Please test pre1: pugs.blogs.com/pre1/Perl6-Pugs-6.2.12.tar.gz | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by audreyt on 24 June 2006.
dduncan but I'll say quickly ... the basis for what I'm mentioning is that we have a Relation, which is a role, that can be implemented either in memory or as a tied object with a db backend; but, if you use several together or do multiple actions on one, they can be rearranged into different actions behind the scenes that are more efficient 00:00
I will not be at YAPC, regrettably 00:01
can't afford it
the rearranging is the optimizer
they still work otherwise, but slower
if not the optimizer, then an elaborate macro
nothingmuch aye 00:02
IMHO macros-for-performance is a bad tradeoff
annotations for performance, otoh
give you both documentation
and speed
(and safety, if they're the good kind)
dduncan either way, I want to be able to implement that functionality in plain Perl 6, aside from the bare minimum of low level bindings
eg, write the optimizer in Perl 6
nothingmuch i think it's still in the distant fugture 00:03
but the compiler toolchain will be in perl 6
so the plugins for the optimizer can be perl 6 too
dduncan meanwhile, I will implement the non-optimized version that looks the same to users, but is slower, doing most work at the perl end 00:04
nothingmuch audreyt: when should i release .46? ~10 PASSes?
obra nothingmuch: should I test the snap against jifty?
dduncan once this is done ... no more SQL ... you write your queries in natural perl 6, with things that look like standard Perl functions like maps and greps and junctions and sets, as if each database table was an in-memory array or something ... that's my goal 00:05
nothingmuch obra: please do
obra url?
nothingmuch pause.perl.org/incoming/YAML-Syck-0.46_01.tar.gz 00:06
obra t/2-scalars...........ok 1/39Use of uninitialized value in concatenation (.) or string at (eval 6) line 1. 00:07
"not a big deal" 00:08
but still
audreyt nothingmuch: release it when we get f2f bandwidth
nothingmuch f2f? 00:09
audreyt face 2 face
nothingmuch ah
sounds good
obra passes tests but hates the dev version 00:10
t/TestApp/t/06-validation..............Argument "0.46_01" isn't numeric in subroutine entry at ../../lib/Jifty/YAML.pm
nothingmuch oi
obra not your problem 00:11
nothingmuch i don't understand the undefined value
but we'll see later
shower time
svnbot6 r10879 | audreyt++ | * At SyntaxNinja++'s suggestion, rename our third-party cabal 00:46
r10879 | audreyt++ | names to pugs-fps and pugs-HsSyck, to avoid clash with incompatible
r10879 | audreyt++ | versions.
svnbot6 r10880 | fglock++ | v6.pm - changed the distribution name to 'v6'; 01:25
r10880 | fglock++ | - POD says that the implementation is incomplete and tells about pugs and parrot
r10880 | fglock++ | - comments welcome - thanks!
colomon audrey, my make test results: 01:29
t/oo/proxy.t 18 3 16.67% 4 6 9
(1 subtest UNEXPECTEDLY SUCCEEDED) (1150 subtests TODO), 1497 subtests skipped.
dduncan I finished smoking pre1 ... shows a bunch of errors 01:45
see m19s28.vlinux.de/iblech/stuff/pugs-...e55c0.html
all the problems seem to be related to t/pugsrun/07-dash-uppercase-v.t just cutting out a quarter of the way through, with the other red squares saying 'stub' 01:47
I will run that test by itself now ...
run in isolation, at about test 13, it dies with the message "Bus error", whatever that means 01:50
I will nopaste a crash report now ... 01:52
pasteling "dduncan" at 24.69.53.198 pasted "crash report for ./pugs t/pugsrun/07-dash-uppercase-v.t ... may be due to memory corruption from local dirty power, though" (70 lines, 3.6K) at sial.org/pbot/17950 01:54
dduncan make of that what you will
I get nearly the same result with multiple runs
if it is corrupted system memory from local dirty power, I'll see if a restart fixes it ... 01:55
dduncan that did it 01:58
all tests pass
I think I need to prioritize the purchase of a power conditioner 01:59
so ignoring all that, we would seem to have no test fails, and one unexpected success, with pre1 02:00
audreyt good night :) 05:53
the trunk should be passing all its tests now
testing time is 10min here
25% or so faster than pre1
dduncan nice to have one of them newer computers ... only 10 minutes 06:13
pasteling "Aankhen" at 203.101.1.39 pasted "Build fails on Win32" (16 lines, 915B) at sial.org/pbot/17958 08:02
cmarcelo hi 08:13
Aankhen`` Since I don't think it showed up earlier... sial.org/pbot/17958 08:42
lambdabot Title: "Paste #17958 from "Aankhen" at 203.101.1.39"
Aankhen`` I'm having trouble building Pugs on Win32.
cmarcelo Aankhen``: from svn repo or it is the release? 08:45
Aankhen`` Repo.
r10887 08:46
cmarcelo just a sec
pasteling "cmarcelo" at 200.232.236.25 pasted "correct references to fps and HsSyck, now they are pugs-fps and pugs-HsSyck" (22 lines, 1.3K) at sial.org/pbot/17959 08:49
cmarcelo Aankhen``: look at my paste. try applying this patch OR change the proper lines in util/build_pugs.pl
Aankhen`` A'right.
cmarcelo tell me if it works. 08:50
cmarcelo spinclad: ping? 08:51
Aankhen`` cmarcelo: It built, thanks. :-) 08:58
cmarcelo great. i'll commit it. probably it wasnt noticed because everyone had fps and HsSyck already installed...
Aankhen`` I thought I did too. :-S 08:59
(I last built Pugs yesterday, which would have been about 30-40 revisions ago, if that helps...)
cmarcelo Aankhen``: do you mind running "ghc-pkg list" to see if they were available? 09:01
pasteling "Aankhen``" at 203.101.1.39 pasted "ghc-pkg list" (12 lines, 618B) at sial.org/pbot/17960 09:02
Aankhen`` "Unpacking tree node parameters" in S06 is making my brain hurt. :-( 09:05
cmarcelo Aankhen``: doesnt really help, but i'm still reading (skimming?) S02.. =o) 09:07
Aankhen`` Hehe.
I jumped into S06 a few days back to check something... have been trying to understand it ever since.
cmarcelo is svnbot6 down? 09:09
meppl gute nacht 09:52
pasteling "cmarcelo" at 200.232.236.25 pasted "patch against r10889 - use Judy.MapSL for IHash implementation" (295 lines, 11.7K) at sial.org/pbot/17962 10:24
cmarcelo audreyt: sial.org/pbot/17962 ... the interface is the same of Judy.Map (which uses JudyHS), but JHS+Iter still segfaulting. 'make smoke' result for patched took almost the same time. in theory JHS will be better on this. comments? 10:31
lambdabot Title: "Paste #17962 from "cmarcelo" at 200.232.236.25"
cmarcelo & 10:51
audreyt evalbot_10889: 1+1 11:36
?eval 1+1
evalbot_10889 2
cmarcelo audreyt: ping? 12:25
spinclad cmarcelo: pong... 17:33
wolverian heh, CPANPLUS doesn't trust jifty. 18:18
MenTaLguY hello 23:54

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!