CVE-2022-45693: https://nvd.nist.gov/vuln/detail/CVE-2022-45693

CVE-2022-41915: https://nvd.nist.gov/vuln/detail/CVE-2022-41915

* Check IPs for LOAD CSV during openStream () ()

* Fix bug with connection state on valid redirects.

If the user has not requested a rounding mode and the precision is 0, we still want to fallback to Java Math.round(), but that does not have the rounding mode 'HALF_UP', but a custom one where it round ties towards positive infinity.

If the user has request a precision >= 1 and no mode, we still want to use 'HALF_UP' to not break current behavior.

Co-authored-by: NLars Korduner <lars.korduner@neotechnology.com>

* [fix] Don't crash DbRepresentation when Node is deleted in parallel thread
Co-authored-by: NHannes Sandberg <hannes.sandberg@neotechnology.com>

* Checkstyle
Co-authored-by: NHannes Sandberg <hannes.sandberg@neotechnology.com>

* Updated IR eagerization analysis to handle property functions.

* Added logic to handle property functions on Map and self references.

* Checks if expression within property function is map and not if property function is map.

In case we end up mutating it from multiple threads again.

These objects contain mutable state in the form of `CachedFunction`s,
which makes them unsafe to use from multiple threads.

Another issue is that `CachedFunction` would accumulate semantic tables
from different planning runs unnecessarily.

Disable TCK tests that use Europe/Stockholm timezone before ~1870 for now as the behaviour is dependent on the tzinfo file in the JVM

Vulnerability reported https://trello.com/c/DJ9qzBDC/252-vulnerabilities-reported-by-external-party

We are removing these tests because they are blocking the CI due to failing in some of the compatibility builds

(Undo commit 5bbf935)

https://www.cve.org/CVERecord?id=CVE-2022-42889

* Add FileURLAccessRuleTest cases

Improve test coverage to include different inputs:

- non escaped
- single escaped
- double escaped
- triple escaped
- overlong escaped

Also introduce tests which highlight a bug in the FileURLAccessRule logic, by allowing paths to be traversed.

* Add LoadCsvAcceptanceTest cases

Added main test cases to this file. Coverage isn't as extensive as in FileUrlAccessTest because these integration tests are slow.

Also fixed some of the existing tests.

* Normalize URI in FileURLAccessRule

The path was being normalised but not the URI. The Java Path and URI classes originate from two different RFC documents which have different concerns and different normalisation algorithms. We need to normalise the URI before we construct a path with it, in order to protect the DBMS from potential vulnerabilities.

* Adapt tests to the Windows OS

(cherry picked from commit 84591dea0a36531175b5a49c2db8b914087cd2e8)

Cleanup FileUrlAccessRuleTest cases

(cherry picked from commit df93740e871c3852b86db92f40cad063757b1eb6)

* Reword 'Unsupported clauses' error message

* Fix test cases

Ignores some failing provided order tests. The test cases are correct, as far
as I understand, but are not fixed as part of this branch.

Use old way of rounding cardinality up to 1 for SemiApply and AntiSemiApply to avoid regressions
Co-authored-by: NArne Fischereit <arnefischereit@users.noreply.github.com>

Fix bug in pipelined runtime where AssertingMultiNodeIndexSeekOperator failed
to perform locking index seeks. This could lead to failures when running
concurrent transactions in queries like `MERGE (n:A:B {id: x})` when we have
a uniqueness constraint on id for both labels A and B.

Fix bug in pipelined runtime where some queries using AssertingMultiNodeIndexSeekOperator would fail incorrectly with `Merge did not find a matching node...`. For example in queries like `UNWIND [0,1] AS id MERGE (n:A:B {id: id})`, if there is a uniqueness constraint on the id property for both A and B labels.

Cherry-picks: