Hillshade geometry update fixes

[TimSylvester]

Fixes a problem where masking changes in hillshade tile geometry are not recognized as updated.

Fixes a bug where vertex attributes on a drawable replaced by ones created earlier are not applied to the underlying buffers. This would have been an issue previously, if the newly-applied attributes were not marked dirty, as with shared static tile geometry.

Fixes a bug where fill updates would apply the fill vertices to the triangulated outline drawables (if enabled), which would lead to errors if they were applied, but they weren't because of the previous item.

Resolves #2838 ?

[github-actions]

Bloaty Results (iOS) 🐋

Compared to main

    FILE SIZE        VM SIZE    
 --------------  -------------- 
  +0.0% +1.44Ki  [ = ]       0    TOTAL

Full report: https://maplibre-native.s3.eu-central-1.amazonaws.com/bloaty-results-ios/pr-2842-compared-to-main.txt

[github-actions]

Benchmark Results ⚡

Benchmark                                                     Time             CPU      Time Old      Time New       CPU Old       CPU New
------------------------------------------------------------------------------------------------------------------------------------------
OVERALL_GEOMEAN                                            +0.0033         +0.0034             0             0             0             0

Full report: https://maplibre-native.s3.eu-central-1.amazonaws.com/benchmark-results/pr-2842-compared-to-main.txt

[github-actions]

Bloaty Results 🐋

Compared to main

    FILE SIZE        VM SIZE    
 --------------  -------------- 
  +0.0% +57.2Ki  +0.1% +19.3Ki    TOTAL

Full report: https://maplibre-native.s3.eu-central-1.amazonaws.com/bloaty-results/pr-2842-compared-to-main.txt

Compared to d387090 (legacy)

    FILE SIZE        VM SIZE    
 --------------  -------------- 
   +27% +31.9Mi  +424% +25.3Mi    TOTAL

Full report: https://maplibre-native.s3.eu-central-1.amazonaws.com/bloaty-results/pr-2842-compared-to-legacy.txt

[louwers]

Better, but now some tiles don't get loaded at all until I change the zoom a bit:

[alasram]

@louwers the behavior in your screenshots is already present in main on Android

[JesseCrocker]

I have not looked at the code, but i tested this in my application, and it fixes the issue.

[TimSylvester]

Looks like we have a longstanding issue with updating the vertices without updating the segments. That's fine in Metal, but leads to problems in GL because the segments retain Vertex Array identifiers, so we get GL errors if we actually replace the vertices alone.

[louwers]

@TimSylvester Render tests are failing unfortunately.