On 12.03.26 13:59, Andrew Cooper wrote:
On 12/03/2026 12:22 pm, Juergen Gross wrote:There is no man page for xenstore-list, add it.Signed-off-by: Juergen Gross <[email protected]> --- docs/man/xenstore-list.1.pod | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100644 docs/man/xenstore-list.1.pod diff --git a/docs/man/xenstore-list.1.pod b/docs/man/xenstore-list.1.pod new file mode 100644 index 0000000000..4bc1ff9846 --- /dev/null +++ b/docs/man/xenstore-list.1.pod @@ -0,0 +1,24 @@ +=head1 NAME + +xenstore-list - list Xenstore key children + +=head1 SYNOPSIS + +B<xenstore-list> [I<OPTION>]... [I<PATH>]... + +=head1 DESCRIPTION + +List direct children of one or more Xenstore I<PATH>s, one child per line. + +=over + +=item B<-p> + +List the full Xenstore path of each listed path. + +=back + +=head1 BUGS + +Send bugs to [email protected], see +https://wiki.xenproject.org/wiki/Reporting_Bugs_against_Xen_Project on how to send bug reports.I had no idea we even had this... Why on earth do we have both xenstore-ls and xenstore-list, and for them to be different?
I would think the use cases are different, but this is pure speculation. xenstore-list might be intended to be used by scripts, while xenstore-ls seems to target human readers of the output.
Even for the options; xenstore-ls uses -f for full path and -p for permissions.
Yes, this is weird. And "xenstore-list -p /" will print: //local //tools which is not what I'd expect. :-(
Having both is gross error on behalf of whomever accepted the code in the first place. It's a bad enough usability problem that I think we should seriously consider deleting the binary rather than allowing accepting it's existance. But as I expect that to be controversial, at a minimum both of the manpages need some kind of "do not confuse $THIS with $OTHER" warning.
Its not as if one of those would cause any harm. Juergen
OpenPGP_0xB0DE9DD628BF132F.asc
Description: OpenPGP public key
OpenPGP_signature.asc
Description: OpenPGP digital signature
