Eric Sandeen <sandeen@xxxxxxxxxxx> 于2020年6月4日周四 下午9:56写道：
>
> On 6/4/20 4:57 AM, Yong Sun wrote:
> > openSUSE and SLE don't support username begin with digit, so it will
> > skip test generic/597 and generic/598 by lack of 123456-fsgqa user.
> > generic/597 and 598 are not test username begin with digit on purpose
> > (different with generic/381). It's will be helpful to use an username
> > begin with non-digit in this case.
> >
> > Signed-off-by: Sun Yong <yosun@xxxxxxxx>
> > ---
> > README | 1 +
> > tests/generic/597 | 7 ++-----
> > tests/generic/598 | 7 ++-----
> > 3 files changed, 5 insertions(+), 10 deletions(-)
> >
> > diff --git a/README b/README
> > index 094a7742..cffa0bc6 100644
> > --- a/README
> > +++ b/README
> > @@ -22,6 +22,7 @@ _______________________
> > - create fsgqa test user ("sudo useradd -m fsgqa")
> > - create fsgqa group ("sudo groupadd fsgqa")
> > - create 123456-fsgqa test user ("sudo useradd 123456-fsgqa")
> > +- create fsgqa-123456 test user ("sudo useradd fsgqa-123456")
> >
> > ______________________
> > USING THE FSQA SUITE
> > diff --git a/tests/generic/597 b/tests/generic/597
> > index 1d87a23a..11846b95 100755
> > --- a/tests/generic/597
> > +++ b/tests/generic/597
> > @@ -41,13 +41,10 @@ _supported_os Linux
> > _require_test
> > _require_sysctl_variable fs.protected_symlinks
> > _require_sysctl_variable fs.protected_hardlinks
> > -# su in _require_user prints warnings about user name starts with a digit,
> > -# discard the warning
> > -_require_user 123456-fsgqa >/dev/null 2>&1
> > -# Do this SECOND so that qa_user is fsgqa, and _user_do uses that account
>
> This comment is still relevant. It's a quirk that _user_do uses the second
> _require_user stated[1], and I wanted to highlight that in the comment.
>
I see, I'll left this comments in v3.
> Otherwise, TBH I guess I'd rather see the user simply be "fsgqa2" or something,
> so it's not easily confused with the 123456-fsgqa user. Not a technical problem
> with this patch though, just a preference.
>
> Alternately, these two tests /might/ be able to run just as well with "root"
> as one user or the other, but I didn't want to risk accidentally getting the
> permissions tests wrong by using a user with elevated privs.... so I'm not
> sure about that option.
>
IMO fsgqa2 is better than root. root may misleading test motivation.
Also It's will be
more convenience to choose when people need a second user in other tests.
> -Eric
>
> [1]
> # check for a user on the machine, fsgqa as default
> #
> _require_user()
> {
> qa_user=fsgqa
> if [ -n "$1" ];then
> qa_user=$1
> fi
>
> ...
>
> _user_do()
> {
> echo $1 | su -s /bin/bash $qa_user 2>&1 | _filter_user_do
> }
>
> tho it might be better to enhance _user_do to take a username as an optional
> (or required) argument, to avoid this complexity/confusion.... that's a different
> patch tho.
+1, this may need a different patch.
Thanks,
Sero
>
> > +_require_user fsgqa-123456
> > _require_user fsgqa
> >
> > -OWNER=123456-fsgqa
> > +OWNER=fsgqa-123456
> > OTHER=fsgqa
