aboutsummaryrefslogtreecommitdiff
path: root/lib
AgeCommit message (Collapse)Author
2004-03-15if the global_init() is called from within curl_easy_init() and returnsDaniel Stenberg
an error code, we now make curl_easy_init fail and return NULL.
2004-03-15Initial commit of the first attempt to make three new *strerror() functions.Daniel Stenberg
No protos in the headers yet and no docs.
2004-03-14fix signed and unsigned warningsDaniel Stenberg
2004-03-13postsize is off_t now, so we typecase it to int before doing normal printfDaniel Stenberg
with it (knowing it won't be larger than what fits in an int)
2004-03-13the postsize is an off_t so use the proper printf format to output theDaniel Stenberg
content-length when doing multipart posts
2004-03-12more variable type fixing for the huge postsDaniel Stenberg
2004-03-12more variable type fixes for the large POST supportDaniel Stenberg
2004-03-12Made the 'postsize' variable an off_t type to be able to hold large fileDaniel Stenberg
sizes if desired
2004-03-12Added CURLOPT_POSTFIELDSIZE_LARGE to offer a large file version of theDaniel Stenberg
CURLOPT_POSTFIELDSIZE option to allow really big HTTP POSTs.
2004-03-11dl and ulspeed are now curl_off_t so typecast them to double when weDaniel Stenberg
return their values
2004-03-11Made max5data() take a curl_off_t size as argument instead of double. ShouldDaniel Stenberg
make the progress meter more accurate for large files. Also made the sprintf usage in that function avoid floating point.
2004-03-11Gisle Vanem's fixes to use CURL_SOCKET_BAD more instead of -1 for sockets.Daniel Stenberg
2004-03-11don't let the EINTR stuff build on windowsDaniel Stenberg
2004-03-10Use more curl_off_t variables when doing the progress meter calculations andDaniel Stenberg
argument passing and try to convert to double only when providing data to the external world.
2004-03-10curl_socket_t mistakes cleanupDaniel Stenberg
2004-03-10turn niflags into a define named NIFLAGSDaniel Stenberg
2004-03-10Nah, ignore the OS define in here. This is being included by ares and itDaniel Stenberg
certainly doesn't need the OS define.
2004-03-10Minor edit to avoid an unreachable break and to remove the extra {} bodyDaniel Stenberg
within the switch.
2004-03-10keep the number of bytes read in a size_t variableDaniel Stenberg
2004-03-10strequal() returns int so we typecast the return to bool when we store theDaniel Stenberg
result as bool
2004-03-10store times in time_tDaniel Stenberg
2004-03-10Jeff Lawson fixed the SSL connection to deal with received signals during theDaniel Stenberg
connect.
2004-03-10Now requires an OS string defined by the config*.h fileDaniel Stenberg
2004-03-10make windows builds use the OS string "i386-pc-win32" instead of just "win32"Daniel Stenberg
to make it more like other OS strings
2004-03-10Make sure SIZEOF_CURL_OFF_T is defined before we check for it being > 4.Daniel Stenberg
It is only undefined when this file is included by others (like ares) and in those cases it doesn't matter.
2004-03-10removed the #if 0'ed pragmas that disable warnings on msvcDaniel Stenberg
2004-03-10moved the curl_socket_t typedef downwardsDaniel Stenberg
2004-03-09Use curl_socket_t instead of int for holding sockets. The typedefs andDaniel Stenberg
defines are in setup.h.
2004-03-09only build with the windows threading trace code ifDaniel Stenberg
DEBUG_THREADING_GETHOSTBYNAME is defined
2004-03-09explicit typecast to visualize that we really want the result of theDaniel Stenberg
operation as a size_t
2004-03-09Tor Arntsen's fix to a AIX build problemDaniel Stenberg
2004-03-09added HAVE_IOCTLSOCKET hereDaniel Stenberg
2004-03-08don't compare signed/unsignedDaniel Stenberg
2004-03-08Commented the Curl_read() arguments.Daniel Stenberg
2004-03-08size_t/int fixDaniel Stenberg
2004-03-08strlen() returns size_tDaniel Stenberg
2004-03-08waitconnect() takes the timeout argument as a longDaniel Stenberg
2004-03-08store size as size_tDaniel Stenberg
use %zd when outputting size_t
2004-03-08don't use 'register'Daniel Stenberg
make strtol() returns get stored in long variables don't mix size_t with int
2004-03-08size_t/int/long fixesDaniel Stenberg
2004-03-08we must not only support long long for Largefile to work, we must haveDaniel Stenberg
a curl_off_t type that is larger than 4 bytes
2004-03-05issue 12 fixDaniel Stenberg
2004-03-05another include to prevent warningsDaniel Stenberg
2004-03-05more fixing to make the progress/getinfo stuff to work properly when doingDaniel Stenberg
file: transfers too
2004-03-05clearly tell that these are ftp response timeoutsDaniel Stenberg
2004-03-04When following to a new URL, we must make sure to call Curl_done() first,Daniel Stenberg
since the current connection must be taken care of properly before we move on. Christopher R. Palmer reported a problem he found due to this mistake.
2004-03-04Andrés García's patch to prevent warnings while compiling with mingw, mainlyDaniel Stenberg
because it is now possible to have both WIN32 and HAVE_CONFIG_H defined.
2004-03-04use size_t for the data, but keep the protos use ssize_t to better fitDaniel Stenberg
with the existing transfer.c code
2004-03-04include the proper header file tooDaniel Stenberg
2004-03-04fix progress data to be updated properly for file: transfers, as reportedDaniel Stenberg
by Jesse Noller