the recent libcurl changes.

when SIZE is used on the file while being in BINARY mode. To work
around that (stupid) behavior, we attempt to parse the RETR response
even if the SIZE returned size zero.

Debugging help from Salvatore Sorrentino on February 26, 2003.

function, as this could easily end up looping for a very long time (more or
less until the whole transfer was done) and no library-using app would want
that.

Found thanks to a report by Kyle Sallee.

header is used, we must wait for a 100-code (or timeout), before we send the
data. The timeout is merely 1000 ms at this point. We may have reason to set
a longer timeout in the future.

than 5 could cause a segfault.

even better (also based on input from Martin).

AIX 3.2.5 and possibly other OSF-like system headers.

data to encode.

good one. This should work fine again.

longer use the SSL functions that store the error message in a static buffer
since that is not very multi-thread friendly.

compilers

data on uploads.

ftp_done() with the previous transfer's value, as Dave Halbakken found out.
He also verified this fixed corrected the problem.

way for emacs, and vim users should provide a similar non-polluting style

removed emacs local-variables stuff

calls!

and this has been verified to correct the problem.

to enable better error-tracking

only allocating the scratch memory buffer once it is needed and not always
in the handle.

was due to a missing fix for the password prompting

"username" will cause libcurl to prompt for password. Bryan Kemp noticed.

test case 136 is added for this