Stephen M. Coakley

Results 115 comments of Stephen M. Coakley

Removing this from the 1.0 milestone, which is cluttered as it is. Persisting or allowing serialization of the cookie jar means we need to pick a particular persistence format and...

Technically you don't _need_ this in order to send multipart forms since a multipart form is just bytes. You can craft the request body yourself or use one of the...

Thanks for working on this! Just a few suggestions on the shape of the API. I think we need to make this a bit more flexible to support the more...

Support for HTTP/0.9 is disabled by default in the underlying libcurl library as of version 7.66.0, so unless you disable the `static-curl` feature and link to an old version of...

Thanks for filing an issue! I'm not actually sure if this is being done by Isahc intentionally or not but is worth digging into. At first guess I suppose maybe...

Proposed API for controlling connection caching per client: ```rust impl HttpClientBuilder { /// Set the size of the connection cache. Setting to `None` disables connection /// caching. fn connection_cache_size(mut self,...

Maybe we could set both `CURLOPT_FRESH_CONNECT` and `CURLOPT_FORBID_REUSE` in one call for the per-request API? Effectively the meaning is that that one request does not partake in the connection cache...

> In the case of HTTP/2, I wonder if `CURLOPT_FORBID_REUSE` also prevents other calls from multiplexing requests into an active benchmarking connection. Looking at the source, it appears the answer...

When inside a redirect it appears that `CURLOPT_FRESH_CONNECT` is completely ignored intentionally: https://github.com/curl/curl/blob/e062043433381fa5fd2f90b2fcc9dc912dbb79f6/lib/url.c#L3577-L3580. After every followed redirect, the existing easy handle is essentially reset to point to the new URL,...