#4893 closed defect (worksforme)
tracd crashes silently when 'autoreload' option is enabled
Reported by: | Emmanuel Blot | Owned by: | Jonas Borgström |
---|---|---|---|
Priority: | low | Milestone: | |
Component: | general | Version: | devel |
Severity: | normal | Keywords: | macosx, verify |
Cc: | Branch: | ||
Release Notes: | |||
API Changes: | |||
Internal Changes: |
Description
For some reason, tracd
daemon may crash (the Python interpreter, indeed).
If the autoreload option is enabled (-r
), the standalone daemon behaves as it simply exits, no error message is dumped to the standard error output nor to the log file.
This error occurs, at least, with Python 2.4 on Mac OS X (10.4.8 / Intel).
When the autoreload is disabled and the same error is reproduced, the error cause is actually printed out:
4676 Bus error
I'll try to reproduce this error with other environments
Attachments (0)
Change History (5)
comment:1 by , 18 years ago
Keywords: | macosx verify added |
---|---|
Milestone: | → none |
comment:2 by , 18 years ago
Resolution: | → worksforme |
---|---|
Status: | new → closed |
comment:3 by , 18 years ago
- (empty) milestone → not yet triaged
- 'none' milestone → valid ticket, but can't be associated to a specific milestone (e.g. it concerns a plugin, an external dependency, or like here, when we have no clear idea about which milestone would be appropriate ;-) )
Ideally, once we use 0.11 here, we'll have a good enough WorkFlow so that we wouldn't need to abuse milestones for doing subtle distinctions like this …
comment:4 by , 13 years ago
Milestone: | not applicable |
---|
comment:5 by , 10 years ago
Keywords: | macosx verify → macosx, verify |
---|
I've not been able to reproduce this with Mac OS 10.4.9 and Python 2.5 - for now.
BTW, what's the difference between the (empty) milestone and the 'none' milestone ?