> Pretty much, 64 bit emits a lot of labels that are not used for jumps
> or anything, so during peephole optimization
> this new version ignores the extra labels for the purpose of pattern
> matching, and leaves them in place.
Just checking: it ignores only labels without any jumps to them, right?
Cheers,
-- Nikodemus
------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_______________________________________________
Sbcl-devel mailing list
Sbcl-...@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sbcl-devel
On 17 September 2010 07:48, Jonathan Smith <jonathan...@gmail.com> wrote:Just checking: it ignores only labels without any jumps to them, right?
> Pretty much, 64 bit emits a lot of labels that are not used for jumps
> or anything, so during peephole optimization
> this new version ignores the extra labels for the purpose of pattern
> matching, and leaves them in place.
Cheers,
-- Nikodemus
> https://sourceforge.net/projects/sbcl-peep-opt/files/
> (file is sbcl-optimizer.diff)
Hi, I wanted to put this on my merge queue for 1.0.50/51, and noticed
that the diff appears broken -- patch says that only garbage was found
at input.
Not sure what's up with that, but if you still have the tree around,
if would be great if you could regenerate the diff with
diff -ruN sbcl-old sbcl-new
It doesn't look _too_ hard to patch in manually if you've lost the
tree... but I'd rather not. :)
Cheers,
-- nikodemus
------------------------------------------------------------------------------
vRanger cuts backup time in half-while increasing security.
With the market-leading solution for virtual backup and recovery,
you get blazing-fast, flexible, and affordable data protection.
Download your free trial now.
http://p.sf.net/sfu/quest-d2dcopy1
Are you looking for the output of this command:
diff -ruN ~/Dropbox/sbcl-1.0.42-peep/ ~/Dropbox/sbcl-1.0.48/ > peephole-diff.txt
or this command:
diff -ruN ~/Dropbox/sbcl-1.0.42 ~/Dropbox/sbcl-1.0.42-peep// >
peephole-diff.txt
I've attached the results of the first command In compressed form. If
you need the second I believe I can do some digging on my home
computer and find a version of sbcl-1.0.42 there.
Attached is a diff of 1.0.42 against the peephole optimizer version
It seems to have applied cleanly against 1.0.48 on my machine.
('patch -p5 -i patchfilename' in the sbcl 1.0.48 directory)
I compiled and ran tests and it was an 'apparent success'.
> I thought about it and decided that the second version would likely be
> easier to apply as a patch.
That was what I meant, yes. Sorry for being unclear. I'll see if I can
review and forward port it over the next week or two.