japhb Sadly, it appears the MoarVM fixes this week did *not* coincidentally fix my input lockup woes: This is Rakudo version 2017.02-224-g68a40f7d9 built on MoarVM version 2017.02-31-g91aab714 00:06
I'd hoped that my weird stuck input thread problem would be fixed by one of the stability fixes, but alas no. 00:07
Ah well, no avoiding work by playing Perl 6 games today. ;-) 00:08
TimToady when you say "input", do you mean a pipe that could wedge because the other end filled its stderr, like precompilation does when you put too much debugging out on stderr? 00:23
00:25 geekosaur joined 01:06 zakharyas joined 01:11 TimToady joined 01:35 agentzh joined 02:48 ilbot3 joined 03:42 agentzh joined 04:40 agentzh joined
samcv ok got full spectest for my case insensitive matching foldcase patch 06:12
\o/
06:36 brrt joined 07:03 domidumont joined 07:45 brrt joined 09:40 Ven joined
jnthn samcv: \o/ 10:15
Will look very soon :)
10:22 Ven joined
samcv thx :) 10:48
10:51 Ven joined 12:06 Ven joined 13:22 Ven joined 13:48 yoleaux2 joined
dogbert17_ shocking silence ... 14:25
what's our take on making patches to 3rd party libs, do we have any? Is referring to MoarVM 549 aka RT #130713 14:32
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130713
14:37 Ven joined
timotimo it'd be nice if we could sanely handle patchsets 15:05
brrt good *, timotimo, dogbert17_ 15:08
dogbert17_ o/ brrt 15:12
timotimo, you found a patch for that expmod problem 15:13
problem seems to be that it has not been merged into libtommath :( 15:14
brrt it's always possible to maintain your own fork, and, it's even pretty easy 15:16
and in general i prefer that above having patchsets
timotimo that's the problem, yes 15:18
ISTR this isn't the first time we had problems where we had to patch a 3rdparty thing
also, there's some distros that are super vigilant about putting 3rdparty stuff or anything sharable really into distro-wide packages 15:19
[Coke] timotimo: we kind of support letting people use their own copies of the 3rd party libs, so that could mean we have a rakudo out there with the same version # that is behaving differently. Not sure how much we want to worry about that. 15:20
(Maybe we just make a super verbose --version)
brrt that's true;, but in fact, we often know better
dogbert17_ I see that some patches have been made earlier to the MoarVM copy of the lib in question 15:21
nine It probably also depends on how obscure the bug is
And how likely that upstream will include the patch soonish.
dogbert17_ github.com/libtom/libtommath/issues/67
timotimo i say we merge that patch in our own fork. 15:25
Latest commit b8527e9 on 10 Oct 2016 15:26
nine Maybe we could get co-maintainership on that repo?
timotimo libtommath is sufficiently slow moving
at least they got the rand fix in
15:35 brrt joined 15:51 brrt joined
timotimo You don’t have permission to fork to these organizations: 16:05
MoarVM
github.com/libtom/libtommath - jnthn wanna push the "fork" button on this page and fork it into MoarVM/ ? 16:06
jnthn Didn't we already have a fork of that? :)
ah, no 16:07
timotimo i looked, we didn't
jnthn hah, you're right
Forkee 16:08
d
timotimo can you give me lots of permissions for this?
for example to edit the description
may not even be necessary, actually *shrug* 16:09
jnthn See if you ahve it now
timotimo no "Edit" button for the description 16:10
no "Settings" tab at the top either
jnthn Now? 16:11
timotimo that's it! :)
jnthn yay 16:12
timotimo um, wait ... our libtommath isn't actually a submodule?!
indeed, we just pushed stuff in there ... 2 years ago
"oops" :) 16:13
timotimo rebuilds and spectests and all that 16:16
japhb TimToady: github.com/ab5tract/Terminal-Print...wInput.pm6 is the input code, github.com/ab5tract/Terminal-Print...es/tris.p6 is the example code that uses it, and github.com/ab5tract/Terminal-Print...-282329480 is the behavior. 16:19
timotimo teh tris! 16:20
japhb [Coke]: Perhaps `make test` for MoarVM should include tests for the things we've patched in our dependency libs, and if someone compiles with a non-blessed version that still has the bugs, we can complain loudly. I'm guessing this would be visible to packagers, at least. And with a good enough error message, we can tell them what patch(es) to cherry-pick. 16:21
timotimo: Yup! :-)
timotimo are you adhering to the official rules on how tetris must behave?
japhb It's closing in on original tetris. It doesn't have the special behaviors for e.g. flipping a T into place and getting extra points for a clear -- but then again, the input problem happens quickly enough that I haven't coded line clears yet, so there you go. Right now it's literally the *original* tetris, that was just a test of fitting blocks together tightly, before he invented the line-clearing to make the game last longer. :-) 16:23
timotimo fair enough 16:25
japhb It does have the feature that was most important to me when I was learning -- wall kick when rotating. You don't get a piece stuck against the side (or a block tower) because you're rotating "the wrong way". 16:28
timotimo right 16:29
i see you know enough about tetris to do this stuff right
japhb Heh
Geth MoarVM: fbb4c27245 | (Timo Paulssen)++ | 128 files
turn libtommath into a submodule, use MoarVM's fork
 16:32
timotimo people will probably have to perl Configure.pl to make all this work again?
and now i'll put in that patch to make the expmod thing work 16:34
Geth MoarVM: 1288e7edde | (Timo Paulssen)++ | 3rdparty/libtommath
update libtommath fork to have expmod hang fix
 16:37
17:23 Ven joined 17:26 zakharyas joined
nwc10 I'm not sure about the reconfigure. "take off and nuke it from orbit" possible 17:35
possibly
Configure.pl gave me this:
Updating submodules .................................... FAIL git error: fatal: destination path '/home/nicholas/Perl/MoarVM/3rdparty/libtommath' already exists and is not an empty directory.
fatal: clone of 'git://github.com/MoarVM/libtommath' into submodule path '/home/nicholas/Perl/MoarVM/3rdparty/libtommath' failed
Failed to clone '3rdparty/libtommath'. Retry scheduled
fatal: destination path '/home/nicholas/Perl/MoarVM/3rdparty/libtommath' already exists and is not an empty directory.
fatal: clone of 'git://github.com/MoarVM/libtommath' into submodule path '/home/nicholas/Perl/MoarVM/3rdparty/libtommath' failed
Failed to clone '3rdparty/libtommath' a second time, aborting
*deleteing* it did
git clean -dxf on the directory didn't help
timotimo :o 17:37
but i already git rm'd that folder
how come it doesn't do this properly?
nwc10 I didn't look too hard, but I guess either
timotimo maybe because there's some .o files?
i wonder what happens if you "make realclean" first
nwc10 1) I misremember just what I tried before deleting it
or
2) yes, files that git ignores
17:38 Ven joined 18:01 domidumont joined 18:09 Ven joined 18:23 Ven joined 18:38 FROGGS joined 18:53 Ven joined 19:34 Geth joined 19:39 Ven joined 19:54 Ven joined 20:08 Ven joined
samcv jnthn, you look at my PR? 20:11
jnthn D'oh, sorry, got zoned in to $other-job task earlier 20:13
It was the first time this week that $other-job was being fun rather than a pain :P
Hm, wonder why Travis is sad... 20:14
samcv: travis-ci.org/MoarVM/MoarVM/jobs/209636303#L471 20:15
Looks like that's a legit compile error
samcv will check that out. any other comments on it though? 20:16
jnthn I did the bit that didn't need me to think first :-)
Looking at the code now :)
h_offset = 0; /* XXX I think this is the right behavior here */ 20:17
I thought we would throw on those
samcv well that's what the other function does 20:18
jnthn Oh, though maybe not
Yeah, throw is regex-engine-nasty I guess
samcv the case sensitive one
jnthn *nod*
Oh, though don't know how this'd come up in the regex engine
Anyway, if it matches what the sensitive one does, thenfine
samcv: Left comments, other than that looks good 20:24
(And my comment is only pointing out why the Travis was upset)
samcv it shows a weird commit number on travis 20:26
that's not the actual commit hash 20:27
odd
thx for showing me. the line numbers were off on the error message
and not sure if it was the correct commit either 20:28
jnthn I'm used to spotting that kinda issue 'cus it upsets MSVC. :) 20:29
samcv high quality enterprise™ product 20:30
jnthn "Nobody uses C any more, we'll focus on C++ only!"
samcv it's amazing how well C has fared as languages go 20:31
jnthn oops, that commit won't work?
Because it uses h_g before assigning it?
samcv sorry :P I just woke up like 20 mins ago 20:32
jnthn :-)
jnthn felt like that all day
samcv maybe should just allocate h_fc_cps once at the top and then not have to allocate it ever again 20:33
jnthn, i always thought it was weird you could initialize variables WITH code but couldn't have any other kind of code 20:38
or i guess it's only functions or something not 100% sure. 20:39
haven't had to use this C standard except on MVM
argh the build is still failing at the same spot 20:52
well. it's still giving the wrong line number.
oh nvm i think it's still showing the old status. will wait for travis to wake up 20:53
jnthn samcv: Yeah, I find it annoying, but "don't build on MSVC" isn't really an option for us either 21:08
samcv yep
21:11 ggoebel joined
mst jnthn: can't see from backlog - what's MSVC filling its diaper over? 21:16
21:38 Ven joined
samcv the usual. code mixed with declarations 21:38
Geth MoarVM: 58457845f5 | (Samantha McVey)++ | src/strings/ops.c
Fix string_equal_at_ignore_case when string `a` changes length (#548)

Fix string_equal_at_ignore_case when strings len changes on fc
When strings foldcase, they can become longer, example 'st' becomes 'st'. There are other examples from non-Latin scripts as well. Previously we foldcased both the needle and haystack and then checked them ... (8 more lines)		 21:46
jnthn samcv++ 21:48
22:23 Ven joined 22:39 Ven joined 22:49 dogbert17 joined 22:58 zakharyas joined 23:09 Ven joined 23:30 geekosaur joined 23:56 Ven joined
Geth MoarVM: d7caeba3c2 | (Timo Paulssen)++ | 3 files
mention debugname, reprname for "requires obj with repr X"
 23:58

Powered by Cro and the Raku® Programming Language.

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

Thank you!