| Message ID | 20210418034958.505267-1-Tony.Ambardar@gmail.com |
|---|---|
| State | New |
| Headers | show |
| Series | [iproute2] ip: drop 2-char command assumption | expand |
On 4/17/21 8:49 PM, Tony Ambardar wrote: > The 'ip' utility hardcodes the assumption of being a 2-char command, where > any follow-on characters are passed as an argument: > > $ ./ip-full help > Object "-full" is unknown, try "ip help". > > This confusing behaviour isn't seen with 'tc' for example, and was added in > a 2005 commit without documentation. It was noticed during testing of 'ip' > variants built/packaged with different feature sets (e.g. w/o BPF support). > > Drop the related code. > > Fixes: 351efcde4e62 ("Update header files to 2.6.14") > Signed-off-by: Tony Ambardar <Tony.Ambardar@gmail.com> > --- > ip/ip.c | 3 --- > 1 file changed, 3 deletions(-) > > diff --git a/ip/ip.c b/ip/ip.c > index 4cf09fc3..631ce903 100644 > --- a/ip/ip.c > +++ b/ip/ip.c > @@ -313,9 +313,6 @@ int main(int argc, char **argv) > > rtnl_set_strict_dump(&rth); > > - if (strlen(basename) > 2) > - return do_cmd(basename+2, argc, argv); > - > if (argc > 1) > return do_cmd(argv[1], argc-1, argv+1); > > popular change request lately. As I responded to Matteo in January: This has been around for too long to just remove it. How about adding an option to do_cmd that this appears to be guess based on basename? If the guess is wrong, fallback to the next do_cmd.
On Tue, 20 Apr 2021 01:26:36 -0700 Tony Ambardar <tony.ambardar@gmail.com> wrote: > The 'ip' utility hardcodes the assumption of being a 2-char command, where > any follow-on characters are passed as an argument: > > $ ./ip-full help > Object "-full" is unknown, try "ip help". > > This confusing behaviour isn't seen with 'tc' for example, and was added in > a 2005 commit without documentation. It was noticed during testing of 'ip' > variants built/packaged with different feature sets (e.g. w/o BPF support). > > Mitigate the problem by redoing the command without the 2-char assumption > if the follow-on characters fail to parse as a valid command. > > Fixes: 351efcde4e62 ("Update header files to 2.6.14") > Signed-off-by: Tony Ambardar <Tony.Ambardar@gmail.com> > --- > v2: (feedback from David Ahern) > * work around problem but remain compatible with 2-char assumption I am ok with this, but if you change the name of command, you can expect some friction (and non support). The original commit was inherited from the original integration of tarball's into BitKeeper. This "feature" was put in by Alexey Kuznetsov back in orignal 2.4 time frame.
On Tue, 20 Apr 2021 at 08:16, Stephen Hemminger <stephen@networkplumber.org> wrote: > > On Tue, 20 Apr 2021 01:26:36 -0700 > Tony Ambardar <tony.ambardar@gmail.com> wrote: > > > The 'ip' utility hardcodes the assumption of being a 2-char command, where > > any follow-on characters are passed as an argument: > > > > $ ./ip-full help > > Object "-full" is unknown, try "ip help". > > > > This confusing behaviour isn't seen with 'tc' for example, and was added in > > a 2005 commit without documentation. It was noticed during testing of 'ip' > > variants built/packaged with different feature sets (e.g. w/o BPF support). > > > > Mitigate the problem by redoing the command without the 2-char assumption > > if the follow-on characters fail to parse as a valid command. > > > > Fixes: 351efcde4e62 ("Update header files to 2.6.14") > > Signed-off-by: Tony Ambardar <Tony.Ambardar@gmail.com> > > --- > > v2: (feedback from David Ahern) > > * work around problem but remain compatible with 2-char assumption > > I am ok with this, but if you change the name of command, you can expect some > friction (and non support). > The renamed binaries are normally accessed via an 'ip' symlink managed at package installation, so shouldn't be an issue. I only discovered the problem while running regression tests on the the underlying binaries, and thought to fix things. > The original commit was inherited from the original integration of tarball's > into BitKeeper. This "feature" was put in by Alexey Kuznetsov back in orignal 2.4 > time frame. Thanks for the feedback and additional detail, Tony
diff --git a/ip/ip.c b/ip/ip.c index 4cf09fc3..631ce903 100644 --- a/ip/ip.c +++ b/ip/ip.c @@ -313,9 +313,6 @@ int main(int argc, char **argv) rtnl_set_strict_dump(&rth); - if (strlen(basename) > 2) - return do_cmd(basename+2, argc, argv); - if (argc > 1) return do_cmd(argv[1], argc-1, argv+1);
The 'ip' utility hardcodes the assumption of being a 2-char command, where any follow-on characters are passed as an argument: $ ./ip-full help Object "-full" is unknown, try "ip help". This confusing behaviour isn't seen with 'tc' for example, and was added in a 2005 commit without documentation. It was noticed during testing of 'ip' variants built/packaged with different feature sets (e.g. w/o BPF support). Drop the related code. Fixes: 351efcde4e62 ("Update header files to 2.6.14") Signed-off-by: Tony Ambardar <Tony.Ambardar@gmail.com> --- ip/ip.c | 3 --- 1 file changed, 3 deletions(-)