pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, elf: etc.) (or perl6: for all) | irclog: irc.pugscode.org/
Set by Tene on 29 July 2008.
meppl good night 00:04
pugs_svn r21740 | lwall++ | [STD] much improved error messages 03:21
r21740 | lwall++ | [methods_on_typed_attrs] class Bar must be predeclared
r21741 | audreyt++ | * Fix Paths_Pugs and Pugs.Prim.Match to look at sourcedir first, and then 03:26
r21741 | audreyt++ | installdir.
audreyt @tell gaal fixed sorry 03:28
lambdabot Consider it noted.
pugs_svn r21742 | Auzon++ | [gsoc_spectest] moved counted.t to spec, added a few more tests. (added 6 tests) 04:49
Auzon vacation (well, tomorrow) &
pugs_svn r21743 | lwall++ | [STD] previous change lost bits of syntax tree 04:53
TimToady (parses 94% of t/ currently)
gaal audreyt: should look in $sourcedir </> blib6/pugs/perl5/lib, I think... 05:16
lambdabot gaal: You have 1 new message. '/msg lambdabot @messages' to read it.
audreyt gaal: hm
lambdabot audreyt: You have 3 new messages. '/msg lambdabot @messages' to read them.
audreyt mind if I ask why :)
I know it's what we used to do
gaal well, my own sourcedir doesn't have HsBridge :) 05:17
audreyt Pugs.Prim.Match currently has
pwd1 <- io $ getDataFileName "perl5/Pugs-Compiler-Rule/lib"
doesn't HsBridge live somewhere there? 05:18
gaal beh, my pbotutil isn't configured on this machine....
sec 05:19
pugs_svn r21744 | audreyt++ | * Use ./Setup to configure/build, not cabal. 05:21
r21744 | audreyt++ | * Also make "upload-smoke" synonym to "smoke-upload"
gaal I'm not understanding this :/ 05:22
the error message is: 05:23
Can't locate Pugs/Runtime/Match/HsBridge.pm in @INC
and the @INC contains the .cabal dir but not sourcedir
... the unshift @INC lines add the cabal dir
though the code looks like it oughta work 05:24
(incidentally, do you have to rm pugs too before make?)
also also: did you mean to create _cabal </> "Pugs-" ++ versnum and return _cabal </> "share" </> "Pugs-" ++ versnum ("share" only in one of these)? 05:25
audreyt ok, a sec 05:26
pugs_svn r21745 | audreyt++ | * Fix paths of Paths_Pugs 05:31
gaal looking...
audreyt++ :) 05:37
pugs_svn r21746 | gaal++ | * minor golfage and remove a debug print
gaal though I can't say I really understand "doesDirectoryExist $ dir </> fn" there 05:38
I mean, why do we </> fn to the dir?
audreyt it checks if the source dir exists
i.e. if we are running a installed (prepackaged) pugs
the sourcedir might not be there
maybe it's an empty directory 05:39
so the dir will be there but not its content
it's just paranoia.
ok with r21747 pugs/rule should work again
bbiab
pugs_svn r21747 | audreyt++ | * build Pugs-Compiler-Rule with pugs 05:40
gaal I'd better go to work.....
pugs_svn r21748 | audreyt++ | * Remove NOTES-pugs-ghc681 as it's out of date. 07:26
r21748 | audreyt++ | (Also, test if git-svn dcommit works for this commit.)
r21749 | lwall++ | [STD and t/] various bugs, STD now parses 96% of t/ 07:37
pugs_svn r21750 | audreyt++ | * Remoge Pugs.sln as Visual Haskell has bitrotten. 08:17
pugs_svn r21751 | pmurias++ | [smop] simple method call works with Mold 08:31
pmurias ruoso: hi 08:33
how does returning values work in smop? 08:34
either setr store the result of the previously evaluated node or the SMOP_DISPATCH returns a value? 08:35
s/store/strores/
pugs_svn r21752 | audreyt++ | * Makefile.PL: Remove config.yml from build requirements for now. 08:50
pmurias s/storores/stores/ 08:51
pugs_svn r21753 | pmurias++ | [m0ld] renamed REAME to README (typoes--) 09:05
ruoso pmurias, SMOP_DISPATCH returns a value so that when that node doesn't recurse, the return value can be taken.... 09:16
but setr is used when you recurse and need to return the value to a previous frame
(then SMOP_DISPATCH returns immediatly) 09:17
(and usually returns SMOP__NATIVE__bool_false)
ruoso pmurias, that allows special casing objects that are allowed to recurse in the C stack (interpreter calls, frame calls etc) 09:20
meaning... 09:21
when you create a frame, or set a frame for running, you're recursing in the C stack and you can't use setr because you're expecting the return from the C stack...
that's why it's the code that creates a new frame that should "setr" on the back frame... 09:22
and the "eval" code of the continuation that should store the result of some specific call in its internal strucuture (in slime, in the slime node "result" attribute) 09:23
pmurias ruoso: i'm just storing it in a register 09:26
pugs_svn r21754 | pmurias++ | [smop] 09:36
r21754 | pmurias++ | Mold - return values, proper refcounting for register content
pmurias ruoso: how should i describe Mold on the wiki is a few examples and explaining what do the 4 opcodes do enough? 09:38
s/decribe/document/ 09:39
ruoso: if you have any questions/concerns about mold feel free to ask and i'll expand the wiki page 10:11
pugs_svn r21755 | pmurias++ | [smop] 10:15
r21755 | pmurias++ | target moved to the first parameter of the call opcode in Mold as it corresponds to $target = ...
ruoso pmurias, you could throw away unused return values immediatly 10:21
pmurias, that is not done in slime just because it's too lame
pmurias, in sm0p you didn't have a way to say a value wasn't being used anyomore... but m0ld have a way for you to know that... 10:23
pmurias, btw... the wiki page is just fine... 10:24
ruoso will come around from time to time...
pmurias ruoso: you mean having a /dev/null register? 10:36
ruoso pmurias, basically yes... but you just need to have a statement without an assignement 10:42
pmurias having contexts would be a different way to solve it 10:49
ruoso pmurias, contexts can be used as an optimization later... 11:41
pmurias, can mold frames be reused? 12:00
oh... now I see it... 12:01
it's the mold "code" that you reuse when creating the frame
pmurias++ # very clean and neat implementatoin 12:02
ruoso beach & 12:06
pmurias is it acceptable haskell style to mix do with >>? 14:25
Arathorn hi all - i have a question about smop: has there been any thought about how the threading semantics of the various embedded runtimes might interact? 14:55
Arathorn whilst i think i understand the idea of a central event loop using polymorphic eval to dispatch invocations to different runtimes 14:56
i don't see how multiple threads within those runtimes could ever return objects to SMOP sanely - especially if SMOP itself was running those runtimes concurrently 14:57
Arathorn gets tied up in concurrency at both the object & meta-object level and starts to boggle
audreyt @tell ruoso thanks for pushing me over the edge to switch to git -- pugs.blogs.com/pugs/2008/08/short-notes-on.html 15:08
lambdabot Consider it noted.
renormalist /quit 17:48
ups
pugs_svn r21756 | pmichaud++ | S29-str/substr.t: 8 new passing spectests for rakudo 18:35
fooandroo pmurias: yes, if you don't overdo it. One place it makes perfect sense is in on-the-fly parsers 20:19
fooandroo for example, do char '\\' ; choice [ char 'n' >> '\n', char 't' >> '\t' ] 20:20
I'm probably missing two "return"s there, but you get the idea. 20:21
pmurias fooandroo: if you are using semicolons in do you can as well to choice [do {char 'n';return '\n'},do {char 't';return '\t'}] 20:55
fooandroo you can use layout to make this pretty: 21:02
char '\\'
choice
[ .... >> .... 21:03
, .... > ....
, .... >> ....
]
s/ > / >> /
you can certainly use do inside the choice, but >> possibly makes it prettier. 21:04
pugs_svn r21757 | pmurias++ | [m0ld] comments,integer constants 21:12

Powered by Cro and the Raku® Programming Language.

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

Thank you!