aboutsummaryrefslogtreecommitdiff
path: root/lib
AgeCommit message (Collapse)Author
2004-06-29First attempt at making the multi interface work when connecting to a hostDaniel Stenberg
that resolves to multiple IP addresses.
2004-06-24oops 5 bytes makes 4 letters plus zero byteDaniel Stenberg
2004-06-24fix warningDaniel Stenberg
2004-06-24the _num_chars() function is not used, removingDaniel Stenberg
2004-06-24to prevent compier warnings, we only declare len if we have codeDaniel Stenberg
that uses it
2004-06-24Gisle cleaned up remaining host resolve re-org issuesDaniel Stenberg
2004-06-24include header for our printfsDaniel Stenberg
2004-06-24reverted bad win32 fixDaniel Stenberg
2004-06-24fixed the MSVC build.Gunter Knauf
2004-06-24length limit the sscanf() parsing to prevent buffer overflowDaniel Stenberg
2004-06-24ah, simplified my latest change moreDaniel Stenberg
2004-06-24fix for systems without gethostbyname_r()Daniel Stenberg
2004-06-24Replaced all uses of sprintf() with the safer snprintf(). It is just aDaniel Stenberg
precaution to prevent mistakes to lead to buffer overflows.
2004-06-24made the Curl_he2ai() take the port number as an int intead, to avoid lotsDaniel Stenberg
of typecasts all over
2004-06-24use snprintf() to be on the safe sideDaniel Stenberg
2004-06-24typecasts to prevent warningsDaniel Stenberg
2004-06-24only use sockaddr_in6 on ipv6-enabled hostsDaniel Stenberg
2004-06-24typecast to prevent warningDaniel Stenberg
2004-06-24use Curl_addrinfo, not 'struct addrinfo'Daniel Stenberg
2004-06-24fixed problems I missed to fix from my cleanupDaniel Stenberg
2004-06-24prevent warningDaniel Stenberg
2004-06-24Source cleanups. The major one being that we now _always_ use a Curl_addrinfoDaniel Stenberg
linked list for name resolved data, even on hosts/systems with only IPv4 stacks as this simplifies a lot of code.
2004-06-23static functions are better not Curl_ prefixed to make their staticDaniel Stenberg
status more obvious
2004-06-23When adding the return -1 to prevent warnings on some compilers, othersDaniel Stenberg
started complaining since it won't be reached... So I removed the call to abort() and just return -1 instead. abort() was wrong to call anyway since this is a library!
2004-06-23prevent a warningDaniel Stenberg
2004-06-22one copy&paste too much, removed the define again.Gunter Knauf
2004-06-22argh - copy&paste error.Gunter Knauf
2004-06-22syncronized with recent getdate.y updates.Gunter Knauf
2004-06-22the hostname variable wasn't assigned and we no longer use itDaniel Stenberg
2004-06-22David Cohen pointed out that RFC2109 says clients should allow cookies toDaniel Stenberg
contain least 4096 bytes while libcurl only allowed 2047. I raised the limit to 4999 now and made the used buffer get malloc()ed instead of simply allocated on stack as before.
2004-06-22Günter Knauf fixed getdate.y to remove a few warnings. I removed theDaniel Stenberg
ifdef'ed test we never ever use anyway.
2004-06-22Moved the "About to connect() to" text to the place where the host name isDaniel Stenberg
actually known, as before this text lied when used in i.e FTP.
2004-06-22Gisle fixed the wildcard checks for certificates.Daniel Stenberg
2004-06-22Gunter's fix to avoid the notorious YYSTACK_USE_ALLOCA warning we get onDaniel Stenberg
several platforms/compilers/yacc versions.
2004-06-22pass an int pointer when it expects an int pointer...Daniel Stenberg
2004-06-21The read callback can now return CURL_READFUNC_ABORT to stop a transfer.Daniel Stenberg
2004-06-21typecasts to prevent compiler warningsDaniel Stenberg
2004-06-19When doing auth negotiations or authprobing, we only consider HTTP codeDaniel Stenberg
<300 to be good.
2004-06-19prevent compiler warningDaniel Stenberg
2004-06-18Gisle's "SSL patch" from June 16th 2004, modified by me as discussed on theDaniel Stenberg
mailing list.
2004-06-18With David Byron's test server I could repeat his problem and make sure thatDaniel Stenberg
POSTing over HTTPS:// with NTLM works fine now. There was a general problem with multi-pass authentication with non-GET operations with CONNECT.
2004-06-16Keep the upload byte counter in an curl_off_t, not an int. 32bits is notDaniel Stenberg
enough. This is most likely the bug Jean-Louis Lemaire reported that makes 2GB FTP uploads to report error when completed. Also padded comments to get them aligned again, only for visibility.
2004-06-16Alexander Krasnostavsky fixed a flaw in the 3rd party transfer code thatDaniel Stenberg
didn't properly check return code.
2004-06-15Fix the auth code to enable us to i.e set DIGEST and then find out that theDaniel Stenberg
server doesn't require any auth at all and then we just continue nicely. We now have an extra bit in the connection struct named 'authprobe' that is TRUE when doing pure "HTTP authentication probing".
2004-06-14we actually build and run fine with libidn 0.4.1 too, so let's not requireDaniel Stenberg
anything newer than that
2004-06-14Allow formposting of files larger than what fits in memory by not reading theDaniel Stenberg
file until it is actually being uploaded. Make sure we build and still work with HTTP disabled - the SSL code might use the boundary string for some random seeding.
2004-06-13moved default: in a switch case to prevent compiler warning that 'request'Daniel Stenberg
might be used uninitialized
2004-06-13provide curl_formfree() even when http is disabled, it does nothing thenDaniel Stenberg
2004-06-13use Curl_strcasestr() when checking wildcard cert namesDaniel Stenberg
2004-06-13added Curl_strcasestr() for case insensitive strstr() searchingDaniel Stenberg