Add support for Go 1.20 and 1.21

[AaronH88]

This PR adds support for Go 1.20 and 1.21,
Updates to the latest version of quic-go along with fixing the breaking changes as a result,
Also introduces a new mutex lock for the packet conns readDeadline field, which was not thread safe before.

This PR also addresses the following CVE's

• CVE-2023-45287
• CVE-2023-39326

[codecov]

Codecov Report

Attention: 5 lines in your changes are missing coverage. Please review.

Comparison is base (4ca9363) 33.30% compared to head (a81c19a) 33.44%.

@@            Coverage Diff             @@
##            devel     #914      +/-   ##
==========================================
+ Coverage   33.30%   33.44%   +0.14%     
==========================================
  Files          40       40              
  Lines        7399     7409      +10     
==========================================
+ Hits         2464     2478      +14     
+ Misses       4792     4788       -4     
  Partials      143      143              

Files	Coverage Δ
pkg/netceptor/packetconn.go	90.10% <100.00%> (+0.57%)	⬆️
pkg/netceptor/conn.go	0.00% <0.00%> (ø)

... and 2 files with indirect coverage changes

[kapouer]

Building receptor on debian/trixie (with or without your patch) gives the same two failing test_workunit.py tests.
Running those two tests independently, or skipping test_mesh.py, let the tests pass.
So it's probably something about a socket not being properly closed between two tests.

[sonarcloud]

Quality Gate passed

Kudos, no new issues were introduced!

0 New issues
0 Security Hotspots
No data about Coverage
0.0% Duplication on New Code

See analysis details on SonarCloud

[AaronH88]

Building receptor on debian/trixie (with or without your patch) gives the same two failing test_workunit.py tests. Running those two tests independently, or skipping test_mesh.py, let the tests pass. So it's probably something about a socket not being properly closed between two tests.

Thanks so much for the hint,
You were correct, the test right before the failing two was meant to clean up a work unit, but the test suite continued too fast, adding a sleep after the cleanup step fixed the problem.

[matoval]

LGTM