github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
pamplemousse I've posted a progress update on the linker: yakshavingcream.blogspot.com/2019/...oh-my.html 06:45
Now, to generate and run the executable, all you have to do is perl6 --compile=foo foo.pl6; ./foo
moritz that's pretty cool! 07:02
Geth MoarVM: ZhongnianTao++ created pull request #1147:
Add gb2312 encode, decode, and decodestream support		 09:10
patrickb Whoot! 09:17
jnthn Wow, lots of cool GSoC things :) 09:45
timotimo o/ 10:52
i kind of wish we could get away with pointing the perl6 -b at the bytecode in memory without having to write to a file first, but i have *no* clue how to do that right, or at all 10:53
other than that, this is very exciting! 10:54
i wonder if linking a library into the elf file will make "is native(Str)" work immediately? i.e. no need to extract the library to a file, no need to find the right filename? since the symbols are already in the elf that's being loaded? or is that different because we're feeding the bytecode to the perl6 binary with -b and that makes all the difference? 10:59
cygx timotimo: when I did my experiment of having a standalone nqp executable, I did it via a 'virtual file' registry 11:18
cf github.com/cygx/MoarVM/commit/2acf...c87536fadc
never got it to work with the Rakudo setting, though
timotimo mhh, that doesn't look bad actually 11:19
but yeah somehow the core setting is really fiddly and i have no idea why?
brrt \o 11:32
timotimo: that is, in fact, one of the reasons I think an ELF-packed binary is so interesting
brrt cygx: that looks pretty cool, but I (think) you could also use cu_from_bytes.... 11:35
timotimo well, now. maybe not back then ...
cygx brrt: it does - the registry is there to translate from file names requested by the NQP module loader to buffers compiled into the executable 11:37
brrt NQP module loader is different from the CU repo? 11:40
I see
cygx the NQP module loader just looks for .moarvm files in the appropriate location 11:41
I hooked into that lookup at the moarvm level
seemed like the easiest approach at the time
brrt I see 11:43
well, either way might work
I think pamplemousse++ has taken this both further and in a different direction than I had anticipated 11:46
and that's actually awesome...
because there's more than one way to do it :-D
timotimo i wonder if we'll want to have -b as a short option available for the perl6 command 11:56
also, it'd be fantastic if someone™ could put in support for multiple -e blocks. in fact, maybe this is the right spot to put -E
i.e. -E means "this is a code block, but the stuff after it isn't arguments to the user's code, it's more perl6 arguments"
cygx while we're making wishes, this might be something I'd like to have eventually: gist.github.com/cygx/ae6735ae9fb89...064c88b769 12:17
timotimo yeah 12:18
ggoebel timotimo: regarding perl6 -b ... -I wonder if memfd_create could be used to write bytecode to memory and pass a file descriptor? 12:34
brrt pamplemousse++ very nice post 12:37
timotimo possibly; for an early experiment you can try using /proc/self/fd/<num here> as the filename for -b to see if the file descriptor you get from memfd_create works right
oh btw, we also have pl6anet.org, which is a place that pulls RSS feeds from a whole bunch of perl6 blogs together 12:38
another good place to see what's being posted on all the blogs we have
pamplemousse Nifty, good to know :) 12:39
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2019/07/22/...-released/ 15:00
samcv lizmat, my mentee has a blog post and PR now, though the blog post is not super detailed. i think due to less familiarity with english 16:07
so i may give that to you to post once he gets back to me about if he wants it posted (guessing answer will be yes)
lizmat samcv: it will be in next week's P6W then, but yeah, all GSOC reports are very welcome! 16:18
Kaiepi do i need a blog for grant work updates? 16:30
timotimo not necessary, you can just have your grant manager post stuff you write per email to the perlfoundation blog 16:37
pamplemousse lizmat++ #p6weekly 17:07
Geth MoarVM: Altai-man++ created pull request #1148:
Init debug socket for Windows		 17:38
timotimo definitely a candidate for the point release ^ 17:54
samcv hmm why don't i see travis anymore on MVM pull requests 20:02
Geth MoarVM: 7a0b5874e2 | Altai-man++ (committed using GitHub Web editor) | src/debug/debugserver.c
Init debug socket for Windows
MoarVM: 682fbf4ea8 | (Jonathan Worthington)++ (committed using GitHub Web editor) | src/debug/debugserver.c
Merge pull request #1148 from Altai-man/patch-1

Init debug socket for Windows
MoarVM/master: 7 commits pushed by ZhongnianTao++, (Samantha McVey)++ 20:03
MoarVM: 87948fe782 | (Ben Davies)++ | Configure.pl
Only use -fno-ret-protector with Clang on OpenBSD if the JIT is enabled

This flag is only needed to make the legojit work, every other part of MoarVM works fine with RETGUARD enabled.
MoarVM: d605648fe0 | (Jonathan Worthington)++ (committed using GitHub Web editor) | Configure.pl
Merge pull request #1146 from Kaiepi/openbsd

Only use -fno-ret-protector with Clang on OpenBSD if the JIT is enabled
Geth MoarVM/2019.07.1: 8 commits pushed by (Ben Davies)++, (Stefan Seifert)++, (Jonathan Worthington)++, (Timo Paulssen)++, Altai-man++, (Samantha McVey)++ 20:30
AlexDaniel samcv: nice! Leave it like this for a bit, I'll double check some things tomorrow 21:14
samcv AlexDaniel, +1 22:18
samcv before the second evaluation deadline i'm going to get a better blog posting from my mentee. I slacked a bit by not requiring formal blog posts, but he was making progress and i was somewhat busy :). Happy with how things are going overall though 22:29
jnthn samcv: Was glad to see a PR on that today. Somebody filed a Cro issue last week or so mentioning about the missing encoding support. :) 22:30
patrickb Hey! I'm currently struggling to figure out which header files I need to include on windows (using the Windows API). Locally stuff builds, but AppVeyor CI complains about not finding processthreadsapi.h and unkown BOOL. 22:44
Any clues on how to check what is needed without many change-push-CI-check cycles? 22:45
AlexDaniel (everyone who makes rakudo better on Windows)++
patrickb I'm working on the perl6 runner programs. 22:46
jnthn patrickb: I don't know off-hand, but are you locally and appveyor using the same toolchain? 22:47
For example, if it's using the Visual Studio toolchain and you have mingw or something... In which case trying to build with whatever appveyor is would at least shorten the cycle 22:48
patrickb both using the cl compiler
I have 2019 and don't know what AppVeyor uses... 22:49
Not true. 2017 is what I use
jnthn Link to appveyor output? 23:00
patrickb I have proceeded two more iterations by now. 23:05
ci.appveyor.com/project/rakudo/rak...m43d101bb5
That's the current one 23:06
still in progress
Got to testing. I guess I got through. 23:07
jnthn Yeah, was gonna say, don't see the error you mentioned :) 23:08

Powered by Cro and the Raku® Programming Language.

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

Thank you!