On 5/13/19 9:20 AM, Max Reitz wrote:
> On 10.05.19 23:45, Nir Soffer wrote:
>> On Sat, May 11, 2019 at 12:19 AM Max Reitz <mre...@redhat.com
>> <mailto:mre...@redhat.com>> wrote:
>>
>> It is possible for an empty file to take up blocks on a filesystem.
>> Make iotest 175 take this into account.
>>
>> Reported-by: Thomas Huth <th...@redhat.com <mailto:th...@redhat.com>>
>> Signed-off-by: Max Reitz <mre...@redhat.com <mailto:mre...@redhat.com>>
>> ---
>> tests/qemu-iotests/175 | 15 +++++++++++----
>> tests/qemu-iotests/175.out | 8 ++++----
>> 2 files changed, 15 insertions(+), 8 deletions(-)
>>
>> diff --git a/tests/qemu-iotests/175 b/tests/qemu-iotests/175
>> index d0ffc495c2..b5652a3889 100755
>> --- a/tests/qemu-iotests/175
>> +++ b/tests/qemu-iotests/175
>> @@ -28,7 +28,8 @@ status=1 # failure is the default!
>>
>> _cleanup()
>> {
>> - _cleanup_test_img
>> + _cleanup_test_img
>> + rm -f "$TEST_DIR/empty"
>> }
>> trap "_cleanup; exit \$status" 0 1 2 3 15
>>
>> @@ -40,18 +41,24 @@ _supported_fmt raw
>> _supported_proto file
>> _supported_os Linux
>>
>> -size=1m
>> +size=$((1 * 1024 * 1024))
>>
>> +
>> +touch "$TEST_DIR/empty"
>> +empty_blocks=$(stat -c '%b' "$TEST_DIR/empty")
>>
>>
>> Maybe extra_blocks?
>
> Why not.
>
>> echo
>> echo "== creating image with default preallocation =="
>> _make_test_img $size | _filter_imgfmt
>> -stat -c "size=%s, blocks=%b" $TEST_IMG
>> +stat -c "size=%s, blocks=%b" $TEST_IMG \
>> + | sed -e "s/blocks=$empty_blocks/nothing allocated/"
>>
>> for mode in off full falloc; do
>> echo
>> echo "== creating image with preallocation $mode =="
>> IMGOPTS=preallocation=$mode _make_test_img $size | _filter_imgfmt
>> - stat -c "size=%s, blocks=%b" $TEST_IMG
>> + stat -c "size=%s, blocks=%b" $TEST_IMG \
>> + | sed -e "s/blocks=$empty_blocks/nothing allocated/" \
>> + | sed -e "s/blocks=$((empty_blocks + size /
>> 512))/everything allocated/"
>>
>>
>> "fully allocated"?
>
> I didn’t like that because that sounds like it only applies to
> preallocation=full.
>
>> Maybe add a helper like this:
>>
>> _filter_blocks() {
>> # Some file systems sometimes allocate extra blocks
>> sed -e "s/blocks=$empty_blocks/nothing allocated/" \
>> -e "s/blocks=$((empty_blocks + size / 512))/everything
>> allocated/"
>> }
>>
>> So we can do:
>>
>> stat -c "size=%s, blocks=%b" $TEST_IMG | _filter_blocks
>>
>> And it is also clear why we need to run sed without looking up the
>> commit message.
>
> Makes sense to me, but I find it a bit awkward to make a filter rely on
> a data value determined outside of the filter... I’ll see what I can do
> to calm my conscience.
>
> Max
>
You can always parameterize the filter so that the relationship is
explicit, no? Does that still feel icky?
--js
