Add caching to guarded_import to boost performance

[jan-jockusch]

Using "import" in Python Scripts triggers the execution of guarded_import each time. This performs a security test and then the import itself. The import is fast because the Python interpreter caches this. The security test is costly, and always yields the same result. This patch inserts a cache to skip this test on repeated executions, speeding up the Python Scripts.

• I signed and returned the Zope Contributor Agreement, and received and accepted an invitation to join a team in the zopefoundation GitHub organization.
• I verified there aren't any other open pull requests for the same change.
• I followed the guidelines in Developer guidelines.
• I successfully ran code quality checks on my changes locally.
• I successfully ran tests on my changes locally.
• If needed, I added new tests for my changes.
• If needed, I added documentation for my changes.
• I included a change log entry in my commits.

---

If your pull request closes an open issue, include the exact text below, immediately followed by the issue number. When your pull request gets merged, then that issue will close automatically.

Closes #

[jan-jockusch]

I cannot push to my branch. Therefore, the changelog is still missing.

[dataflake]

Please try the push again, I just deleted a broken branch protection rule that wantd to cover all possible branches, which makes no sense.

[jan-jockusch]

Please try the push again, I just deleted a broken branch protection rule that wantd to cover all possible branches, which makes no sense.

Yes! Push has worked. Thank you!

[perrinjerome]

This seems a good idea, but for correctness I believe this cache should be invalidated when allow_module is called.

[jan-jockusch]

I'll look into adding clearing the cache on allow_module as suggested.

For motivation: This shaves off 4% of time spent in our scenario, with many scripts in Data.fs utilizing import. YMMV, but I believe this is the near the max value achievable in terms of performance gains, since we use import rather voraciously.