BUG: Clear CONTEXT_THREAD_KEY when destroying the PJ_CONTEXT

[dgburr]

The switch to thread-local contexts in #1419 stores a pointer to the newly created PJ_CONTEXT using Thread Specific Storage. However the value stored in TSS is not cleared when the PJ_CONTEXT is destroyed, meaning that a subsequent call to pyproj_context_create() will return a pointer to an invalid PJCONTEXT, leading to a segmentation fault.

My specific use case is in a multithreaded QT app with symptoms similar to #1509. However this issue can occur in any multithreaded app which performs multiple calls to pyproj_context_create() from the same thread.

• [❌] Tests added
• [✅] Fully documented, including history.rst for all changes and api/*.rst for new API

[dgburr]

@snowman2: should I also try to fix the linting issue in this PR, or should I create another one?

[snowman2]

@snowman2: should I also try to fix the linting issue in this PR, or should I create another one?

A separate PR would be perfect.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.84%. Comparing base (b78a50d) to head (d0fdb83).
⚠️ Report is 19 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1541      +/-   ##
==========================================
- Coverage   96.55%   95.84%   -0.72%     
==========================================
  Files          20       20              
  Lines        1827     1827              
==========================================
- Hits         1764     1751      -13     
- Misses         63       76      +13     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[snowman2]

Thanks @dgburr 👍