improve `fetch` timeout docs

[shriram]

Running

bad-fetch = F.fetch("https://raw.githubusercontent.com/NO-SUCH-URL")
bad-fetch satisfies Ei.is-right

works, but GitHub returns a 400 instead of 404. A URL like

fetch("https://www.pyret.org/NO-SUCH-URL")

ought to help, but it was breaking. Now it works (email: "bad fetching"), so improve the docs w/ this, e.g., look for a 404 in the string.

[jpolitz]

This should work:

include fetch
import either as E

check:
    # See https://httpstat.us/
    result = fetch("https://httpstat.us/200?sleep=30000")
    result satisfies E.is-right
    to-repr(result) satisfies string-contains(_, "Abort")
end

[shriram]

@jpolitz I think the issue is a bit different. This works:

string-contains(to-repr(F.fetch("https://raw.githubusercontent.com/NO-SUCH-URL")), "40")

is true, but

string-contains(to-repr(F.fetch("https://www.pyret.org/NO-SUCH-URL")), "40")
string-contains(to-repr(F.fetch("https://cs.brown.edu/NO-SUCH-URL")), "40")

both return false. It would be nice if the server's return code were part of the error, so that the user can match against it, debug problems, etc. So I think this isn't actually a documentation issue as much as an implementation one.