github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
nwc10 good *, #moarvm 07:10
jnthn good morning 09:05
nwc10: Remind me, is anything blocking on me for the hash branch at this point? Will have some time to look at it soon if so.
nwc10 I'm not aware of anything blocking 09:16
I believe that I have made commits that answer your questions
jnthn releasable6: status
nwc10 *don't* merge it (as-is) because it has that commit message that starts 'XXX'
releasable6 jnthn, Next release in ≈4 days and ≈9 hours. 2 blockers. Changelog for this release was not started yet
jnthn, Details: gist.github.com/74a76b4b4b4e61b6ed...b41ae63b4e
jnthn *nod* 09:17
nwc10 and I have a suspicion that we should wait for 4 days and 10 hours before merging it :-)
the NQP tests (and implied NQP language spec) PR is also "interesting" if you have time
I was about to go out shopping. (back in an hour I guess) 09:18
we're fine for beer, but critically low on sherry
(and milk and other stuff)
jnthn oh deer
*dear
bah
nwc10 use more 'coffee' ?
jnthn Whenever I drink sherry my thoughts are "I wish this was port instead" :) 09:19
nwc10 I have some white port, but not the red stuff
Andrea's dad has ruby port, but I think that he mostly uses it for cooking.
anyway, I'm told that sherry casks are always in short supply for whisky making, so I'll keep doing my bit to help 09:20
(I have no idea how accurate this statement is. I was told it by someone who IIRC had been told it at a whisky tasting. It might be true.)
jnthn Ah, the red stuff is what I like :) 09:24
nwc10 jnthn: also, I don't know how to (so have not) tested the debug server or the heap profiler, which IIRC get touched as a side effect of hashes 11:49
jnthn Looked over the commits; agree they address all my comments. 12:30
[Coke] wonders if whisky vs. whiskey is a UK/US thing.
jnthn Hmm...I just installed App::MoarVM::HeapAnalyzer and tried to run it on the produced heap profile output and it reports malformed UTF-8 12:34
timotimo: Any idea about ^^?
nwc10 [Coke]: it's certainly a Scotland vs Ireland thing 12:38
jnthn: \o/
jnthn The debugger seems busted though 12:39
nwc10 and also, I didn't have any better suggestions for the debug server thing than I'd already put in the comment in the code (which appeared near your question to timotimo )
:-(
I don't know enough to fix that. (As in, even where to start)
jnthn Process finished with exit code 139 (interrupted by signal 11: SIGSEGV) 12:41
nwc10 I don't even know where TFM is for the debug server, so I don't know how to start it 12:42
jnthn There's a command line client but I don't know how to use it :P
nwc10 so you use netcat?
jnthn No, I use it from the IDE :)
gist.github.com/jnthn/bef407c806db...042f869c70 12:47
Seems to happen when it hits a breakpoint
Ah, well, more to the point, when it is handling a follow-up request after hitting a breakpoint
nwc10 this is a regression in the branch? 12:48
hashtable=0x0 12:49
jnthn Well, I think that thing can legitimately be null' 12:50
And indeed adding a null check there fixes things
nwc10 yes, and IIRC NULL is a "valid" UT hash
jnthn Yeah, I think so to 12:52
jnthn Or at least, not distinguishable from an empty hash 12:52
Shall I push a commit fixing it?
nwc10 yes please
Geth MoarVM/AAAA-A-better-hash: 2af3ffbf60 | (Jonathan Worthington)++ | src/debug/debugserver.c
Don't try to use a NULL debug locals hash

Under the previous hash implementation, the hash pointer being NULL was interpreted as an empty hash, and thus a lookup would just come back NULL and there was no problem. Under the new one, we should NULL-check it explicitly. Fixes a crash in the debug server.		 12:54
jnthn I looked if there's any other instances of the problem with the debug locals, but it seems that was the only one 12:55
nwc10 I had a look in that file, and couldn't spot any others 12:56
jnthn OK
I think for now that's all I have :)
nwc10 time, or bugs? 12:57
jnthn Bugs.
I have time but want to spend it on my RakuAST grant, which I barely touched in the last month, though mostly 'cus I had the audacity to go on vacation :)
moritz jnthn: speaking of which, it's mid of the month, and nominally time for the grant report 12:59
nwc10 expect a report with pictures of Swiss hills?
jnthn haha, yeah...
moritz jnthn: there's bascially two options: 1) publish a "I did nothing" grant report, or 2) we move the grant report the start of the month, and I'll publish one start of September 13:00
[Coke] a grant report of 'was busy with things including vacation' is an acceptable report. :)
jnthn moritz: I was going to write it at the end of this week and account for what I did this week in it so it'll be mildly more exciting :)
moritz jnthn: works fine for me

Powered by Cro and the Raku® Programming Language.

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

Thank you!