01:36 zakharyas joined 01:48 ilbot3 joined 06:19 domidumont joined 06:24 domidumont joined 09:05 zakharyas joined 09:20 domidumont joined 11:32 geekosaur joined 12:57 zakharyas joined 13:56 pyrimidine joined 15:39 zostay joined 15:40 FROGGS joined 15:50 domidumont joined
FROGGS is able to reproduce github.com/MoarVM/MoarVM/issues/418 15:50
gah, that's silly -.- 15:53
FROGGS--
mst FROGGS: be happy 15:55
generally the stupidest mistakes are the easiest to fix
it's intelligent mistakes taht are terribad to unfuck
FROGGS aye
that's what I am facing here :o) 15:56
mst ah. I assumed "that's silly" meant it was a sufficiently stupid mistake to be easy
ah well
FROGGS aye
I also tell my colleagues from time to time to be not to clever when writing code...
because I dont want to debug "clever" stuff in the middle of the night 15:57
so, here I am facing a simple problem... I just forgot to care about something
nine Soooo...what's the reason? 18:20
arnsholt I spent more or less all day yesterday and up to lunch today figuring out why the MIME envelope my code output was incorrect 18:26
Turns out that when you say boundary="$sep" in the header, the parts are separated by --$sep
Except that was really hard to discern from the example I had, given that $sep started with four dashes already 18:27
timotimo oh
arnsholt >.<
timotimo ;(
nine Great example!
arnsholt Serves me right for not reading the spec properly right away =)
And I'm yak shaving, reimplementing our HTTP response parser
The joys of working in an environment where libraries for this kind of thing can't necessarily be expected =) 18:28
FROGGS arnsholt: if boundary is "a", then you have one or more "--a" and an "--a--" at the end btw 18:41
nine: I'll push in a few minutes 18:42
arnsholt FROGGS: Oh, I didn't know about the --a-- one
Is that optional? 18:43
The MIME parser I'm looks like it accepts only a --a at the end, anyways =)
FROGGS arnsholt: example: gist.github.com/FROGGS/9a425efc09a...d557c6d8ef 18:44
stackoverflow.com/questions/4656287...e-boundary 18:45
seems not to be optional
arnsholt Indeed. Maybe the parser I'm talking to is lenient, because this is a common misencoding or something =) 18:46
Or it could be a bug in the parser. I think it's a lazy pull parser, so it might not trigger the check for correct termination 18:47
dalek arVM: 17f33bf | FROGGS++ | src/core/nativecall_libffi.c:
handle C++ constructors on libffi, closes #418
 18:49
FROGGS yeah, I guess parsers are very tolerant in that area, so that things "work"
geekosaur they have to be, yes. there's far too much garbage created by too many mail clients 18:52
FROGGS yeah, it the same with electronics: be pedantic when outputting stuff and be tolerant when working with inputs 18:53
psch i think that's an old jargon file thing, isn't it? 18:54
or maybe an RFC, not sure
moritz but don't exagerate, otherwise you'll interpret error messages as flight data and explode Ariane 5
FROGGS :S
psch ah, Postel's law 18:55
there's trevorjim.com/postels-law-is-not-for-you/ arguing against it 18:56
also already old...
geekosaur yep. mime, and later html, are actually examples of how postel's law goes wrong
psch it does kind of fit with MIME-headers..?
geekosaur once you combine it with the tragedy of the commons,
dalek arVM: 0475ede | niner++ | src/core/frame.c:
Fix uninitialized arg_names of an MVMCallsite

Could be the reason for
  ==12118== Conditional jump or move depends on uninitialised value(s)
  ==12118== at 0x40BA2D0: MVM_callsite_destroy (callsite.c:100)
reported by dogbert17++		 19:04
arVM: 2ea6426 | lizmat++ | src/core/frame.c:
Merge pull request #419 from niner/master

Fix uninitialized arg_names of an MVMCallsite
arnsholt geekosaur: This is over HTTP not SMTP (SOAP with MTOM, if you must know), but yeah. I suspect there's an element of Postel involved 19:05
nebuchadnezzar FROGGS++ thanks for #418 19:28
I'm testing the build right now, lets get a break for dinner 19:29
FROGGS awesome!
nwc10 good *, FROGGS 19:31
dalek arVM: d912378 | niner++ | src/6model/reprs/CUnion.c:
Fix CUnion::get_attribute treating inlined attrs as pointers

valgrind reports a Conditional jump or move depends on uninitialised value when checking if cobj is not NULL. That's because inlined attributes are not actually pointers but inlined values which may not have been initialized. The attribute is still there though. Easily visible because the handlers for the inlined cases don't actually use cobj. Fixed by moving the inlined cases up and handling them separately.		 19:32
arVM: 542b899 | lizmat++ | src/6model/reprs/CUnion.c:
Merge pull request #420 from niner/master

Fix CUnion::get_attribute treating inlined attrs as pointers 3297508 | niner++ | src/core/nativecall_dyncall.c: Fix 32 bit issue with rw args in NativeCall callbacks
We allocated the size of a pointer for storing a 64 bit integer. Reported by valgrind as Invalid write of size 4 in MVM_nativecall_invoke
  (nativecall_dyncall.c:491)
FROGGS hi nwc10 19:33
arVM: c8d33ae | FROGGS++ | src/core/nativecall_dyncall.c:
Merge pull request #421 from niner/master

Fix 32 bit issue with rw args in NativeCall callbacks
arVM: 0bdbd6e | FROGGS++ | src/core/nativecall_libffi.c:
use better throw away type for void nativecalls

The mem before (values[body->num_args]) is not allocated. valgrind++ for spotting it.		 20:12
21:18 travis-ci joined
travis-ci MoarVM build failed. Tobias Leich 'handle C++ constructors on libffi, closes #418' 21:18
travis-ci.org/MoarVM/MoarVM/builds/165617900 github.com/MoarVM/MoarVM/compare/a...f33bf03586
21:18 travis-ci left
timotimo nine: thanks for all those! 21:23
22:13 travis-ci joined
travis-ci MoarVM build failed. lizmat 'Merge pull request #419 from niner/master 22:13
travis-ci.org/MoarVM/MoarVM/builds/165622323 github.com/MoarVM/MoarVM/compare/1...a64267e2e6
22:13 travis-ci left
timotimo travis is failing hard 23:29
23:37 travis-ci joined
travis-ci MoarVM build passed. lizmat 'Merge pull request #420 from niner/master 23:37
travis-ci.org/MoarVM/MoarVM/builds/165629390 github.com/MoarVM/MoarVM/compare/2...2b8995ce31
23:37 travis-ci left

Powered by Cro and the Raku® Programming Language.

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

Thank you!