As reported by Andreas Beckmann in Debian Bug #920995 (https://bugs.debian.org/920995):
> during a test rebuild of spatialite/experimental I noticed that it hangs
> during the virtualknn test. After killing that process the build
> continued and succeeded. I think this has so far happened more than
> once, not sure if on amd64 or or i386 or both.

I've confirmed the issue by rebuilding the spatialite (5.0.0~beta0-1~exp1) Debian package in an up to date Debian experimental (amd64) chroot.

The check_virtualknn process takes 70%-80% CPU, but strace show no activity.

test/check_virtualknn.log contains only:

 CREATE VIRTUAL TABLE "knn" error: table knn already exists

Fixed by Commit 90180e065d

the apparent cause of this bug was some obscure regression introduced by SQLite 3.25 (and following versions).

just rearranging the internal sequence of the tests seems to resolve any issue.