Gateway API: fix for failing GatewayWithAttachedRoutes conformance test

[izturn]

Signed-off-by: Gang Liu gang.liu@daocloud.io

fix #5920

[codecov]

Codecov Report

Merging #5961 (1cb7d00) into main (44dd9f4) will decrease coverage by 0.02%.
Report is 5 commits behind head on main.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5961      +/-   ##
==========================================
- Coverage   78.69%   78.68%   -0.02%     
==========================================
  Files         138      138              
  Lines       19687    19675      -12     
==========================================
- Hits        15493    15481      -12     
  Misses       3890     3890              
  Partials      304      304              

Files	Coverage Δ
internal/dag/gatewayapi_processor.go	93.74% <100.00%> (+<0.01%)	⬆️
internal/gatewayapi/listeners.go	100.00% <100.00%> (ø)

[sunjayBhatia]

just a requested change on the release note so far but from a quick look at the test changes/code changes this looks like it achieves what looks like the right outcome for the most part, though ends up being a little repetitive

With this change, we're processing Listeners/Routes multiple times now, once looping through the Listeners for finding which ones are valid, then going through each Listener/Route another time for seeing if the Routes can attach/intersect hostnames, then doing the Listener/Route analysis again for the purposes of calculating attached route counts. It would be good to coalesce some of this logic together since it seems like extra processing we could reduce.

Maybe this looks like not returning early from computeListener and including the error of why the listener is invalid/if it should still be allowed to have attached routes in the returned listenerInfo struct

This isn't something the upstream conformance tests are checking yet, but if this doesn't exist would be good to ensure a test exists for a Gateway Listener that has an invalid configuration in its route selection (e.g. invalid namespace selector due to a field missing or invalid route kind) but a Route with a parent ref that would be valid should still not count

[izturn]

ping @sunjayBhatia

[skriss]

Couple nits but otherwise looks pretty good to me (I'm heavily relying on the tests here for correctness!)

[skriss]

LGTM, will leave for @sunjayBhatia