matryer
commented
9 years ago Very cool - I wonder if it’s worth benchmarking the two solutions so we can understand the difference under load?
On 30 Sep 2015, at 13:10, Rene Zbinden notifications@github.com wrote:
Hi Mat
I watched your London gopher talk and became aware of the response package. Great talk BTW.
I like the respond package, but what I didn't like, was the usage of locks in the options handler.
I rewrote the package to use https://github.com/nbio/httpcontext https://github.com/nbio/httpcontext for contexts. It stores the request context directly in the http.Request by mutating the request.Body, avoiding the use of a global mutex and per-request teardown.
And it passes your tests :-)
You can view, comment on, or merge this pull request online at:
https://github.com/matryer/respond/pull/3 https://github.com/matryer/respond/pull/3 Commit Summary
rewrite to use nbio/httpcontext instead of map and locks File Changes
M respond.go https://github.com/matryer/respond/pull/3/files#diff-0 (53) Patch Links:
https://github.com/matryer/respond/pull/3.patch https://github.com/matryer/respond/pull/3.patch https://github.com/matryer/respond/pull/3.diff https://github.com/matryer/respond/pull/3.diff — Reply to this email directly or view it on GitHub https://github.com/matryer/respond/pull/3.
zbindenren
Hi Mat
I watched your London gopher talk and became aware of the response package. Great talk BTW.
I like the respond package, but what I didn't like, was the usage of locks in the options handler.
I rewrote the package to use https://github.com/nbio/httpcontext for contexts. It stores the request context directly in the http.Request by mutating the request.Body, avoiding the use of a global mutex and per-request teardown.
And it passes your tests :-)