[Ecls-list] Release ahead, report problems

[Juan Jose Garcia-Ripoll]

Fixed. I removed the bogus inline forms -- which leaves the moderately
inefficient ones. It would be nice to improve the inline code to handle also
floor with fixnums and multiple return values.

On Mon, Feb 8, 2010 at 2:00 PM, Alexander Gavrilov <angavrilov at gmail.com>wrote:

> Hi, there is a bug in code inlining:
>
> (defun test (a b)
>  (declare (fixnum a b))
>  (multiple-value-bind (q r) (floor a b)
>    (declare (fixnum q r))
>    (print q)
>    (print r)))
>
> > (test 3 2)
>
> 1
> 1
> 1
>
> > (compile 'test)
> > (test 3 2)
>
> NIL cannot be coerced to a C int.
>
>
> This happens because only the calculation for the 0th value is
> inlined, and thus r contains garbage.
>
>
> On Wed, Feb 3, 2010 at 11:43 PM, Juan Jose Garcia-Ripoll
> <juanjose.garciaripoll at googlemail.com> wrote:
> > I want to produce a new release. For that I am officially beginning a
> test
> > phase -- I will have to update the test farm, unable yet to install
> Samium's
> > new system --, and will look into the following problems
>



-- 
Instituto de Física Fundamental, CSIC
c/ Serrano, 113b, Madrid 28006 (Spain)
http://juanjose.garciaripoll.googlepages.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/ecl-devel/attachments/20100208/6a90d736/attachment.html>