Re: [PATCH] fuzz: fix unbound variable in build.sh

[Thomas Huth]

On 07/09/2021 14.51, Alexander Bulekov wrote:
> On 210907 1432, Thomas Huth wrote:
> > On 07/09/2021 13.08, Alexander Bulekov wrote:
> > > /src/build.sh: line 76: GITLAB_CI: unbound variable
> > > Fix that.
> > >
> > > Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
> > > ---
> > >
> > > This change is in preparation to revert:
> > > 7602748c ("qemu: manually build glib (#5919)") on OSS-Fuzz.
> > > Reverting as-is produces an unbound variable complaint when we try to
> > > build the fuzzers in the OSS-Fuzz container.
> > >
> > >    scripts/oss-fuzz/build.sh | 2 +-
> > >    1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/scripts/oss-fuzz/build.sh b/scripts/oss-fuzz/build.sh
> > > index 98b56e0521..5ddc769c9c 100755
> > > --- a/scripts/oss-fuzz/build.sh
> > > +++ b/scripts/oss-fuzz/build.sh
> > > @@ -73,7 +73,7 @@ if ! make "-j$(nproc)" qemu-fuzz-i386; then
> > >              "\nFor example: CC=clang CXX=clang++ $0"
> > >    fi
> > > -if [ "$GITLAB_CI" != "true" ]; then
> > > +if [ -z ${GITLAB_CI+x} ]; then
> >
> > My bash-foo is really not the best, but shouldn't there be a colon in there,
> > i.e. ${GITLAB_CI:+x} ?
>
> I think the difference is that GITLAB_CI+x only checks if GITLAB_CI is
> set, while GITLAB_CI:+x checks that it is set and non-null.
> https://pubs.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html#tag_18_06_02
>
> I don't think that makes much of a difference here.

TIL, and I agree that it does not make a difference here (and if it would, 
your variant is certainly better).

Reviewed-by: Thomas Huth <thuth@redhat.com>