KIO autotests: repair JobTest::moveDestAlreadyExistsAutoRename
ClosedPublic

Authored by dfaure on Mar 21 2020, 3:23 PM.

Details

Summary

It failed for me, and for FreeBSD on CI.
I guess the Linux CI has it all on the same partition, so the moving
across partition case isn't really tested there.

See explanations in the code comment.

Test Plan

Now jobtest passes locally

Diff Detail

Repository
R241 KIO
Branch
master
Lint
No Linters Available
Unit
No Unit Test Coverage
Build Status
Buildable 24025
Build 24043: arc lint + arc unit
dfaure created this revision.Mar 21 2020, 3:23 PM
Restricted Application added a project: Frameworks. · View Herald TranscriptMar 21 2020, 3:23 PM
Restricted Application added a subscriber: kde-frameworks-devel. · View Herald Transcript
dfaure requested review of this revision.Mar 21 2020, 3:23 PM
apol accepted this revision.Mar 21 2020, 4:02 PM
apol added a subscriber: apol.

I confirm that the test gets fixed.
Also the change makes sense.

It seems to me that both else branches end the same. From list = ... it can be moved out, no?

This revision is now accepted and ready to land.Mar 21 2020, 4:02 PM

They are not the same.

(3,4)
vs
(2,4)

dfaure closed this revision.Mar 21 2020, 4:09 PM

It was failing locally for me too, both when I had /tmp as tmpfs and when I changed it to be a regular dir on /; /home is a separate partition on my system. So I didn't understand how the test passed on the CI.

FTR, I build KIO on a regular mechanical/spinning HDD, but / and /home are on an SSD. Maybe that's why? the test wants two separate partitions on the same desk? that doesn't make sense though...

It was failing locally for me too, both when I had /tmp as tmpfs and when I changed it to be a regular dir on /; /home is a separate partition on my system. So I didn't understand how the test passed on the CI.

Because on the CI (which is a LXC container), /home isn't a separate partition.

So the tests for "moving between different partitions" are actually testing (again) moving on the same partition, on CI, because there is just one partition.

FTR, I build KIO on a regular mechanical/spinning HDD, but / and /home are on an SSD. Maybe that's why? the test wants two separate partitions on the same desk? that doesn't make sense though...

Not sure what you mean by desk, but I hope the above explains the reason why the test was passing on CI and failing for most of us locally.

It was failing locally for me too, both when I had /tmp as tmpfs and when I changed it to be a regular dir on /; /home is a separate partition on my system. So I didn't understand how the test passed on the CI.

Because on the CI (which is a LXC container), /home isn't a separate partition.

So the tests for "moving between different partitions" are actually testing (again) moving on the same partition, on CI, because there is just one partition.

FTR, I build KIO on a regular mechanical/spinning HDD, but / and /home are on an SSD. Maybe that's why? the test wants two separate partitions on the same desk? that doesn't make sense though...

Not sure what you mean by desk, but I hope the above explains the reason why the test was passing on CI and failing for most of us locally.

Oops, I meant disk.

That makes sense; most of us would have /home on a separate partition.