Hi Volker,

On 10 Jan., 15:46, Volker Braun <[email protected]> wrote:
> Running doctests should approximate a long-running Sage session imho. So if
> you leak memory, say, it should start to hurt after a while.

If you have a memory leak then garbage collection wouldn't change it:
Leaking data can, by definition, not be collected. So, it would hurt
anyway...

And of course, garbage collection will run during the tests (and also
in any long-running Sage session) anyway.

The point is: For some (not all) pairs of tests A and B, it matters
whether the next garbage collection after finishing A happens before
starting B or happens in the middle of B.

So, we are in the same situation as with randomness: For some (not
all) tests, the initial state of the random generator matters.

In other words: If we fix the initial state of the random generator
then why don't we fix the initial state of (say) the coercion caches
as well, in a way that would certainly *not* hide a memory leak?

Cheers,
Simon

-- 
To post to this group, send an email to [email protected]
To unsubscribe from this group, send an email to 
[email protected]
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Reply via email to