launchpad.net/juju-core/worker/uniter/jujuc tests fail with gccgo
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| juju-core |
Fix Released
|
High
|
Unassigned | ||
Bug Description
The output isn't super clear but this might just be an error being reported differently?
2013-11-19 03:31:40 INFO juju juju-log.go:66 u/0: msg
-------
FAIL: server_test.go:92: net_juju_
server_test.go:96:
c.Assert(err, jc.Satisfies, os.IsNotExist)
... obtained *os.PathError = &os.PathError{
... func(T) bool func(error) bool = (func(error) bool)(0x7f31b9c
-------
PANIC: server_test.go:142: net_juju_
... Panic: Fixture has panicked (see related PANIC)
OOPS: 32 passed, 1 FAILED, 7 MISSED
--- FAIL: TestPackage (0.25 seconds)
FAIL
FAIL launchpad.
| John A Meinel (jameinel) wrote | #1 |
| Changed in juju-core: | |
| importance: | Undecided → High |
| status: | New → Triaged |
| Michael Hudson-Doyle (mwhudson) wrote | #2 |
| Michael Hudson-Doyle (mwhudson) wrote | #3 |
| Michael Hudson-Doyle (mwhudson) wrote | #4 |
| Changed in juju-core: | |
| status: | Triaged → Fix Released |
This does appear to be an error reporting failure. However, we are using the OS functions provided (hopefully implemented correctly by gccgo).
We are doing an _, err = os.Open(path), and then expecting it to be os.IsNotFoundEr
Now on *my* machine errcode 0x6 is ENXIO which doesn't look like ENOENT to me.
Given the description here: http://
Getting ENXIO sounds *really* strange (the device is not connected)? Is it possible that this is a failure because of how your VM is being set up?
Either that or GCCGO is mapping error codes incorrectly somehow.