Add gateway conformance test workflow

[hannahhoward]

Since the official test of trustless gateway lives in ipfs/gateway-conformance, adding the test here.

Still working on this, especially cause we're not currently passing.

[codecov-commenter]

Codecov Report

Merging #452 (2c7fb51) into main (2cf1121) will increase coverage by 0.11%.
The diff coverage is n/a.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #452      +/-   ##
==========================================
+ Coverage   75.73%   75.85%   +0.11%     
==========================================
  Files          87       87              
  Lines        6677     6677              
==========================================
+ Hits         5057     5065       +8     
+ Misses       1335     1332       -3     
+ Partials      285      280       -5     

see 7 files with indirect coverage changes

[rvagg]

If you end up changing the way that trustless-utils handles ranges, you may find that the e2e test in here fails; because frisbii is using the same logic to serve that data, so it's a matter of making the changes flow through the set of things, frisbii first and then lassie. i.e. e2e failing here isn't necessarily a problem and should be easily fixable if we agree that the change is correct and should be bubbled up through the stack.

[parkan]

this is a worthwhile effort but given the time passed + divergence on all fronts + concerns about whether this is the right interface at which to evaluate conformance I think we should close; I don't have enough context to make the call though, so perhaps @rvagg you could chime in?

[rvagg]

Would be good to do, the conformance tests track the spec properly now and will pick up issues if you're not tracking the spec - I don't track it actively and they still make chances so it's one way of being aware of those changes. @lidel showed up and added it to frisbii and it was much easier than I expected it would be and it's been helpful there: ipld/frisbii#214, it caught ipld/go-trustless-utils#62 which I'd have no reason to know was a spec change without those tests.

[parkan]

closing and re-implementing