| Ticket UUID: | 3389684 | ||
| Title: | clock-33.11* timing sensitive | ||
| Type: | Bug | Created on: | 2011-08-10 16:25:00 |
| Submitter: | dgp | Assigned to: | kennykb |
| Subsystem: | 06. Time Measurement | Severity: | |
| Priority: | 1 Zero | Last modified: | 2011-08-18 04:36:32 |
| Status: | Open | Closed by: | |
| Resolution: | None | Closed on: | |
| Version: | obsolete: 8.6b2 | ||
| Description: | ||||
Tests clock-33.11* hang on any system where Tcl runs sufficiently slowly that every pass through the [while] body takes more than a millisecond. I find that to be true on mainstream, modern systems when valgrinding. | ||||
| User Comments: | ||||
dgp added on 2011-08-11 01:12:23:
Only valgrinding with --enable-symbols=mem is so slow. And that's apparently a completely wrong config to do anything useful with, so ignore this until you have nothing else to do. :) ferrieux added on 2011-08-18 04:36:32:
Also note that these two tests, even when they work, have a strong sampling bias. | ||||
