Robert Dodier <robert.dodier at gmail.com> writes:
> On 2013-05-27, Raymond Toy <toy.raymond at gmail.com> wrote:
>> I don't have any insight on this issue, but perhaps we should arrange
>> it so that the share testsuite is also run with run_testsuite? Maybe
>> an additional arg to run_testsuite to run the share tests as well? Or
>> maybe just always run these tests with run_testsuite?
>
> I dunno -- some of the code in share is pretty flakey -- I'm not sure if
> we can expect the share test suite to run successfully. I guess I'm
> inclined to leave it as it stands, with share_testsuite being separate
> from run_testsuite.
Hmm. I'd be much happier if we ensured the testsuite at least ran to
completion and then just marked a load of expected errors.
There is a problem with run time though. On my machine, the fastest lisp
at running the testsuite is SBCL, which takes 173 seconds (vaguely
modern laptop; 5.5GiB memory; 2 x 2.3GHz processors). CLISP takes
something like a factor of eight longer, if I remember correctly. I
don't know how long the share testsuite takes.
Maybe one option is to add a "share" keyword argument to run_testsuite,
taking values "no, yes, only" or something. That would make running the
share stuff a bit easier, at least. (Not sure what the default value of
the argument should be...)
Rupert
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 315 bytes
Desc: not available
URL: <http://www.math.utexas.edu/pipermail/maxima/attachments/20130527/ee81079b/attachment.pgp>;