github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
lizmat github.com/scandum/quadsort # interesting ? 11:39
comments on HN news.ycombinator.com/item?id=22322967
jnthn Ok, let's try and sort out this guard tree refactor... 13:24
sena_kun jnthn: good luck. 13:27
Geth MoarVM/derived-specializations: 9b0910d1db | (Jonathan Worthington)++ | 3 files
Revert "More information in dupe arg guard add error"

This reverts commit 72c1d55f488876f3680358edf2a33cebee2748f4. We're going to elimiante this code soon anyway, and the change makes the upcoming refactor harder.
13:41
MoarVM/derived-specializations: e76d700f67 | (Jonathan Worthington)++ | 4 files
Stash callsite and type tuple in spesh candidate

So we have them available for re-generating the arg guard whenever we wish to do that.
Guest4508
.oO(The great guard tree refactor)
13:44
Geth MoarVM/derived-specializations: 49e2ae6f88 | (Jonathan Worthington)++ | 3 files
Update API for updating the arg guards

To suggest we'll now be regenerating them, and to provide the full candidate list. This doesn't actually put in place the new algorithm yet, just makes sure we have the data being passed down to do it.
13:56
Guest4508 jnthn: don't forget to grab a coffee break 14:28
jnthn Well, tea in the afternoon :) 14:58
nwc10 and the beer fridge is primed? 15:03
oh, I missed that. What is this "afternoon" concept? :-)
japhb nwc10: It's in between "lunch food torpor" and "dinner food torpor" 15:24
Geth MoarVM/derived-specializations: cc927e6bcb | (Jonathan Worthington)++ | src/spesh/arg_guard.c
WIP on arg guard tree building refactor
17:04
MasterDuke two things i saw in HN today: github.com/carolemieux/perffuzz (a modified afl that finds input to detect performance problems) and arxiv.org/abs/2002.05599 (experimenting with sorting networks for the base case of sorting algorithms) 18:21
MasterDuke nine: i was doing a little googling about that setjmp/longjmp clobbering warning. it looks like it's triggered by github.com/MoarVM/MoarVM/commit/d7...314374ba2c 18:51
s/it looks/i think without having actually tested that/ 18:52
the warning seems to be about values being set that could be overwritten when returning to the site of a setjmp 18:53
nine MasterDuke: oh it's definitely that commit
I just don't know what to do about is
MasterDuke ah
nine Well I'm pretty sure a refactor thaf moves around a lot of the code would fix it... 18:57