Re: NQP JVM prototype faster than Perl 5

2013-02-02 Thread Darren Duncan
Well, my assumption was that I would start off by producing some kind of source code on disk that is the same kind as that which defines Rakudo, which your existing toolchain could then compile and execute. Similar to how my current implementation actually produces equivalent Perl 5 source code

Re: NQP JVM prototype faster than Perl 5

2013-02-02 Thread Matthew Wilson
Did you mean to use $z in the say output of the nqp and perl versions of the microbenchmark, or did you mean to run it twice? On Sat, Feb 2, 2013 at 10:43 AM, Nicholas Clark wrote: > This is surprising, interesting and pleasing... > > There's some example NQP code to time calculating Fibonacci

Re: NQP JVM prototype faster than Perl 5

2013-02-02 Thread Matthew Wilson
Right, but (sorry if this is already clear) there's no textual QAST language; you currently must be running in NQP to create the node objects. If you wanted to use a Perl5-bootstrapped Muldis D to produce QAST, you could write it to some serialization on disk, then you would need to deserialize th

Re: NQP JVM prototype faster than Perl 5

2013-02-02 Thread Darren Duncan
Okay, I understand, I would target QAST. So once I've implemented over Perl 5, and bootstrapped where possible, I'll work on targeting QAST. Thank you. -- Darren Duncan On 2013.02.02 4:29 PM, Matthew Wilson wrote: QAST is the protocol rakudo and NQP use to send abstract syntax trees to the V

Re: NQP JVM prototype faster than Perl 5

2013-02-02 Thread Matthew Wilson
QAST is the protocol rakudo and NQP use to send abstract syntax trees to the VM-specific compiler. PIRT is the target for parrot, and JAST is the target for nqp-jvm. You can view its source in nqp's source under src/QAST, as well as note src/NQP/Actions.pm, which generates the QAST nodes while pa

Re: NQP JVM prototype faster than Perl 5

2013-02-02 Thread Matthew Wilson
I'm sorry; I was just filling out what the last letter in the acronym QAST is. Search for QAST instead. On Sat, Feb 2, 2013 at 4:21 PM, yary wrote: > This doesn't happen very often. Google tells me > > Your search - *QASTree nqp* - did not match any documents. > > where does one read more abou

Re: NQP JVM prototype faster than Perl 5

2013-02-02 Thread yary
This doesn't happen very often. Google tells me Your search - *QASTree nqp* - did not match any documents. where does one read more about this QAStree? -y

Re: NQP JVM prototype faster than Perl 5

2013-02-02 Thread Matthew Wilson
absolutely, though to be precise you'll want to compile to QASTrees in the NQP environment (not to NQP code itself). On Sat, Feb 2, 2013 at 2:32 PM, Darren Duncan wrote: > So here's a question ... > > Is it now, or should it soon be, reasonable to treat NQP as its own entity > that is a build ta

[perl #116613] Parent types augmented at run time no longer make augmented methods available to sub types

2013-02-02 Thread via RT
# New Ticket Created by Steve Schulze # Please include the string: [perl #116613] # in the subject line of all future correspondence about this issue. # https://rt.perl.org:443/rt3/Ticket/Display.html?id=116613 > Some time between Oct 2012 and Feb 2013 Rakudo stopped making methods added at

Re: NQP JVM prototype faster than Perl 5

2013-02-02 Thread Darren Duncan
So here's a question ... Is it now, or should it soon be, reasonable to treat NQP as its own entity that is a build target for a programming language, instead of treating Parrot as a build target? After getting the reference implementation of my language working over Perl 5 (because it is ma

NQP JVM prototype faster than Perl 5

2013-02-02 Thread Nicholas Clark
This is surprising, interesting and pleasing... There's some example NQP code to time calculating Fibonacci sequences. I've tweaked it a tiny bit to take an optional count on the command line. Unfortunately the NQP JVM prototype doesn't yet set @ARGS, so this isn't that useful. Anyway, the code i