Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2016-08-21

| Channels | #perl6-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:01 sivoais joined #perl6-dev
00:15 Zoffix__ /nick Zoffix\
00:15 Zoffix__ /nick Zoffix
01:13 perlfriend1 joined #perl6-dev
01:48 ilbot3 joined #perl6-dev
01:48 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today
02:28 lizmat joined #perl6-dev
02:48 Gh0st joined #perl6-dev
02:49 Guest47408 help
02:49 Guest47408 !help
02:49 Guest47408 :|
02:50 Guest47408 !test test
02:50 Guest47408 !help
02:50 ShimmerFairy Guest47408: what do you need help for?
02:50 Guest47408 i don't really need help i just wanna view commands
02:50 Guest47408 in other IRC rooms !help get's you a list of commands
02:50 Guest47408 *sometimes
02:50 Guest47408 *gets
02:51 ShimmerFairy We don't have a bot that offers services like that. We just have the Perl 6 evaluation bot, and then a couple others for reporting on various dev-related things.
02:51 ShimmerFairy m: say "Hello, Guest47408, this Rakudo on MoarVM!"
02:51 camelia rakudo-moar 668dc5: OUTPUT«Hello, Guest47408, this Rakudo on MoarVM!␤»
02:51 Guest47408 ah i see
02:51 ShimmerFairy *this is    (oops :P)
02:52 Guest47408 so how do i go about registering my username?
02:52 Guest47408 !nick Gh0st
02:52 Guest47408 :\
02:52 mainfram joined #perl6-dev
02:53 skids joined #perl6-dev
02:53 geekosaur /msg nickserv help
02:53 ShimmerFairy Guest47408: that's something you do via the 'nickserv' service, which is a freenode-wide thing. (try to message nickserv with something like  /msg nickserv help  )
02:53 Guest47408 sweet thanks shimmerfairy
02:54 Guest47408 REGISTER
02:54 mainfram left #perl6-dev
02:55 ShimmerFairy Guest47408: if you can, I'd recommend opening a separate window with nickserv (e.g. /query nickserv I believe). That way you don't risk accidentally typing passwords and such into a chat window :)
02:55 Guest47408 :)
03:01 neato joined #perl6-dev
03:02 neato left #perl6-dev
04:01 lizmat joined #perl6-dev
04:42 rana_ joined #perl6-dev
04:56 ggoebel joined #perl6-dev
05:33 lizmat_ joined #perl6-dev
07:04 lizmat joined #perl6-dev
07:31 TheLemonMan joined #perl6-dev
07:51 TheLemonMan 'morning
07:51 yoleaux2 20 Aug 2016 16:16Z <pmurias> TheLemonMan: the protected block *is* always generated as if it's result was gonna be used
07:52 TheLemonMan RT#125181 can be closed as a proper error message is now shown
07:53 TheLemonMan RT#127012 can also be closed as the problem has has been fixed in moarvm
08:33 dalek roast/6.c-errata: 8d57b03 | jnthn++ | S17-lowlevel/lock.t:
08:33 dalek roast/6.c-errata: Correct test that did unprotected array writes.
08:33 dalek roast/6.c-errata:
08:33 dalek roast/6.c-errata: It's only safe to grab an *already vivified* element of an array and
08:33 dalek roast/6.c-errata: write to it in a cross-thread way. Since @out's elements were never
08:33 dalek roast/6.c-errata: vivified, this led to concurrent binds - and thus resizes - to the
08:33 dalek roast/6.c-errata: array, and thus incorrect results and crashes.
08:33 dalek roast/6.c-errata:
08:33 dalek roast/6.c-errata: The SEGV that sometimes happened needs addressing too; that already
08:33 dalek roast/6.c-errata: has an RT issue and will be covered by another test that actually
08:33 dalek roast/6.c-errata: wants to create that condition, rather than this one which did it by
08:33 dalek roast/6.c-errata: accident.
08:35 nine Zoffix: cherry-picked jnthn++'s fix for lock.t in 6.c-errata
09:07 lizmat hmmm.... I just had t/spec/S05-substitution/subst.t test 180 ("code in replacement part of s/// has correct scoping") flap on me as part of spectest
09:08 lizmat actually, it seems to flap for me about 20% of the time when run by itself  :-(
09:08 lizmat commute&
09:25 pmurias joined #perl6-dev
09:26 pmurias jnthn: it was needed when a lexical was compiled down to a register
09:27 pmurias jnthn: so if the last thing in the protected part of a try was a lexical stored in a register, it was getting overwriten
09:37 RabidGravy joined #perl6-dev
10:00 AlexDaniel joined #perl6-dev
10:01 El_Che left #perl6-dev
10:17 sherwool joined #perl6-dev
10:30 pmurias jnthn: I fixed the leak
10:30 dalek nqp: 4dcc311 | (Pawel Murias)++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
10:30 dalek nqp: Avoid register leak.
10:30 dalek nqp: review: https://github.com/perl6/nqp/commit/4dcc311077
11:00 synopsebot6 joined #perl6-dev
11:28 tadzik joined #perl6-dev
12:41 Zoffix Nice. 40 comments on HackerNews https://news.ycombinator.com/item?id=12328885  "Always wondered how coding in Perl is, guess I will have some fun with it tomorrow :)"
12:47 timotimo Zoffix: my perl6-and-stuff search lane in tweetdeck is full of tweets about your post. that's pretty good :)
12:47 Zoffix :)
12:48 masak Zoffix++ # blog post
12:48 timotimo multiple screen-pages full
12:48 Zoffix I marked as testsneeded the two tickets TheLemonMan mentioned as closable
13:01 nwc10 joined #perl6-dev
13:02 nwc10 Zoffix++ # working to make the flesh creatures obsolete :-)
13:02 nwc10 left #perl6-dev
13:02 Zoffix :)
13:02 Zoffix nwc10++ spotting what went wrong right away
13:09 mst Zoffix: ... so you ended up *not* replacing the tag like I suggested :(
13:10 mst the whole point of 'outweigh the risks' was because inevitably some crusty greybeard was going to yell about published history
13:10 mst but then you went to #git which is all greybeard all the time
13:10 mst and listened to the wrong answer
13:10 mst oh well
13:13 masak fow what it's worth, I don't think I'd have replaced a published tag, either
13:13 Zoffix I didn't know he was a greybeard... I should've asked him to teach me some Thu'um :P
13:15 Zoffix mst, well, the manpage for the tool called that 'the insane approach' and this was right after I sent a release announcement to—hopefully—an umptillion people. I didn't want to make matters worse
13:16 mst I would've replaced it as shortly after a release as possible
13:16 mst since annoying a few people -today-
13:16 mst beats the fact that there's now a steaming turd in published history which is going to emit maggots over people for years
13:17 mst then again, my release tooling also leaves pushing the tag to origin to the *last* step - it uploads to cpan first even
13:17 mst specifically to avoid ever needing to make this decision
13:17 gfldex joined #perl6-dev
13:18 mst masak: hard rules are for deontologists and virtue ethicists; I prefer to use consequentialist reasoning for such choices :D
13:24 Zoffix After a 20-minute run in a loop, t/spec/S05-substitution/subst.t finally failed for me. So I don't get the 20% of the time like lizmat, but there's something wrong: https://gist.github.com/zoffixznet/09d3d3d58fe2d976ee4d172dd18df3af
13:25 Zoffix jnthn, FWIW ^ that's the s/// block scoping that you fixed a couple of days ago. RT#128809
13:25 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128809
13:33 masak mst: I think the chances of anyone going for the bad tag are low in the first place. since I don't see the old tag mattering much, I prefer not to overwrite published history. does that qualify as consequentialist reasoning?
13:34 mst masak: nope, you've hidden the non-consequentialst part behind "I prefer not"
13:34 mst unless you're attempting to make a case via preference utilitarianism
13:35 masak I have no idea if that's what I'm attempting
13:35 mst but then I suspect your preferences are still outweight by the people who prefer normal looking bisect results etc.
13:36 masak that... may be a good point, yes
13:36 mst in general, I find tweaking published history now to be somewhat analgoisable to a sharp pain in your foot
13:36 mst whereas a botched tag becomes a dust speck that homes in on would be contributors eyes forever
13:37 mst and also requires developer time every so often to explain -why- there's an odd tag there or whatever
13:37 mst since you can't attach an explanation without also altering published history
13:37 mst (or at least, I have not found such a way anybody ever notices)
13:38 masak *nod*
13:39 masak I've certainly gone from being very wary about using `push --force` to using it several times per week
13:39 [TuxCM] his is Rakudo version 2016.08.1-3-g668dc5c built on MoarVM version 2016.08
13:39 [TuxCM] csv-ip5xs       10.056
13:39 [TuxCM] test            15.436
13:39 [TuxCM] test-t           7.311
13:39 [TuxCM] csv-parser      16.169
13:40 mst really though, the key take away here is "tag and push the tags last, after you've done everything else"
13:40 nine I use it several times per hour when developing ;) But that's because I develop in branches that I'm quite sure noone else is working on at the same time.
13:40 mst Zoffix: ^^ and that take away is actionable for your code :)
14:01 MasterDuke has anyone had a problem with temp files created by File::Temp disappearing earlier than they should?
14:01 yoleaux2 19 Aug 2016 12:49Z <unmatched}> MasterDuke: would you add yourself to https://github.com/rakudo/rakudo/blob/nom/CREDITS so the entiry shows up as you like in release announcement, please?
14:13 nine jnthn: moarvm.org lists 2016.07 as the lastest
14:40 Zoffix Now after an hour run, t/spec/S05-substitution/subst.t's 's/// has correct scoping' test failed again, this time with a segfault: https://gist.github.com/zoffixznet/8b9f2ea3b1a03c5559aa1de6a1b1a939
15:59 timotimo Zoffix: seeing about a thousand people proclaim "i botched a perl 6 release" on twitter ... seems really weird
16:00 timotimo it's like they all stand in support of zoffix
16:00 timotimo "i, too, botched a perl 6 release"
16:16 b2gills I agree with mst, it would have been better to fix the tag as soon as humanly possible. I think it might still be better to fix it than leave it
16:17 b2gills I would have also created a new tag like what was done
16:19 Zoffix thousand people... :S
16:24 b2gills There are two groups of people, those who already pulled that tag, and those who didn't. Those who pulled already have the wrong one. fixing it wouldn't matter much to them anyway. Fixing it would matter to those who don't have it yet
17:34 zengargoyle joined #perl6-dev
18:50 * unmatched} dissapears for a couple of weeks
18:51 unmatched} Can be reached via twitter if needed
19:00 vendethiel joined #perl6-dev
19:17 go|dfish joined #perl6-dev
19:49 brrt joined #perl6-dev
20:19 mst WTF
20:19 mst nqp --show-config gives me the moar prefix
20:19 mst but not the nqp one
20:28 tadzik joined #perl6-dev
20:30 ggoebel joined #perl6-dev
20:31 lizmat joined #perl6-dev
20:41 lizmat joined #perl6-dev
20:44 mst MUHAHAHAHAHAHAHAHA I think I might have rakudo building
20:50 pmurias mst++ # working on the build system which is something I don't want to touch with a ten foot pole
20:51 mst my ten foot pole is specially sharpened
20:53 mst note this is a hack that will then get backported in parts to master later
20:53 geekosaur hopefully also integrated, and with catch-guards integrated (consider what you need to kill a wild boar without it being reciprocal)
20:56 mst the -I stuff was just Makefile.in hackery
20:56 mst moar --libpath options are more complicated
20:57 mst but I can hack that easily enough
20:58 mst BUT IT FUCKING BUILT
21:02 arnsholt All hail mst! Tamer of build systems and wrangler of code =)
21:03 mst assuming a single --libpath passed to moar was the nasty bit
21:03 mst BUT create-moar-runner.pl actually knows how to do it right
21:05 mst I'll just have to feed it better options
21:06 mst however, that's enough for one day, will fix the remainder tomorrow before packing up to fly :)
21:13 lizmat joined #perl6-dev
21:28 synopsebot6 joined #perl6-dev
22:20 skids joined #perl6-dev
22:34 MasterDuke is there a way to trigger a GC run?
22:41 jdv79 i think there might be an nqp op
22:42 MasterDuke hmm, nothing in the ops doc about it
22:43 jdv79 idk.  maybe im wrong or its not doc'd?
22:44 MasterDuke but that does sound familiar, maybe it exists, but (deliberately?) undocced?
22:44 * jdv79 is on a "phone" so cant be of any more help:(
22:50 MasterDuke m: use nqp; nqp::force_gc()
22:50 camelia rakudo-moar 668dc5: ( no output )
23:24 cognominal joined #perl6-dev

| Channels | #perl6-dev index | Today | | Search | Google Search | Plain-Text | summary