On Wed, 2 Dec 2020, Daniel Stenberg wrote: > To me, that's probably one of the bigger philosophical concerns about this > idea. curl uses libcurl, libcurl doesn't know about curl.
> But then nobody else than libcurl knows exactly what has been set for a > transfer so only libcurl itself can do a decent conversion... > I'm personally both for and against this idea! =) Think of it as a functionality present in many applications: "copy as curl" and also a great orthogonal feature to the "--libcurl" option. In software I maintain, we use libcurl for nearly all network communication and nearly every aspect of that is controlled by the user of the software. It's quite complicated. Now, if there's a problem curl is the number 1 test tool - if curl can speak to the server, the error is on our side (either in code or in front of the monitor) otherwise it's the server (it's never a curl bug ;) ). Right now we are struggling each time to construct a test command from the final set of options. Having libcurl generate it, would save a ton of time each time I have to debug stupid mistakes... -- Jakub
------------------------------------------------------------------- Unsubscribe: https://cool.haxx.se/list/listinfo/curl-library Etiquette: https://curl.se/mail/etiquette.html
