Tcl Source Code

View Ticket
Login
Ticket UUID: 2997967
Title: interp-36.7 timing sensitive?
Type: Bug Created on: 2010-05-07 00:51:41
Submitter: dgp Assigned to: hobbs
Subsystem: 20. [interp] Severity:
Priority: 5 Medium Last modified: 2010-05-07 14:16:39
Status: Open Closed by:
Resolution: None Closed on:
Version: obsolete: 8.6b1.1
Description:
On two occasions in recent days testing the HEAD
on busy machines, I've seen this failure:

==== interp-36.7 SlaveBgerror sets error handler of slave [1999035] FAILED
==== Contents of test case:

    slave eval {
        variable done {}
        after 0 error foo
        after 10 [list ::set [namespace which -variable done] {}]
        vwait [namespace which -variable done]
    }
    set result

---- Result was:
untouched
---- Result should have been (exact matching):
foo
==== interp-36.7 FAILED


suspect the test is timing sensitive.
User Comments:
ferrieux added on 2010-05-07 14:16:39:
Indeed replacing the [after 10] by [after 0] makes it 100% reproducible.
Consequences:

 (1) it indicates that ex-aequo timers are not evaluated in order of insertion. Not sure whether we are breaking any written contract here, but it would be nice to have this extra guarantee.

 (2) as an emergency measure, replacing the [after 10] in the test by [after 1000] is a quick and efficient fix, and doesn't even slow down the test suite in the normal case since this timer fires only in case of failure.