Re^5: MoarVM update

[Jenda]

You promised solid beta of Perl6 in less than a year, yet one of the prerequisities (i.e. NQP) is to take years. So what is that solid beta gonna run on? Seems it's not gonna be MoarVM.

[raiph]

I didn't promise a solid beta.

I didn't say NQP would take years.

[Jenda]

"Based on what I know, I'd guess that MoarVM will run Rakudo Perl 6 enough to compile and run itself before YAPC::NA 2014 in 9 months." - raiph in Re^2: MoarVM update

"I can see the P6 team aiming at having a solid 6.0.0 beta that adds p5interop, compact arrays, concurrency, async IO, unicode, macros, module versioning, better libs, better module installer, much better performance, complete documentation, and user support by YAPC::NA 2014." - raiph in Re^4: A $dayjob Perl 6 program that runs 40x faster on the JVM than on Parrot

"NQP on MoarVM is not a week away from passing the NQP testsuite. At a guess I'd say it'll be a few years before NQP passes 100% of its test suite on MoarVM." - raiph in Re^2: MoarVM update







The celebrity confuse-all of the Perl-6 cult I guess.

[raiph]

if they are one week from passing all tests

"NQP on MoarVM is not a week away from passing the NQP testsuite."

Right. I stand by that. You made a mistake. I felt it needed correction.

"At a guess I'd say it'll be a few years before NQP passes 100% of its test suite on MoarVM."

Right. I stand by my forward looking statements about Perl 6.¹ Given the constant accusations of spin, I decided to start with the absolute worst case I could think of, namely passing 100% of the test suite, and allowing I might be underestimating how hard that would be. (Fwiw, subsequent conversations with a MoarVM dev suggest I was being far too conservative.)

"Based on what I know, I'd guess that MoarVM will run Rakudo Perl 6 enough to compile and run itself before YAPC::NA 2014 in 9 months."

Right.¹

One of the devs I'd call a core MoarVM dev told me (before my comment above) that he thinks it's 50/50 that Rakudo Perl 6 will compile and run itself on MoarVM this year.

Update: Rakudo on MoarVM compiled and ran itself a couple months after my above comment.

P5Interop is set to be developed on MoarVM first (with $10k chasing its completion). This is one of many factors that suggest there will likely be pressure to develop MoarVM rapidly.

Update: MoarVM did indeed develop rapidly, including the pieces necessary to allow nine to rapidly (in about one month!) develop comprehensive P5Interop on MoarVM (for free, using a less sophisticated approach than the one diakopter proposed).

The official Perl 6, Rakudo and NQP lead dev, Patrick Michaud, pretty much stopped working on Perl 6 for the last 2 years. He is set to return this year, perhaps this month. \o/ Patrick will almost certainly be happy to help develop Rakudo on MoarVM.

Update: Patrick has appeared on #perl6 occasionally but that's about it. Otoh he attended APW this weekend and aiui he's going to work on the GLR.

jnthn is currently the de facto lead P6 dev. There are things in his life that suggest he really wants to bring P6 to a satisfying point about 9 months from now.

jnthn has been teaching ever more people around him to work on the NQP toolchain (eg the NQP workshop this weekend in Germany is being attended by 15 devs). Some of these folk are starting to make NQP and MoarVM commits, and I anticipate an uptick after the workshop.

Etc.

Update: I don't know if jnthn thinks P6 is at a satisfying point but he shows no signs of slowing down his P6 development efforts. There are 58 committers to the NQP repo, with 29 of those having landed commits so far in 2014, and 31 MoarVM committers with 19 of those landing commits so far in 2014.

---

¹ My hope is that, if someone reads what I say carefully they will find it to be clear, reliable, and without spin. For forward looking statements about Perl 6:

• I generally try to be specific about what I'm saying might happen (so the outcome will typically be either yes it happened, or no it didn't)

• I am always conservative (in the hope I might build a track record over several years of getting my forward looking statements consistently right)

• If I say "before", I mean any time between now and the date I suggest.

• I frequently qualify what I am saying with things like "aiui", or "imo", or by calling something a "guess" or an "estimate". These qualifications indicate how strongly (or weakly) assertive I intend my statement to be.