On Mon, Jun 18, 2018 at 8:16 AM, Jim Meyering <[email protected]> wrote: > On Sun, Jun 17, 2018 at 9:41 PM, Pádraig Brady <[email protected]> wrote: >> On 17/06/18 21:16, Jim Meyering wrote: >>> On Sun, Jun 17, 2018 at 8:44 PM, Pádraig Brady <[email protected]> wrote: >>>> I'll apply the attached full patch soon. >>> >>> Thanks for doing all that. I like it. >>> One nit: this says "on a separate device": >>> >>> + --preserve-root[=all] do not remove '/' (default); with 'all', >>> skip\n\ >>> + any command line argument on a separate >>> device\n\ >>> >>> Perhaps it should say something like "on a separate device from its >>> parent" ? Otherwise, one might erroneously interpret "separate" as >>> meaning "different from the device of other command line arguments". >>> >>> I wondered how the code tested for this attribute, and had to read the >>> code to find that it stats both the command-line argument, A, and >>> "A/..", and if they have different device IDs, then A is eligible to >>> be skipped when this new option is in effect. >> >> Yes the texinfo mentions the parent. >> I'll expand --help to mention that also: >> >> --preserve-root[=all] do not remove '/' (default); >> with 'all', skip any command line argument >> on a separate device from its parent > > Thanks, clearly I missed the texinfo addition :-) > Speaking of that, please use the active voice, not passive, i.e., change this > > When @samp{all} is specified, any command line argument that > is on a separate file system to its parent is also skipped. > > to this: > > When @samp{all} is specified, skip any command line argument that > is not on the same file system as its parent.
Actually, please also change s/skip/reject/; that is more appropriate, since that rejection results in a nonzero exit status.
