[Ecls-list] Re: One more fix

Paul F. Dietz dietz at dls.net
Fri Dec 17 06:43:01 UTC 2004

Juan Jose Garcia Ripoll wrote:

> This solves a number of errors in Paul's test suite, but ECL has not yet 
> reached the desired stability under the random tester.

All the tests in gcl/ansi-tests/misc.lsp pass now, but the random tester
does show strange behavior if you run it long enough.  The failure rate
starts to go up, and eventually things become completely corrupted.
I suspect there's a gc safety bug somewhere.


More information about the ecl-devel mailing list