allow passing headers to subgraphs

[mikaelkaron]

see platformatic/platformatic#1966

[mcollina]

Can you add a test?

[mikaelkaron]

Can you add a test?

Of course. That’s why it’s a WIP ;)

[mikaelkaron]

A few notes on the implementation:

1. Is it enough to just have a blacklist of headers to remove
2. Should there also be a "whitelist" of headers to keep
3. Is there a need to transform headers
4. Should there be a way to inject whitelist/blacklist/transformer from the outside

So I'm not sure about 1, I'm hoping that it's OK to just forward all headers (except a few) but I can already think about problems this may cause downstream.

The problem with 1 can be partially solved with 2, but it adds complexity that may not be needed at this stage.

I don't think we should transform headers here - if it's needed it should happen before we get here.

Providing some configuration would probably be a good idea - I'll try to add this without adding complexity. I'm hoping this can be simple enough to be serialized and provided in a platformatic configuration in the end.

[mcollina]

IMHO having a callback to transform the headers would solve all the configurability problems. However, keep this implementation as simple as possible, and we can add that later on if needed.

[mcollina]

Any help with the tests?

[mikaelkaron]

Sorry. I’ll try to get to it today.

[mcollina]

lgtm

[mikaelkaron]

Sorry @mcollina - just pushed a change renaming the test to align with other test names