Age | Commit message (Collapse) | Author |
|
describes, Bob Schader's research seems to agree.
|
|
|
|
|
|
the request isn't GET. Presumably, this is how it should be made.
|
|
taking place
|
|
|
|
for what HTTP request that is being used. The old bit-style fields are still
in there as well.
|
|
GET resume. Other resumes are upload-wise and don't care about this header
in the download stream
|
|
|
|
|
|
|
|
|
|
|
|
unfolded telwrite() instead of being a separate single function
|
|
|
|
Curl_write() and Curl_read() are here
|
|
|
|
Some functions changed prototype due to this change as well.
|
|
socket. Curl_ConnectHTTPProxyTunnel changed prototype.
|
|
|
|
|
|
|
|
|
|
added lots of explaining comments
|
|
|
|
curl will return an error when such an address is used.
|
|
|
|
|
|
|
|
and download.h
|
|
|
|
|
|
|
|
|
|
set, as those systems are likely to have it already set in a system header
and this prototype has proven to cause problems on SCO systems.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
firewalls (like FW-1) seems to dislike split-up writes at times...
|
|
|
|
|
|
|
|
|
|
on >= 400 errors
|