On 07/12/2016 10:45 AM, John Ferlan wrote:
On 07/01/2016 07:37 AM, Cole Robinson wrote:
> This series reworks test:///default open handling to use the same
> XML parsing helpers that a non-default test://$PATH open uses.
> Saves some code, and makes it easier to extend test:///default,
> for example in patch #4 which we want for testing nodedev events
>
> Cole Robinson (4):
> test: Introduce testOpenParse
> test: Move testOpenDefault definition later
> test: Have test:///default open use file parsing helpers
> test: Add scsi vport nodedev to test:///default
>
> src/test/test_driver.c | 326 +++++++++++++++++++++----------------------------
> 1 file changed, 141 insertions(+), 185 deletions(-)
>
Nice...
Something for a "todo" list - adding snapshots for domains (to test
testParseDomainSnapshots) and volumes for pools (to test
testOpenVolumesForPool) for defaultConnXML.
w/r/t: 4/4... If 'wwnn' and 'wwpn' weren't supplied, then
virRandomGenerateWWN should have done the trick; however,
testParseNodedevs passes a NULL virt_type so the generation fails. Part
of me wonders if should we create a TEST_DRIVER_OUI using "200000" as
the prefix just to "prove" that the code path works. Your call though -
easy enough to mock up a virRandomGenerateWWN or modify the non mocked
version to support the test driver.
ACK series regardless of how you handle 4/4.
Thanks for the reviews. I'd pushed the patches as posted. I agree getting some
extra coverage for the WWN generation would be a nice addition, but this is
blocking some of the incoming GSOC work and I'm short on time at the moment...
Thanks,
Cole