Jean-Christophe Amiel

Hi @danielbprice I agree, we need to revisit the multipart form data to support all possible use cases (particularly we should be able. to do what can be done with...

@theoforger I've the impression that the right option to disable reusing connection is not [CURLOPT_TCP_KEEPALIVE](https://curl.se/libcurl/c/CURLOPT_TCP_KEEPALIVE.html), but something like [CURLOPT_FRESH_CONNECT](https://curl.se/libcurl/c/CURLOPT_FRESH_CONNECT.html) or [CURLOPT_FORBID_REUSE](https://curl.se/libcurl/c/CURLOPT_FORBID_REUSE.html). I've to dig to understand the difference but `CURLOPT_TCP_KEEPALIVE`...

It think it may also depends on the HTTP version protocol. I don't think there is any reuse with a server responding HTTP/1.0, you should try with a "known" host...

Don't think there is a need to `keepalive` / `no-keepalive`. We can begin to implement `CURLOPT_FRESH_CONNECT` (`--fresh-connect`) and that should be sufficient to expose an option to be able to...

@fabricereix one thing to note it that there is no curl cli equivalent of `CURLOPT_FRESH_CONNECT`, it's just a libcurl option.

@theoforger Of course, let us just check something that will be merged without problem, so we don't waste your time this time!

Hi @shawnallen85, No advancement for this, and not a priority. It seems to me that these parametized workflows can be done with a proper script without any Hurl syntax addition...

Hi @robjtede The `arbitrary_precision` flag has been introduced in f3909c21cfe8a28750fde4138c17ce09d9b65bb4: the main issue was to address JSON number precision in JSONPath for instance: Given a JSON response body like this:...

@RapidPencil I understand your point but I'm really really reluctant to do this. I'm not at all happy with having a different Hurl's output for a given input. Our priority...

@robjtede @RapidPencil we have looked around this issue (we still want to have a minimal reproductible sample if you have something like that it would be cool). If we want...