Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2017-12-18

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

All times shown according to UTC.

Time Nick Message
01:30 NEDERLA606 joined #moarvm
01:37 zostay left #moarvm
02:16 tbrowder joined #moarvm
02:21 icecre112 joined #moarvm
02:22 tbrowder left #moarvm
02:58 ilbot3 joined #moarvm
02:58 Topic for #moarvm is now https://github.com/moarvm/moarvm | IRC logs at  http://irclog.perlgeek.de/moarvm/today
06:53 Morfent joined #moarvm
06:55 Morfent joined #moarvm
07:22 brrt joined #moarvm
08:02 domidumont joined #moarvm
08:03 brrt joined #moarvm
08:07 domidumont joined #moarvm
09:31 brrt joined #moarvm
09:40 brrt good * #moarvm
09:40 brrt so, i'm annoyed by the fact that ASAN on macOS is not showing me the same failures that ASAN on fedora *is*
09:40 brrt i know, different compilers, platforms etc
09:40 brrt so now i'm going to create a VM so I can actually get some useful work done :-P
10:08 brrt joined #moarvm
10:36 domidumont1 joined #moarvm
10:50 brrt joined #moarvm
11:02 Voldenet joined #moarvm
11:02 Voldenet joined #moarvm
11:06 dogbert2 joined #moarvm
11:26 samcv does someone want to work on the changelog for me this month? i'm sick and don't feel up to doing the whole thing. though i can do all of my own commits at least
11:27 lizmat samcv: will see what I can do
11:29 lizmat samcv: so is it the plan to change the wiki and then transplant that into the ChangeLog file ?
11:29 samcv yeah
11:30 samcv ping me when you or whoever is close to done and i'll make sure my own commits are properly represented :)
11:30 AlexDaniel samcv++ lizmat++
11:37 lizmat AlexDaniel: don't ++ me just yet  :-)
11:47 lizmat samcv did like 80+% of MoarVM commits this month
11:48 lizmat with most of them related to ucd2c.pl
11:48 lizmat samcv: I've gone through the commits of last month and documented what I could
11:48 lizmat HTH  :-)
11:49 samcv thanks! will look after i have some tea
11:50 lizmat good plan: I will also get some  :-)
12:36 brrt joined #moarvm
12:44 squashable6 joined #moarvm
12:44 reportable6 joined #moarvm
12:44 unicodable6 joined #moarvm
12:44 greppable6 joined #moarvm
12:44 benchable6 joined #moarvm
13:14 samcv ok so i've at least narrowed down part of the issue i've been working on fixing with ucd2c.pl. it seems that it may be the binary search algorithm or something. it's getting the wrong bitfield row and at least that is happening and causing it to start failing massively when i add in codepoints not in UnicodeData.txt
16:05 brrt joined #moarvm
16:39 zakharyas joined #moarvm
16:41 * samcv is happy she's finally making some inroads on fixing the source of the issue she's been trying to hunt down for this month
16:43 brrt \o/
16:43 brrt tell us more
16:44 samcv well for the last few weeks i've been improving ucd2c.pl to ultimately try and track down this error. most of the changes i've made haven't resulted in code changes. one significant one. is so all points are stored in $POINTS_BY_CODE we used to have $points_by_code and $points_by_hex. so i made all use one refhash, $POINTS_BY_CODE (since i renamed everything to be uppercase for globals)
16:45 brrt uhuh
16:45 samcv and so then next, all points have a $POINTS_BY_CODE->{20}->{next_point} and this used to be set only for UnicodeData.txt, and so we had effectively one sub setting that and another one also setting it
16:46 samcv and i wanted everything to use the same method to set it. and it's very problematic. because that assumes input is one ordered, and two it's just error prone
16:46 samcv if i wanted to extend functionality so we could add properties for codepoints not in UnicodeData.txt
16:47 samcv also another thing that was error prone was adding codepoints to @$PLANES. so what it does now it iterate through $POINTS_BY_CODE refhash once we've put all the data into it, setting {next_point} and adding it to @$PLANES
16:48 brrt uhuh
16:48 samcv so then we make sure anything in $POINTS_BY_CODE has those values set properly. since though it worked fine normally, when i started to try and make more changes it started breaking things. i think because there must have been some minor "mistake" with next_point or the planes and it missed some or something
16:49 samcv anyway that's been fixed, and now i've been trying to track down the same problem really, but finally have figured out where it is located. and it's MVM_codepoint_to_bitfield_row
16:52 samcv but i believe the binary search if else's are missing certain codepoints
16:53 samcv possibly because it has separate binary search sections for codepoints < 0x10000
16:53 samcv and i think the @$EXTENTS may cause it then to miss those overlapped codepoints. and maybe some other issue
17:14 AlexDaniel samcv: any ETA for the moarvm release? I want to adjust myself accordingly
17:17 timotimo adjust yourself before you bust yourself
17:21 samcv AlexDaniel: a day or so possibly. since i've made progress with this i may want to release after i release a fix. though it's technically not a regression
17:22 AlexDaniel that's OK for me, so I'll check back tomorrow
17:22 AlexDaniel in my UTC+02:00 tomorrow :)
17:22 samcv ok cool :)
17:44 domidumont joined #moarvm
18:01 geospeck joined #moarvm
18:05 zakharyas joined #moarvm
18:09 coverable6 joined #moarvm
18:09 nativecallable6 joined #moarvm
20:57 committable6 joined #moarvm
20:58 bisectable6 joined #moarvm
20:58 bloatable6 joined #moarvm
20:58 quotable6 joined #moarvm
20:58 releasable6 joined #moarvm
20:58 evalable6 joined #moarvm
20:58 statisfiable6 joined #moarvm
20:59 lizmat and another Perl 6 Weekly hits the Net: https://p6weekly.wordpress.com/2017/12/18/2017-51-thirty-years/
21:01 ChanServ joined #moarvm
21:04 moritz lizmat++
21:26 lizmat perhaps interesting for MoarVM devs: http://mattwarren.org/2017/12/15/How-does-.NET-JIT-a-method-and-Tiered-Compilation/ ?

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