was tricky to reproduce : http://trac.parrot.org/parrot/ticket/2188
19:40 Eclesia plobsing: In was stack the error source is at : (exptest.wx:39)   and in the other: (friend.wx:9)
19:40 Eclesia plobsing: making debugging painfull :/
19:41 dalek TT #2188 created by Eclesia++: Missing path in exception stack trace
19:41 dalek TT #2188: http://trac.parrot.org/parrot/ticket/2188
19:41 Eclesia whiteknight: it's the error I talked about a few days ago with you
19:45 nine whiteknight: that's an awful lot of threading code you get rid of in your branch...
19:46 dalek parrot/all-hll-test: fff0ace | cotto++ | tools/dev/all_hll_test.pl:
19:46 dalek parrot/all-hll-test: add parrot-gmp, make verbose output nicer, add PID to the tmp dir
19:46 dalek parrot/all-hll-test:
19:46 dalek parrot/all-hll-test: dukeleto++ for suggestion $$ and parrot-gmp
19:46 dalek parrot/all-hll-test: review: https://github.com/parrot/parrot/commit/fff0acec9d
19:46 cotto_work nine: it's cathartic
19:47 cotto_work merged and deleted
19:47 dalek parrot: eb7a585 | cotto++ | tools/dev/all_hll_test.pl:
19:47 dalek parrot: better status tracking, less verbose output
19:47 dalek parrot: review: https://github.com/parrot/parrot/commit/eb7a585d7a
19:47 dalek parrot: fff0ace | cotto++ | tools/dev/all_hll_test.pl:
19:47 dalek parrot: add parrot-gmp, make verbose output nicer, add PID to the tmp dir
19:47 dalek parrot:
19:47 dalek parrot: dukeleto++ for suggestion $$ and parrot-gmp
19:47 dalek parrot: review: https://github.com/parrot/parrot/commit/fff0acec9d
19:47 dalek parrot: f9b4167 | cotto++ | tools/dev/all_hll_test.pl:
19:47 dalek parrot: Merge branch 'all-hll-test'
19:47 dalek parrot: review: https://github.com/parrot/parrot/commit/f9b416751b
19:49 nine cotto_work: I'm quite curious if that tasklet/messaging thing will work. In theory it should be much less intrusive while still giving most or all of the benefits.
19:51 dalek parrot: f5ee33e | cotto++ | tools/dev/all_hll_test.pl:
19:51 dalek parrot: check ARGV definedness
19:51 dalek parrot: review: https://github.com/parrot/parrot/commit/f5ee33e642
19:55 plobsing nine: a good test would be to write a chameneous-redux benchmark and see that the whole thing didn't fall flat (like erlang does)
20:03 dukeleto plobsing: chameneous-redux ?
20:09 plobsing it's one of the shootout benchmarks
20:10 plobsing it is a parallel problem on which erlang is uncharacteristically slow
20:17 preflex joined #parrot
20:21 mj41 tryin to debug my failing Padre trunk ... failing probably in Wx::ScintillaTextCtrl
20:29 ttbot Parrot fff0acec MSWin32-x86-multi-thread make error http://tt.taptinder.org/cmdinfo/46859
20:31 mj41 sorry, wrong channel
20:58 bluescreen joined #parrot
21:03 dalek rakudo/nom: 2bac6a9 | (Martin Berends)++ | / (2 files):
21:03 dalek rakudo/nom: [lib/Test.pm] improve accuracy of test timing, pmichaud++ for diagnosing
21:03 dalek rakudo/nom:
21:03 dalek rakudo/nom: This change affects the results in docs/test_summary.times that are
21:03 dalek rakudo/nom: created by 'perl tools/test_summary.pl --timing'.  Prior to this
21:03 dalek rakudo/nom: commit the time calculated for each test included the time taken by
21:03 dalek rakudo/nom: print or say to produce the TAP output.  A recent apparent slowdown
21:03 dalek rakudo/nom: in test performace was traced to 'say' taking longer due to added
21:03 dalek rakudo/nom: functionality.  With this commit the time spent producing TAP output
21:03 dalek rakudo/nom: is excluded from test times, making them shorter and more accurate.
21:03 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2bac6a9877
21:27 perlite_ joined #parrot
21:44 dalek parrot: 4893b85 | cotto++ | tools/dev/all_hll_test.pl:
21:44 dalek parrot: write logs sooner, make tail -f on the stdout log work
21:44 dalek parrot: review: https://github.com/parrot/parrot/commit/4893b8537c
22:08 RobertLJ1 joined #parrot
22:22 kid51 joined #parrot
22:23 rfw joined #parrot
22:25 rurban joined #parrot
22:37 dalek parrot: 9765019 | cotto++ | tools/dev/all_hll_test.pl:
22:37 dalek parrot: don't lose status information when a stage fails
22:37 dalek parrot: review: https://github.com/parrot/parrot/commit/976501957e
22:39 kid51 t/src/checkdepend.t is giving me a FAIL for frontend/parrot2/prt0.pbc
22:39 kid51 Failed test 'frontend/parrot2/prt0.pbc.c (no rule found for this file).'
22:39 kid51 But I can't find any instance of the string prt0 in the code base.
22:42 NotFound I think realclean doesn't clean all in the frontend_parrot2 branch
22:42 pjcj joined #parrot
22:43 NotFound And then, that files create the fails when you switch to master.
22:43 dalek parrot: 1c090ee | jkeenan++ | tools/dev/all_hll_test.pl:
22:43 dalek parrot: Add note that System::Command must be installed from CPAN.
22:43 dalek parrot: review: https://github.com/parrot/parrot/commit/1c090eef7c
22:43 cotto_work what NotFound++ said
22:44 * kid51 will investigate
22:45 dalek parrot: 42e9165 | cotto++ | tools/dev/all_hll_test.pl:
22:45 dalek parrot: more succinct final output
22:45 dalek parrot: review: https://github.com/parrot/parrot/commit/42e91655a3
22:45 kid51 So, what I have in frontend/parrot2/ is a residue of something from a branch?
22:45 kid51 Has that branch been merged yet?
22:46 kid51 And, while I'm on the subject, where is this README2 file coming from?
22:46 NotFound The Select PMC tests, I think.
22:46 cotto_work kid51: it hasn't been merged
22:47 NotFound It also needs realclean attention.
22:48 NotFound t/dynpmc/select.t
22:48 NotFound $P0.'open'('README2', "w")
22:49 plobsing select shouldn't be being tested on files
22:49 plobsing select on files is not a meaningful operation
22:58 kid51 In config/gen/makefiles/root.in, in the definition of 'clean', I see Makefile variables such as:
22:58 kid51 $(RM_F) $(NQP_CLEANUPS) $(PGE_CLEANUPS) $(TGE_CLEANUPS)
22:58 kid51 But I don't see where those Makefile variables are defined or assigned to.
23:00 tadzik do we have something like open3() in Parrot?
23:01 plobsing kid51: I'd hazard those are found in included definitions files
23:03 kid51 plobsing: Correct.  They are files which 'ack' fails to detect (at least to the extent that I know how to use 'ack')
23:05 kid51 cotto_work: Given that more and more people are writing Perl programs in tools/dev/, etc. that make use of post-Perl 5.8 features, what's your receptivity to boosting Parrot's required Perl 5 version?
23:08 NotFound Why do we need more perl 5 tools? I guess we have enough languages
23:09 kid51 Example: cotto just wrote tools/dev/all_hll_test.pl
23:09 cotto_work NotFound: laziness, mostly.  Now that you mention it, it's tempting to rewrite that script in winxed.
23:09 cotto_work kid51: I'll talk to him about that.
23:09 kid51 That's not the only instance.
23:10 cotto_work kid51: I didn't bother sticking to 5.8.x for that because it's not a part of our build infrastructure, i.e. not required to get a working parrot.
23:10 kid51 But we could simplify some of our Perl 5 code by taking advantage of the core modules newly bundled with 5.10, etc.
23:11 cotto_work but I continue to say that if there's a compelling reason to bump the minimum version that outweighs the potential (and actual) disadvantages, I'm all for it
23:12 cotto_work simplifying existing code isn't a great reason unless it's been causing us some kind of pain
23:14 cotto_work I like simpler code, but not for its own sake.
23:14 NotFound If ain't broke, don't fix it.
23:14 cotto_work yes
23:16 cotto_work kid51: if bumping the version will make frequently-updated Perl 5 code nicer, that'd be a good reason.
23:19 rurban_ joined #parrot
23:43 dalek parrot/whiteknight/frontend_parrot2: 2a8073d | jkeenan++ | config/gen/makefiles/root.in:
23:43 dalek parrot/whiteknight/frontend_parrot2: Add 3 files to 'clean' targets.
23:43 dalek parrot/whiteknight/frontend_parrot2: review: https://github.com/parrot/parrot/commit/2a8073d5b9

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

Parrot | source cross referenced