On 11/14/17 2:52 PM, Dave Chinner wrote: > On Tue, Nov 14, 2017 at 07:31:34AM -0600, Eric Sandeen wrote: >> >> >> On 11/14/17 12:32 AM, Nikolay Borisov wrote: >>> >>> >>> On 14.11.2017 00:22, Eric Sandeen wrote: >>>> On 11/13/17 4:05 PM, Nikolay Borisov wrote: >> >> ... >> >>>>> You are right this is a bit ugly, but it seems more consistend to me, >>>>> rather than something like >>>>> >>>>> xfs_io -c "fiemap -r -n 10 -v 4k 10k" for example. And the reason why >>>>> I'm using this hackish way and not declaring r as r: in getopt is >>>>> because getopt doesn't recognise when a parameter takes more than 1 >>>>> argument. >>>> >>>> Sorry for chiming in so late after all the go-rounds, but: >>>> >>>> Why not just drop -r entirely, and make fiemap go into ranged mode iff a >>>> range is specified at the end of the command, i.e.: >>>> >>>> fiemap [ -alv ] [ -n nx ] [ offset length ] >>> >>> V2 was like that but it required some changing to the generic code which >>> detects the presence of this feature and Dave objected hence v3. >> >> Ok, sorry for being so behind and rehashing old discussions. Let me look >> at that. > > I suggested "-r" because of the fact that you can't detect optional > parameters in fstests without jumping through really nasty hoops > involving regex line noise. Yeah, but I'd rather have nasty stuff in xfstests than in xfsprogs interfaces. See also my suggestion to try mapping past EOF as a detection method. -Eric > Cheers, > > Daves. > -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
