In the current Logtalk git version, the testing automation scripts, logtalk_tester.sh and logtalk_tester.ps1, now accept a -j option for concurrent run of tests sets. This is essential for quick development feedback given the growing number of test sets (+660) and tests (+16K) in the Logtalk distribution. A typical call is:
$ cd ~/logtalk
$ logtalk_tester -p trealla -j 8 -o minimal
But I'm seeing a small number of random test sets failing or crashing that cannot be reproduced when run in isolation. Recently, the same issue surfaced and was fixed in another Prolog system. The bug was a race condition when starting multiple Prolog processes concurrently due to use at startup of temporary directories/files. Wondering if the same issue may be present in Trealla Prolog.
In the current Logtalk git version, the testing automation scripts,
logtalk_tester.shandlogtalk_tester.ps1, now accept a-joption for concurrent run of tests sets. This is essential for quick development feedback given the growing number of test sets (+660) and tests (+16K) in the Logtalk distribution. A typical call is:But I'm seeing a small number of random test sets failing or crashing that cannot be reproduced when run in isolation. Recently, the same issue surfaced and was fixed in another Prolog system. The bug was a race condition when starting multiple Prolog processes concurrently due to use at startup of temporary directories/files. Wondering if the same issue may be present in Trealla Prolog.