[PATCH blktests] nvme/046: change nvme io-passthru command option from -o to --opcode

Daniel Wagner dwagner at suse.de
Fri Feb 2 01:03:32 PST 2024


On Thu, Feb 01, 2024 at 06:17:01PM +0000, Chaitanya Kulkarni wrote:
> On 2/1/24 07:20, Keith Busch wrote:
> > On Wed, Jan 31, 2024 at 03:34:14PM +0100, Daniel Wagner wrote:
> >> On Fri, Jan 26, 2024 at 03:10:37PM +0900, Shin'ichiro Kawasaki wrote:
> >>> A recent commit in nvme-cli v2.6 changed the single letter of the
> >>> --opcode option from -o to -O. This caused the failure of nvme/046. To
> >>> make the test case work regardless of the nvme-cli version, replace -o
> >>> with --opcode.
> >>>
> >>> Signed-off-by: Shin'ichiro Kawasaki <shinichiro.kawasaki at wdc.com>
> >> Sorry about this. We introduced a couple of global flags, -v (verbose)
> >> and -o (output) and this changed a bunch commands. This was done in the
> >> quest to add JSON output format support add to all commands.
> >>
> >> I will refrain from doing more like this for the 2.x cycle.
> > Changing intput/output is always going to trip someone up. Maintaining
> > backward compatibility vs. improving the design is a challenge.
> >
> > IMO, though, scripts should always use the long options. The short
> > options are convenient for hacking away at a terminal, but they're not
> > readily readable. The long options are more descriptive for when you
> > have to maintain the script in the future.

Good point. I'll keep that in mind.

> I'll keep in mind while reviewing patches, let's make the policy to use
> the long options for blktests at least.

Yes, makes sense.

> Daniel, is it possible to update nvme mailing list when you release new 
> version
> of nvme-cli/libnvme so someone can run the blktests and make sure things
> are stable ?

Sure, will do.

I hope to get blktests hooked up into the CI build, so that we can catch
those problems earlier. Dennis made good progress recentely on this
project. Let's see how it goes.

FWIW, you can also get notification from github whenever there is a new
release. As far I can tell all distributions rely on this feature
nowadays. E.g. when I do a release a few minutes later Arch usually
ships the new version :)



More information about the Linux-nvme mailing list