TST: Fix tests for RGF93 v1 -> ETRS89-FRA [RGF93 v1]

[phaarnes]

• Fix tests for renamed coordinate operation in PROJ 9.7+

• ✅ Closes TST: Test failures due to RGF93 v1 -> ETRS89-FRA [RGF93 v1] #1559

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.84%. Comparing base (b78a50d) to head (06d63cf).
⚠️ Report is 35 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1560      +/-   ##
==========================================
- Coverage   96.55%   95.84%   -0.71%     
==========================================
  Files          20       20              
  Lines        1827     1879      +52     
==========================================
+ Hits         1764     1801      +37     
- Misses         63       78      +15     

☔ 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.

[phaarnes]

Hi @snowman2,

I investigated the CI failure and found an issue with using version checks for this case.

The "Test PROJ and Cython Latest" workflow builds PROJ from the main branch (git clone https://github.com/OSGeo/PROJ.git), which results in a development version like 9.8.0dev. This version parses as >= 9.7.1, so PROJ_GTE_971 evaluates to True.

However, the EPSG:1671 name change ("RGF93 v1 to WGS 84 (1)" → "ETRS89-FRA [RGF93 v1] to WGS 84 (1)") comes from an EPSG database update, not the PROJ software version itself. The dev version from main has a high version number but may not have the updated EPSG database with this name change.

If my interpretation is correct, this means we can't reliably use PROJ version checks for EPSG database-dependent changes when testing against development versions like this.

Would it be acceptable to use an in check that accepts both names instead? Like in the initial commit?

assert co.name in (
    "RGF93 v1 to WGS 84 (1)",
    "ETRS89-FRA [RGF93 v1] to WGS 84 (1)",
)

[snowman2]

If my interpretation is correct, this means we can't reliably use PROJ version checks for EPSG database-dependent changes when testing against development versions like this.

The EPSG versions are linked to the PROJ version. So, you can reliably check the PROJ version for EPSG changes.

[snowman2]

https://github.com/OSGeo/PROJ/milestones

Looks like for this comparison, we need a PROJ_GTE_980

[phaarnes]

@snowman2
Thanks! I added PROJ_GTE_980 now, and the test CI tests passes again now.

[snowman2]

Thanks @phaarnes 👍