Perl 6 - the future is here, just unevenly distributed

IRC log for #perl11, 2014-09-09

| Channels | #perl11 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
02:46 willthechill joined #perl11
02:56 tokuhirom left #perl11
07:20 basiliscos joined #perl11
08:16 travis-ci [travis-ci] RPerl build failed. Will Braswell says 'Namespace Hack, Prepare To Implement RPerl OO Tests Part 19'
08:16 travis-ci [travis-ci] http://travis-ci.org/wbraswell/rperl/builds/34789679 https://github.com/wbraswell/rperl/compare/5d02bdef77a9...ae76956fb965
09:07 basiliscos joined #perl11
16:08 travis-ci joined #perl11
16:08 travis-ci [travis-ci] perl11/potion#754 (master - 89b316e : Reini Urban): The build passed.
16:08 travis-ci [travis-ci] Change view : https://github.com/perl11/potion/compare/20cb1028b849...89b316e7eb2e
16:08 travis-ci [travis-ci] Build details : http://travis-ci.org/perl11/potion/builds/34828899
16:08 travis-ci left #perl11
16:35 basiliscos joined #perl11
17:32 j_ey joined #perl11
17:32 j_ey https://github.com/perl11/potion/issues/74 is so confusing
17:56 rurban thanks for the msg vs bind hint
17:57 j_ey Im.. still totally baffled :)
17:58 j_ey I closed my terminal, and repopened, and it seems to work now?
17:58 rurban -Dt helps maybe
17:58 rurban (tracing)
17:59 rurban But for me the results are consistently correct
17:59 rurban so it's a bit hard to repro. there are also similar weird failures with other tests sometimes
18:00 j_ey I cant repro it any more :(
18:00 j_ey now that I restarted my shell
18:01 rurban $ bin/potion -B -I test/classes/sub.pn => (nil, nil, 9, 81)
18:01 rurban being repro with -B is a big plus
18:01 j_ey thats the interp?
18:01 rurban -B is bytecode, yes
18:02 rurban at least I found the crazy GC bug today
18:03 rurban Are you the famous british speedcuber?
18:03 j_ey lol, yes
18:03 rurban :)
18:04 rurban well, I'd certainly could need some help and hints how to attack this problem. But I got much more time now. volatile quirks was my theory
18:05 j_ey since I can't even repro it now
18:05 j_ey Im not sure how to help
18:05 rurban compiling with DEBUG usually helps
18:06 rurban So I came to the bad gcc/clang optimizer and volatile idea. Or some some missing volatile
18:07 j_ey it's working all the time now :(
18:08 rurban thanksfully you can easily change config.inc manually, and remake it pretty fast
18:08 j_ey compile.c is from AST->vm instructions right?
18:08 j_ey where's the "entry" point?
18:08 rurban yes
18:09 rurban potion_source_asmb
18:09 j_ey thanks
18:10 rurban which is called by the compile method for AST objects
18:10 j_ey hm
18:10 rurban it's very similar to lua, just a bit easier to read I think
18:10 j_ey that seems to be used only inside compile.c
18:10 j_ey I mean something that calls *into* compile.c
18:12 rurban that would be the compile method being called in the frontend. potion.c
18:12 rurban potion_cmd_exec(P, buf, NULL, compile, "");
18:12 j_ey thanks
18:12 rurban code = potion_send(code, PN_compile, potion_str(P, filename),
18:12 rurban compile ? potion_str(P, compile): PN_NIL);
18:13 rurban code is the AST, PN_compile is the method call into the compiler
18:14 rurban the AST is of type PN_SOURCE
18:14 rurban PN_TSOURCE actually
18:14 j_ey whats the "compile ? potion_str : PN_NIL" for?
18:16 rurban that's an optional string (compiler method) being passed to the compiler backend. --compile=c for the c compiler backend (not yet)
18:16 rurban usually NULL.
18:16 rurban I want to have external compiler libs
18:18 rurban This part is not really super elegant so far
18:19 rurban a --compile option calls the dump or dumpbc method on the code then, not compile
18:20 rurban --compile=exe would be the native code compiler, which would add linker segments to the jit'ted code
18:20 j_ey ah
18:20 rurban elf and PE coff only i think. that's the plan
18:21 j_ey its a bit confusing
18:22 j_ey its like the compiler is written in potion objects
18:23 rurban And the compile.c itself just calls the asmb methods on the source and arg parts, but not on the other types so far. So it's pretty hard to maintain and extend.
18:23 j_ey what are thr "other types"?
18:24 rurban and macros are missing. they would fit into the AST_MSG part, with a hash lookup if the method is a macro or function call
18:24 rurban all the AST_ subtypes.
18:25 rurban enum PN_AST
18:26 rurban so  far I have cases in AST_MSG for some should-be-macros, like loop, while, if, else, ...
18:27 rurban if (t->part == AST_MSG && PN_S(t,0) == PN_if) { ... should really be a macro defined in some simplier language which understand the AST tree
18:30 rurban AST_MSG is the most important AST type, as almost everything is a method call. most of the keywords, like class, self, if, while, return, ...
18:30 rurban so it's a pretty functional language
18:31 rurban and each methodcall is a message being sent to the target object (like in smalltalk)
18:31 rurban so it should support threads very easily (but not yet started)
18:37 j_ey sounds cool
18:37 rurban did you come from ruby or somewhere else?
18:38 j_ey ruby was one of the first languages I learnt
18:38 j_ey where I knew _why from
18:39 rurban I never met him. I just saw his house in Provo but didn't dare to knock
18:39 j_ey heh
18:40 j_ey I was in the #camping irc channel a lot
18:40 rurban freenode? nobody there
18:41 j_ey I guess it died
18:41 j_ey it was like 5 years ago
18:51 dalek p2/p2: bcb24a3 | rurban++ | core/gc.c:
18:51 dalek p2/p2: GC: fix ptr->vt < P->vts->len check for user objects
18:51 dalek p2/p2:
18:51 dalek p2/p2: len < 100 but ptr->vt starts at PN_TNIL.
18:51 dalek p2/p2: It caused example/binarytrees.pn errors
18:51 dalek p2/p2: review: https://github.com/perl11/p2/commit/bcb24a3109
18:55 rurban could also be a strange alignment issue
18:55 travis-ci joined #perl11
18:55 travis-ci [travis-ci] perl11/p2#193 (p2 - bcb24a3 : Reini Urban): The build passed.
18:55 travis-ci [travis-ci] Change view : https://github.com/perl11/p2/compare/74c00514b1a0...bcb24a310910
18:55 travis-ci [travis-ci] Build details : http://travis-ci.org/perl11/p2/builds/34843665
18:55 travis-ci left #perl11
20:17 basiliscos joined #perl11
20:25 basiliscos joined #perl11
20:35 basiliscos joined #perl11
20:56 basiliscos joined #perl11
22:35 travis-ci [travis-ci] RPerl build passed. Will Braswell says 'Namespace Hack, Prepare To Implement RPerl OO Tests Part 20'
22:35 travis-ci [travis-ci] http://travis-ci.org/wbraswell/rperl/builds/34861797 https://github.com/wbraswell/rperl/compare/ae76956fb965...d9d2d540c0eb
22:49 willthechill joined #perl11

| Channels | #perl11 index | Today | | Search | Google Search | Plain-Text | summary