aboutsummaryrefslogtreecommitdiff
path: root/lib
AgeCommit message (Collapse)Author
2003-06-10Daniel Kouril's patch that adds HTTP negotiation support to libcurl wasDaniel Stenberg
added.
2003-06-06When doing very big GET requests over HTTPS, we need to add some extraDaniel Stenberg
funky logic in order to make re-tries work fine with OpenSSL. This corrects the problem David Orrell noticed.
2003-06-06Just moved around some logic in Curl_write() to make it easier to debug.Daniel Stenberg
2003-06-02Make the Content-Length info override the Connection: close header, so thatDaniel Stenberg
libcurl will stop reading when the number of bytes have arrived and not wait for a closed socket.
2003-06-02work-around SSL implementation flaws better, pointed out in bug reportDaniel Stenberg
#745122.
2003-06-02make a more descriptive error message when CURLE_HTTP_RETURNED_ERROR isDaniel Stenberg
returned
2003-05-28Posting static data using POST and chunked encoded now also appends theDaniel Stenberg
data to the initial request buffer, if the total post data is less than 100K.
2003-05-27Rudy Koento experienced problems with curl's recent habit of POSTing data inDaniel Stenberg
two separate send() calls, first the headers and then the data. I've now made a fix that for static and known content that isn't to be chunked-encoded, everything is now sent in one single system call again. This is also better for network performance reasons.
2003-05-27Another socks5-fix. Make sure that when we use a socks-proxy, it is not theDaniel Stenberg
same as using a httpproxy so we must make sure to better check for http proxies before we do HTTP proxy stuff. This included authorization and URI usage in the request etc.
2003-05-23Ricardo Cadime found a socket leak when listing directories withoutDaniel Stenberg
contents. Test cases 144 and 145 were added to verify the fix. Now we deal with return code 450 properly and other codes also do proper cleanup.
2003-05-23Rudy Koento's problem fixed, test case 66 verifies this.Daniel Stenberg
2003-05-23include digest.h for protoDaniel Stenberg
2003-05-23DJGPP fix by Gisle VanemDaniel Stenberg
2003-05-22warning-free is betterDaniel Stenberg
2003-05-22Better Digest stuffDaniel Stenberg
2003-05-22Added CURLOPT_HTTPDIGEST supportDaniel Stenberg
SOCKS5 fix as suggested by Jis in bugreport #741841.
2003-05-22proper header addedDaniel Stenberg
2003-05-22hush the compilerDaniel Stenberg
2003-05-22Initial Digest support. At least partly working.Daniel Stenberg
2003-05-21David Balazic pointed out the lack of checks for a valid %XX code whenDaniel Stenberg
we unescape a string. We now check and decode only valid %XX strings.
2003-05-21 Gisle Vanem made curl build with djgpp on DOS.Daniel Stenberg
2003-05-20Gisle Vanem's code for not trusting h_aliases to always be non-NULLDaniel Stenberg
2003-05-15Skip any preceeding dots from the domain name of cookies when we keep themDaniel Stenberg
in memory, only add it when we save the cookie. This makes all tailmatching and domain string matching internally a lot easier. This was also the reason for a remaining bug I introduced in my overhaul.
2003-05-14George Comninos provided a fix that calls the progress meter when waitingDaniel Stenberg
for FTP command responses take >1 second.
2003-05-13Setup and use CURL_INADDR_NONE all over instead of INADDR_NONE. We setupDaniel Stenberg
the define accordingly in the hostip.h header to work nicely all over.
2003-05-13before using if2ip(), check if the address is an ip address and skip it ifDaniel Stenberg
it is.
2003-05-12before checking for network interfaces using if2ip(), check that the givenDaniel Stenberg
name isn't an ip address
2003-05-12 Dan Fandrich changed CURLOPT_ENCODING to select all supported encodings ifDaniel Stenberg
set to "". This frees the application from having to know which encodings the library supports.
2003-05-12when we have accepted the server's connection in a PORT sequence, we setDaniel Stenberg
the new socket to non-blocking
2003-05-12avoid the write loopDaniel Stenberg
2003-05-12incoming proxy headers shall be sent to the debug function has HEADERs notDaniel Stenberg
DATA
2003-05-09support for CURLOPT_FTP_USE_EPRT addedDaniel Stenberg
2003-05-06 Kevin Delafield reported another case where we didn't correctly check forDaniel Stenberg
EAGAIN but only EWOULDBLOCK, which caused badness on HPUX. We also check for and act the same on EINTR errors as well now.
2003-05-02If there is a custom Host: header specified, we use that host name toDaniel Stenberg
extract the correct set of cookies to send. This functionality is verified by test case 62.
2003-05-01corrected a comment about gzip not being supportedDaniel Stenberg
2003-05-01Andy Cedilnik fixed some compiler warningsDaniel Stenberg
2003-05-01ourerrno became Curl_ourerrno() and is now available to all libcurlDaniel Stenberg
2003-05-01Use the proper Curl_ourerrno() function instead of plain errno, for betterDaniel Stenberg
portability. Also use Andy Cedilnik's compiler warning fixes.
2003-04-30when making up the list of path parts, save the last entry pointing to NULLDaniel Stenberg
as otherwise we'll go nuts
2003-04-30if there's a cookiehost allocated, free that tooDaniel Stenberg
2003-04-30stop parsing Host: host names at colons tooDaniel Stenberg
2003-04-30modified to the new cookie function protoDaniel Stenberg
2003-04-30extract host name from custom Host: headers to use for cookiesDaniel Stenberg
2003-04-30Many cookie fixes:Daniel Stenberg
o Save domains in jars like Mozilla does. It means all domains set in Set-Cookie: headers are dot-prefixed. o Save and use the 'tailmatch' field in the Mozilla/Netscape cookie jars (the second column). o Reject cookies using illegal domains in the Set-Cookie: line. Concerns both domains with too few dots or domains that are outside the currently operating server host's domain. o Set the path part by default to the one used in the request, if none was set in the Set-Cookie line.
2003-04-30changes need for the new ftp path treatment and the new cookie codeDaniel Stenberg
2003-04-30David Balazic's patch to make the FTP operations "do right" according toDaniel Stenberg
RFC1738, which means it'll use one CWD for each pathpart.
2003-04-29 Peter Kovacs provided a patch that makes the CURLINFO_CONNECT_TIME work fineDaniel Stenberg
when using the multi interface (too).
2003-04-25RSAglue.lib is no longer needed with recent OpenSSL versionsDaniel Stenberg
2003-04-22Dan Fandrich corrected the error messages on "bad encoding".Daniel Stenberg
2003-04-22Dan Fandrich's gzip bugfixDaniel Stenberg