recent fluff

The initial HTTP request can now be sent in multiple parts, as part of the
regular transfer process. This required some new tweaks, like for example
we need to be able to tell the tranfer loo

The initial HTTP request can now be sent in multiple parts, as part of the
regular transfer process. This required some new tweaks, like for example
we need to be able to tell the tranfer loop to not chunky-encode uploads
while we're transferring the rest of the request...

show more ...

send_buffer is no more here

added test56, nearly 100KB big!

Test case for sending insanely big HTTP requests. Mainly done this way to
make sure that it isn't all sent off in one single send() but instead
really tests the multiple-part-send logic.

more logging, now logs the full response too, basic support for dealing
with chunked transfer-encoding uploads added

A normal POST now provides data to the main transfer loop via the usual
read callback, and thus won't put a lot of stress on the request sending
code (which currently does an ugly loop).

The fread() callback pointer and associated pointer is now stored in the
connectdata struct instead, and is no longer modified within the 'set' struct
as previously (which was a really BAAAD

The fread() callback pointer and associated pointer is now stored in the
connectdata struct instead, and is no longer modified within the 'set' struct
as previously (which was a really BAAAD thing).

show more ...

Added a default headers section and also made some minor details more
up-to-date with recent changes.

better errno include and no extern

read and write as much as possible until end of data or EWOULDBLOCK before
returning back to the select() loop. Consider this a test so far.

deal with spaces in name and value tags a lot better!

changed proto for Curl_krb_kauth()

Solaris needs errno as an extern int.

make WIN32 defined for Borland properly, as told by Alexander J. Oss

called SSLCERTS now

7.10.3-pre2

The waiting for the 226 or 250 line expected to come after a transfer is
complete is now only made for 60 seconds and if no data was received during
those 60 seconds, we store a special error

The waiting for the 226 or 250 line expected to come after a transfer is
complete is now only made for 60 seconds and if no data was received during
those 60 seconds, we store a special error message (preparing to make this
a special error code) as this most likely means that the control connection
has died while we were transferring data.

show more ...

missing }

clarified

Jeff pointed out this flaw in the example

-@ is no longer an official shortcut for --create-dirs

don't officially use -@ for --create-dirs, only use the long form

clarify the DEBUGFUNCTION data not being zero terminated

Curl_GetFTPResponse() takes a different set of parameters and now return a
proper CURLcode. The default timeout for reading one response is now also
possible to change while running.