On 2017年09月15日 23:48, David Sterba wrote:
On Fri, Sep 15, 2017 at 09:24:19PM +0800, Qu Wenruo wrote:
I'm going to review & merge this series to devel. Tests and
documentation should be updated to make the usecase clear.
I'm happy to address any comment, both code and doc/test.
For the tests I'd like to see:
* with file target, try non-existent file, zero-length file, too-small
file and large-enough
No problem.
But it would be better to define the expected behavior first.
I'll show the behavior difference below to see if you're OK with it.
For non-existent file:
New: Fail, no such file at the very beginning
Old: Create new one
Zero-length:
New: Fail, too small file at the very beginning
Old: Expand file and continue mkfs
Too-small:
New: Fail, showing -28 errno half way
Old: Same as zero length.
Large-enough:
New: Success, without shrink
Old: Success, shrink
* for block device target, I think loop device would do, too-small and
large-enough
Same as above too-small and large-enough case.
* after the test, verify that the files have same size and contents
I'll reuse convert facilities to do that.
And I'll add one extra test case:
No enough privilege to read some file in the directory:
New: Fail gracefully
Old: Fail with BUG_ON.
If you're OK with the behavior difference, I'll create new test cases
for mkfs --rootdir.
Thanks,
Qu
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
