On 05/19/2009 06:02 PM, Karel Zak wrote:
Applied. I did some changes to your test. Unfortunately, the "Delete
extended partition" returns a different MD5 sum. Please,
Totally my bad. I overlooked the failure in the original submitted file.
The correct md5sum (as indicated by running with set -x) is:
-16a9e8009ee95d2e2341e6657acee4c2 doslabel.img
+fc3cdb12326656d7996b09b6f76973e7 doslabel.img
it's better to use "ts_finalize" function to check diffs between
expected/fdisk/doslabel and real output at output/fdisk/doslabel
OK.
Please, use git-send-email. Really. Your mail client produces
malformed mails (patches).
I will. Not really using git that often, so I did not even have
email/name set up for proper patch submission. Already caught up on
those little details. The client does not malform the patch, but it
certainly does break long lines into smaller. I should've realised that
before sending. :/
you need to add, TS_CMD_FDISK to commands.sh.in :-)
What i hate about this test is that fdisk, on each run, calls partition
rediscovery ioctl() (which takes several seconds to complete), making
the test run incredibly long, even though it does not make sense (it's
called on a binary file, and no partitions will be discovered anyway).
Before adding the test to the default list of tests, I would either: a)
add a commandline option to explicitly forbid running partition
rediscovery b) check if we're running on a file rather than block
device, and forbid using all ioctls internally.
Anyway, for now the tests should primarily make sure that no regression
happens during the rewrite/refactoring, so running them manually seems
sufficient.
Zdenek
--
To unsubscribe from this list: send the line "unsubscribe util-linux-ng" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
