[m-rev.] for post-commit review: Handle ENOTEMPTY == EEXIST in MR_errno_name().

Zoltan Somogyi zoltan.somogyi at runbox.com
Wed Apr 30 13:16:13 AEST 2025



On Wed, 30 Apr 2025 13:11:15 +1000, Peter Wang <novalazy at gmail.com> wrote:
> > > diff --git a/tools/generate_errno_name b/tools/generate_errno_name
> > > index 2337815b2..089deb3d9 100755
> > > --- a/tools/generate_errno_name
> > > +++ b/tools/generate_errno_name
> > > @@ -49,7 +49,6 @@ ERANGE
> > >  EDEADLK
> > >  ENAMETOOLONG
> > >  ENOLCK
> > > -ENOTEMPTY
> > >  ELOOP
> > >  ENOMSG
> > >  EIDRM
> > 
> > This change means that on systems where ENOTEMPTY exists
> > and has a value that is NOT shared with other errnos, get_system_error_name
> > won't recognize it.
> 
> I don't see how, e.g. on Linux, ENOTEMPTY and EEXIST have different
> values, and both would be recognized.

Yes, you are right. I wrote that before I read the second half of the diff,
and then forgot to go back to delete it when I recognized I was wrong.

Zoltan.


More information about the reviews mailing list