Should we allow lower-case header names in io.pedestal.service.test/test-servlet-request ?

[daemianmack]

Re: https://github.com/pedestal/pedestal/blob/master/service/src/io/pedestal/service/test.clj#L96-97

I spent a bit of time trying to pass a "content-type" header to io.pedestal.service.test/response-for before realizing I needed to pass a "Content-Type" header instead.

Do we suppose this is a sufficiently-common error that it's worth searching for arbitrarily-cased headers in the test-servlet-request fn?

If so, I'll submit a PR.

[avescodes]

Hmm, so we'd be lenient consider things like "content-type" or :content-type or :Content-Type. I wonder how far sweeping this leniency would have to be given we normally deal in maps?

[timewald]

The goal of the faux servlet test infrastructure is to ensure that your
service code works with the data it would get from a servlet, so the goal
should be to emulate that data. That said, the servlet-interceptor
lower-cases header names when it builds the headers map for the request
map. So this seems like a reasonable change in response-for. The main risk
to this change is that if, at some point, servlet-interceptor did something
more with headers, some tests that used lower-case headers might break. I
would prefer we use correct casing for HTTP headers, as per the spec, but
if this is a real pain point, I'm not against making the test jig a little
more lenient.

Tim-

On Fri, May 31, 2013 at 11:41 AM, Ryan Neufeld notifications@github.comwrote:

Hmm, so we'd be lenient consider things like "content-type" or
:content-type or :Content-Type. I wonder how far sweeping this leniency
would have to be given we normally deal in maps?

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/79#issuecomment-18752926
.

[avescodes]

It sounds like the answer was a round-about "no."