Closed Bug 372845 Opened 17 years ago Closed 16 years ago

Floating point issues on "Linux qm-rhel02 dep unit test"

Categories

(Release Engineering :: General, defect)

x86
Linux
defect
Not set
major

Tracking

(Not tracked)

RESOLVED INVALID

People

(Reporter: bzbarsky, Unassigned)

Details

The regression test for bug 372770 is failing on "Linux qm-rhel02 dep unit test" while passing on both other test boxen and locally.  The error when it fails is consistent with the FPU producing bogus results, as far as I can tell.

Are there FPU precision issues on this box?  If so, could we resolve them?

If not, what can I do to figure out why this is failing?  Is there any way to run an interactive debug session on this box?

Note that I'm disabling the tests for now so the tree can go green...
robcee, any idea what's up here?
Pretty sure this should be on build for troubleshooting as this is not a simple "restart tinderbox".  Moving...
Assignee: server-ops → build
Component: Server Operations: Tinderbox Maintenance → Build & Release
QA Contact: justin → preed
Actually these are administered by robcee; moving to Core/Testing :)
Assignee: build → nobody
Component: Build & Release → Testing
Product: mozilla.org → Core
QA Contact: preed → testing
Version: other → unspecified
Note that if we need to fix the code, we should fix the code from bug 372770 and from bug 372782.
I have no idea if there are any FPU precision issues on the VM and unsure how I'd check such a thing (other than throwing around some large floating point numbers and checking the results). As it's a VM with some other machines running on the same host, it's unlikely a hardware issue.

so, I'm looking for guidance on how to get this fixed. Boris, if you need debugging access to the machine, we can probably set that up.
> and unsure how I'd check such a thing 

Run a debug build and see whether you get the FPU precision warnings from the JS engine?  Or duplicate the code they use to test this in a small app and run it?

If you can get me access to just debug on that machine, that would rock.
Assignee: nobody → preed
Wrong window; didn't mean to reassign to me.
Assignee: preed → nobody
Tests are no longer running on qm-rhel02. Closing.
Status: NEW → RESOLVED
Closed: 16 years ago
Resolution: --- → INVALID
Component: Testing → Release Engineering
Product: Core → mozilla.org
QA Contact: testing → release
Version: unspecified → other
Product: mozilla.org → Release Engineering
You need to log in before you can comment on or make changes to this bug.