hoodwink.d enhanced
RSS
2.0
XHTML
1.0

RedHanded

YARV is Merging #

by why in cult

Just following up on the Denver Accord: Koichi’s promise was good and he is wrapping up his work on the recent PARALLEL branch of YARV and has updated trunk to sync with Ruby’s 1.9 HEAD. This is an enormous job (as Matz said before) Koichi is trying to keep up with a moving target and YARV significantly diverges at eval.c.

In his letter to ruby-core and yarv-dev-en, Koichi sounds very elated that finally YARV is being merged into Ruby. This means Ruby development will be moving to a new svn repository this month and that YARV will no longer be called YARV because it’ll be called by a new and wonderful name: Ruby. And then we spend the next year working on Ruby 1.9.1.

Yes, so, Koichi is in real need of testers on all platforms.

 mkdir yarv_TRUNK && cd yarv_TRUNK
 svn co http://www.atdot.net/svn/yarv/trunk yarv
 cd yarv
 autoreconf -ivf
 cd ..

 mkdir build && cd build
 ../yarv/configure --prefix=/opt/yarv
 make
 make test
 make install

Post bugs to yarv-dev-en with the word [BUG] in the subject.

said on 06 Nov 2006 at 01:21

Watch out for the case change between yarv_TRUNK and YARV _TRUNK up there in the first line. Macs won’t notice.

said on 06 Nov 2006 at 02:08

This is great news, even though it’s completely expected. Our Christmas present will be in our hands in no time.

said on 06 Nov 2006 at 05:22

nertzy: Should have skipped the && and used cd $! on the next line instead?

said on 06 Nov 2006 at 09:15

I’ve been running some informal tests with ruby -rbenchmark -e "puts Benchmark.measure {}" and YARV ’ed Ruby is a bit… slower. What gives? I thought YARV -infused Ruby was going to be my Red Ryder BB gun this Christmas?

said on 06 Nov 2006 at 09:16

I should note that I actually am putting code in that block. Like 'a'*1_000_000_000

(Interestingly enough, passing it empty like that is faster in YARV .)

said on 06 Nov 2006 at 10:01

why’s build script in a single file:

http://topfunky.net/svn/shovel/yarv/build_yarv.sh

said on 06 Nov 2006 at 10:48

Apparently YARV ’s String#* isn’t very fast yet? I dunno…

said on 06 Nov 2006 at 11:05

Excellente.

I constantly scour the internets for new YARV information yet _why always gets the scoops before anyone else.

said on 06 Nov 2006 at 22:35

Now the big question does the (FF) Sandbox work with YARV ?

said on 07 Nov 2006 at 02:54

I’m going through the course running with my fly Dual 1Ghz G4 and fairly standard setup (I did install Ruby 1.8.4, but that doesn’t count).

Configuring… no problem there.

Compiling right now… all going well… still going… dl.c taking its time… very good… a few trivial warnings (pointer targets in assignment differ in signedness)... done.

Uh oh. Make test basically said “arghhhhhhhh”: not ok assignment 47 -- ../yarv/sample/test.rb:91:in `<main>' and it says it about 100 times.

I just copy-pasted it into the terminal, so it ran make install as well, which responded “nothing to be done” for a number of the entries.

Trying to run Yarv, though, through IRB , seems to work fine. I only tested with little functions and objects, but will give it a good workover tomorrow.

If what I exprienced was truly a bug, and not just dimwittedness on my part, I will submit a bug promptly.

Thanks for the good news, friend!

M.T.

said on 07 Nov 2006 at 11:00

JFMiller: no.

said on DD Mon YYYY at HH:MM

* do fancy stuff in your comment.

PREVIEW PANE