stmx test suite problem on ECL

stmx testsuite sometimes hangs on ECL, sometimes crashes with "out of memory" error.

http://common-lisp.net/project/cl-test-grid/library/stmx.html

Sep 05 '13 22:09 avodonosov

Commit f1b9fa233a18f551d60f7ce87e8f3c9cfc2508ef should fix this bug. I guess it will be picked up by cl-test-grid only after the next Quicklisp update. My suggestion is to mark this bug as "needsinfo" or "question" until cl-test-grid tests the next Quicklisp update of STMX.

Sep 23 '13 20:09 cosmos72

I have checked out the lates stmx and running tests on "quicklisp 2013-08-13 + stmx.f1b9fa233a".

BTW, good status name for tickets where fix is commited, but hasn't yet been verified could be "fix-commited"

Sep 23 '13 22:09 avodonosov

Good suggestion, thanks. About the test suite, see also issue #5 - I fixed what I could, but there are known problems in ECL 13.5.1 with bordeaux-threads 0.8.3

Sep 23 '13 23:09 cosmos72

I have performed the stmx tests on the lisps I have on our linux test machine. Here is the report: http://common-lisp.net/project/cl-test-grid/stmx/stmx.f1b9fa233a.html

Sep 23 '13 23:09 avodonosov

So it seems in the very recent version of ECL, ecl-13.5.1-44c86e08, everything passed. Elder versions have some problems.

BTW, to clarify the the report above. "quicklisp 2013-08-13 + stmx.f1b9fa233a" is not affected by the issue #5 as we only tested stmx this time, which means bordeaux-threads FASLS were not compiled before stmx for some other library. Therefore the compilation failures are absent for "quicklisp 2013-08-13 + stmx.f1b9fa233a" and we see some test case failures instead.

Sep 23 '13 23:09 avodonosov

On recent ECL snapshots that fix ECL bug causing issue #5, test suite seems to pass more often but it sometimes still hangs - both with and without recent STMX fixes.

I suspect a race condition, investigating it.

Sep 26 '13 23:09 cosmos72