Implement header line folding (fixes #37, #68) #107
Conversation
nox
force-pushed
the
line-folding
branch
3 times, most recently
from
73bfcd9
to
dd936b8
Compare
| assert_eq!(response.reason.unwrap(), "OK"); | ||
| assert_eq!(response.headers.len(), 1); | ||
| assert_eq!(response.headers[0].name, "Line-Folded-Header"); | ||
| assert_eq!(response.headers[0].value, &b"hello \r\n \r\n there"[..]); |
There was a problem hiding this comment.
I see, this was what I was wondering: how exactly does it get exposed. Am I right in assuming that this implies it will "simply" capture the text from the first line until the end of all obs-folded lines? So, if this were hello: there\r\n mister\r\n baguette\r\n, it would scoop it all as b"there\r\n mister\r\n baguette"?
That seems elegant. Just to consider what I'd thought of, I assumed this feature would instead use up an entry in the &mut [Header] slice for each line. If the line started with obs-folding, then the parse would just grab the name from the previous entry. So it'd look something like this:
assert_eq!(resp.headers[0].name, "hello");
assert_eq!(resp.headers[0].value, &b"there");
assert_eq!(resp.headers[1].name, "hello");
assert_eq!(resp.headers[1].value, &b"mister");
assert_eq!(resp.headers[2].name, "hello");
assert_eq!(resp.headers[2].value, &b"baguette");Whatever we end up using, explaining that is what it does would be useful as a sentence/paragraph in the commit description.
There was a problem hiding this comment.
Line folds are equivalent to spaces, but multiple entries of a single header are equivalent to a single entry combined with commas, so we can't return multiple entries.
There was a problem hiding this comment.
Oh right.... gross...
Welp, then I think this is the best we can do. I think it'd be good to better explain, perhaps with an example, in the documentation for that builder method.
There was a problem hiding this comment.
I added an example to the builder method.
No description provided.