Richard Marshall

Got a start on looking through the implementation this morning and will take a look at the updates as well when I can. In the meantime I noticed a few...

While doing some more looking at the results of formatting various real world VCL files I notice comments getting lost in a few places. Given how critical it is for...

I hope to have some free time this weekend to get back to reviewing this.

Definitely a lot to look through, will start reviewing the changes in the next day or two when I have a bigger block of free time.

Put together this fiddle / falco test file to evaluate as many not set conditions I could come up with in a reasonable amount of time. https://fiddle.fastly.dev/fiddle/3f8ea76a Running these tests...

The `LenientString`/`VString` types are handling two use-cases; one for handling the `(null)` expansion, and the other to store a collection of values for the header object representation (`req.http.foo:bar`). ## Multiple...

> @richardmarshall [richardmarshall@3212a48](https://github.com/richardmarshall/falco/commit/3212a486de66f7426a0e516b2ae39da1ed17dae3) is a great approach to consider the (null) string! could you send me a PR for this branch of `feat/re-implementation-for-header-value` please? Sure will do, but probably won't...

I got started on rebasing my experimental changes onto this branch when I got derailed by discovering what seems to be a bug in Fastly's not_set handling for header fields....

Ended up running into another issue that opened up a whole new problem area for our compatibility with Fastly's header field implementation. Currently in our implementation setting the main header...

> Oh really, I also did not know such behavior in dealing with comma-separated header values. Yeah it was quite a surprise when I first saw it. Further experimentation found...