aboutsummaryrefslogtreecommitdiff
path: root/lib
AgeCommit message (Collapse)Author
2004-02-13Ben Greear's SO_BINDTODEVICE patch that binds to a network interface "evenDaniel Stenberg
more" when the previous approach. Known to work on Linux, possibly on other platforms as well.
2004-02-13Greg Hewgill found out 'contentlength' wasn't big enough to hold a large file!Daniel Stenberg
2004-02-13I made the same fix here, that Tor already did in the ftp.c code. To makeDaniel Stenberg
sure this doesn't get weird on 64bit archs.
2004-02-13Tor Arntsen's fix for the bad (64bit wise) typecast when using gmtime()Daniel Stenberg
2004-02-12Make hostcache_fixoffset() take a long for offset, to fully work with 64bitDaniel Stenberg
archs, also no longer typecast pointers to ints as that is a nono on 64bit systems.
2004-02-12removed the subchar variable, it was only set and never usedDaniel Stenberg
2004-02-12removed the ldaptext variable, it was only set and never usedDaniel Stenberg
2004-02-12removed the nth variable, it was only set and never used anywayDaniel Stenberg
2004-02-12No longer receive the return code in ConnectionKillOne() that wasn't dealtDaniel Stenberg
with anyway and thus caused picky compiler to warn.
2004-02-11#if-check for SIGALRM before assuming it is presentDaniel Stenberg
2004-02-09Modified the default HTTP Accept: header to only be Accept: */*Daniel Stenberg
2004-02-09Oops. I broke the flow with the previous commit.Daniel Stenberg
2004-02-09some annoying compilers warn about "(void)foo;" lines so we avoid themDaniel Stenberg
2004-02-09Tor Arntsen's patch for working around a notorious bug in the AIX5Daniel Stenberg
getaddrinfo() implementation.
2004-02-06A custom Host: header is only considered if the request is not made byDaniel Stenberg
following a location. After discussions with Tim Baker.
2004-02-06The MIPSPro compiler complains on constructs such as "(void)foo;" soDaniel Stenberg
we avoid it where possible.
2004-02-06Make sure DynaGetFunction() returns a function pointer, not a data pointer.Daniel Stenberg
The standards don't actually allow typecasts between data and functions so some picky compilers warn about this.
2004-02-05Andrés García's updatesDaniel Stenberg
2004-02-05if an empty 'transfer-encoding:' header is provided, we switch off theDaniel Stenberg
chunky coding of uploads
2004-02-05Gisle Vanem fixed a windows compiler warningDaniel Stenberg
2004-02-05options we get as longs need to be typecasted when assigned to preventDaniel Stenberg
picky compiler warnings
2004-02-05fix return type to silence compiler warningsDaniel Stenberg
2004-02-05prevent warning from that picky MIPSpro compilerDaniel Stenberg
2004-02-05compiler warning fix, compare struct pointers of the same typeDaniel Stenberg
2004-02-05use the timeout options when waiting for the server to connect when usingDaniel Stenberg
PORT Provide better error messages to allow debugging if one if the ipv6-related name functions fail in the ftp_use_port() function. This might help us diagnose the problems on AIX. Also make sure getaddrinfo() uses NULL and not "0" for the service argument.
2004-02-05hide the pack_hostent proto if ipv6 is enabled, as figured out by Tor ArntsenDaniel Stenberg
2004-02-03the unused quit-function didn't pass a correct variable type to theDaniel Stenberg
response reading function
2004-02-03added the missing ares numerical version initialiserDaniel Stenberg
2004-02-02undef accept before defining it, since AIX 5.2 has it as a define!Daniel Stenberg
2004-02-02now provides c-ares version info in both version callsDaniel Stenberg
2004-02-02adjusted to the modified ares_strerror() functionDaniel Stenberg
NOTE that this breaks ares-compatibility, we have now officially taken the turn into the c-ares path. We will now officially depend on c-ares for asynch name resolves.
2004-02-02Dirk Manske fixed the ares usage even more. We could get a timeout from aresDaniel Stenberg
as well, and when failing and not getting a timeout we now include the error message ares can provide us with.
2004-02-02set the 'retry' bit to TRUE when the connection is about to be retried,Daniel Stenberg
this allows the HTTP code to *not* return a failure just because no data has been received from the server
2004-02-02clear the sockerror if no error was returnedDaniel Stenberg
2004-02-02we include errno.h to truly know if we have ECONNRESET or notDaniel Stenberg
2004-02-02Timeout slow ares name lookups. This is based on the patch brought byDaniel Stenberg
Dirk Manske, but modified by me.
2004-02-02when including ca-bundle.h, don't look in the current dir first, simply useDaniel Stenberg
the search path since we want the build-version rather than the one in the source dir
2004-01-30make sure the connection is closed when it was detected reset!Daniel Stenberg
2004-01-30Somewhat crude attempt at fixing the test 91 failures. I commit this nowDaniel Stenberg
so that the automatic testing hosts will test these changes over the weekend.
2004-01-30only do the verifyconnect() clear magic on mpeixDaniel Stenberg
2004-01-29added verifyconnect proto and use it correctly in the waitconnect functionDaniel Stenberg
2004-01-29moved the definitions of IOCTL_3_ARGS to setup.hDaniel Stenberg
2004-01-291. changed order of two include files to build fine on MPE/iXDaniel Stenberg
2. now reads the socket error before check connect status, also to make us run fine on MPE/iX
2004-01-29Dan Fandrich's cleanup patch to make pedantic compiler options cause lessDaniel Stenberg
warnings. Minor edits by me.
2004-01-28Chunked-transfers should have an additional CRLF after the final 0 CRLFDaniel Stenberg
sequence.
2004-01-27CONNECT response headers are now passed back as "regular" headersDaniel Stenberg
2004-01-27very big transfers now get nicer progress displayed after 9999 megabytes haveDaniel Stenberg
been transfered!
2004-01-26when saving in a cookie jar fails, include the file name in the error messageDaniel Stenberg
to make it easier to track down
2004-01-23Proxy username and password on persistant connections could easily getDaniel Stenberg
messed up. Vincent Bronner detected this.
2004-01-23check the arguments to the socks5 function, as the name and password mightDaniel Stenberg
be NULL pointers, and if non-NULL if now support zero-length names/passwords