Le 12 janv. 2014 à 22:20, Rich Gray <[email protected]> a écrit :

> Although it is probably the closest code for this error, CURLE_UPLOAD_FAILED 
> seems wrong.  In my mind, it implies a wire error,

Just a precision: the corresponding error string is "Upload failed (at 
start/before it took off)", which I found appropriate here at first.

> whereas it would really be a library parameter/setup  error.

If we go ahead with the ability to report a misuse (which I would like to 
discuss - see my last post), it could indeed be even better to use a very 
explicit error code.

> if the road to a new return code is taken, perhaps add more than one:
> 
> CURLE_MISSING_PARAMETER
> CURLE_INVALID_PARAMETER
> CURLE_INCONSISTENT_PARAMETER

It makes sense. But to be honest I would prefer introducing a single one to 
first address the (minor) issue I reported here. And I'm not so sure if the 
better solution is to report the misuse and fail early (which introduces a bit 
of complexity - e.g for auth requests with multiple phases), or do as HTTP PUT 
does so far, i.e simply not send any Content-Length header if the post size is 
unknown (again, see my last post).


-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette:  http://curl.haxx.se/mail/etiquette.html

Reply via email to