Daniel Stenberg [Tue, 8 Jan 2002 23:19:32 +0000 (23:19 +0000)]
this seems to correct the SSL reading problem introduced when switching
over to non-blocking sockets, but this loops very nastily. We should return
back to the select() and wait there until more data arrives, not just blindly
attempt again and again...
Sterling Hughes [Tue, 8 Jan 2002 04:30:59 +0000 (04:30 +0000)]
1) the dns_cache_timeout should be an integer, not a bool
2) in the curl_dns_cache_entry structure, timestamp should be
a time_t instead of an integer (although I doubt it matters).
Daniel Stenberg [Mon, 7 Jan 2002 23:05:36 +0000 (23:05 +0000)]
As identified in bug report #495290, the last "name=value" pair in a
Set-Cookie: line was ignored if they didn't end with a trailing
semicolon. This is indeed wrong syntax, but there are high-profile web sites
out there sending cookies like that so we must make a best-effort to parse
them.
Sterling Hughes [Mon, 7 Jan 2002 20:52:32 +0000 (20:52 +0000)]
Make cach'ing work with threads now, there are now three cases:
- Use a global dns cache (via setting the tentatively named,
CURLOPT_DNS_USE_GLOBAL_CACHE option to true)
- Use a per-handle dns cache, by default
- Use a pooled dns cache when in the "multi" interface
Daniel Stenberg [Fri, 4 Jan 2002 09:38:52 +0000 (09:38 +0000)]
an unconditional occurance of inet_ntoa() now uses inet_ntoa_r() on all
platforms that have such a function.
This affects multi-thread running libcurls on IPv4 systems that have VERBOSE
switched on. The previous version was risking that another thread overwrote
the data before it was read out in this thread. There could possibly also
be a slight risk that the data isn't zero terminated for a short while and
thus could cause the thread to crash...
Daniel Stenberg [Fri, 4 Jan 2002 09:17:52 +0000 (09:17 +0000)]
The buffer in ftp_pasv_verbose(), used for gethostbyaddr_r(), is now defined
to become properly 8-byte aligned on 64-bit archs. Philip Gladstone reported.
Daniel Stenberg [Thu, 3 Jan 2002 08:07:29 +0000 (08:07 +0000)]
Changed how -I/--head works when --include is also used... Test case 104
stopped working after the dec-20 fixes that now supports FTP operations to
skip the transfer phase.
Daniel Stenberg [Thu, 20 Dec 2001 11:22:01 +0000 (11:22 +0000)]
If nobody is set we won't download any FTP file. If include_header is set,
we return a set of headers not more. This enables FTP operations that don't
transfer any data, only perform FTP commands.
Daniel Stenberg [Tue, 11 Dec 2001 00:48:55 +0000 (00:48 +0000)]
when the file name given to -T is used to build an upload path, the local
directory part is now stripped off and only the actual file name part will be
used
Daniel Stenberg [Wed, 5 Dec 2001 06:47:01 +0000 (06:47 +0000)]
Jon Travis suggested fix. when CURLOPT_HTTPGET is used we must assign
set.upload to FALSE or else we might still get an upload if the previous
operation was an upload!
Daniel Stenberg [Mon, 3 Dec 2001 13:48:59 +0000 (13:48 +0000)]
As Eric Lavigne pointed out, the ftp response reader MUST cache data that
is not dealt with when we find an end-of-response line, as there might be
important stuff even after the correct line. So on subsequent invokes, the
cached data must be used!
Daniel Stenberg [Mon, 3 Dec 2001 13:46:56 +0000 (13:46 +0000)]
test case 126 added, this uses RETRWEIRDO that makes the FTP server send two
responses at once, to excerise the part of curl to make sure it can cache
(parts of) responses properly.