Fix failure in mkdir() during startup

[guedou]

Fixes #4546

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.62%. Comparing base (c38a5de) to head (4556f1a).
Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #4558   +/-   ##
=======================================
  Coverage   81.61%   81.62%           
=======================================
  Files         356      356           
  Lines       85540    85540           
=======================================
+ Hits        69816    69819    +3     
+ Misses      15724    15721    -3     

Files with missing lines	Coverage Δ
scapy/main.py	70.37% <100.00%> (ø)

... and 6 files with indirect coverage changes

[gpotter2]

This issue is a bit ridiculous. But alright.

[KepptnKool]

Will there be a patch level release containing the fix in the near future?

[ctheune]

This issue is a bit ridiculous. But alright.

I'd love to understand why it's ridiculous. It's totally possible that I am missing a /s here ... ;)

To clarify: this is a real problem. We had two concurrent processes in our test suite that ended up racing each other quite often exactly on this line and caused stochastic failures ... so I'm very happy to see this has been caught and released already! :) So, big thanks from me!

[gpotter2]

The code is

if not path.exists():
    path.mkdir()

The fact that this can fail with an "already exists" error is ridiculous to me.

[whoisxy]

The code is

if not path.exists():
    path.mkdir()

The fact that this can fail with an "already exists" error is ridiculous to me.

@gpotter2 TOCTOU

[gpotter2]

If we were in a multi processed code, sure. If you somehow managed to start multiple instances of Scapy, maybe. But that's not the case isn't it?

Closing though.