stable-sort temporary vector retention

[Christophe Rhodes]

Report from Keith James

 affects sbcl
 status confirmed
 importance medium
 tag sort performance space patch
 done

>>>>> "Christophe" == Christophe Rhodes <email address hidden> writes:

    Christophe> In this day and age, I'm not sure
    Christophe> why stable-sort / vector-merge-sort doesn't simply
    Christophe> create a new vector for itself, which would at a
    Christophe> stroke eliminate both this leaks and worries about
    Christophe> threadsafety

I have modified my sbcl by removing *merge-sort-temp-vector* and
replacing it with a fresh vector each time. The result is that memory
use is much more manageable for stable-sorting vectors of big data.
The test suite runs successfully.

I've attached a patch for your consideration.

[Nikodemus Siivola (nikodemus)]

In SBCL 1.0.36.18.