aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2004-10-12#include "strerror.h" to get the strerror protoDaniel Stenberg
2004-10-12recent fixesDaniel Stenberg
2004-10-12one item fixed, one added for 7.12.2 and two more for 7.12.3Daniel Stenberg
2004-10-12add proper error message when send() failsDaniel Stenberg
2004-10-11removed trailing whitespaceDaniel Stenberg
2004-10-11SO_NOSIGPIPEDaniel Stenberg
2004-10-10another lame attempt to avoid the "warning: will never be executed" warningDaniel Stenberg
by gcc 3.4
2004-10-10MSVC uses 'i64' suffix for 64-bit sizes.Gisle Vanem
2004-10-10attempt to make the configure output appear in the build log when it runsDaniel Stenberg
on my solaris 2.7 box too (currently unknown perl version)
2004-10-10set the SIZEOF_LONG define to get the new lib/strtoofft.h fineDaniel Stenberg
2004-10-10If long is 8 bytes we can use strtol() to get 64 bit numbers and won't needDaniel Stenberg
our strtoll() replacement function.
2004-10-10Prevent a longjmp warning by moving the rc assign within Curl_resolv().Daniel Stenberg
Andy Cedilnik reported. Warning on HP-UX?
2004-10-10Use LL suffix for long long constants if the compiler supports it, to preventDaniel Stenberg
warnings.
2004-10-10If long long is supported, check if [num]LL is supported for numericalDaniel Stenberg
constants.
2004-10-08--enable-ares on AIX has problemDaniel Stenberg
2004-10-08Use 'struct lifreq' and SIOCGLIFADDRDaniel Stenberg
2004-10-08killed trailing whitespaceDaniel Stenberg
2004-10-08if basename was found, check for a prototype and if none was found, provideDaniel Stenberg
our own in the formdata.c file to prevent warnings on systems without it
2004-10-08prevent warning with comparison between signed and unsignedDaniel Stenberg
2004-10-07use curlx_strnequal() from the private lib sources instead of strncasecmp()Daniel Stenberg
for maximum portability
2004-10-07use curl_strnequal(), not strncasecmp()Daniel Stenberg
2004-10-07Gisle's new fix, the old file:// leakDaniel Stenberg
2004-10-07use tld_strerror() only if previously detected, since otherwise we can'tDaniel Stenberg
work with libidn < 0.5.6
2004-10-07check for tld_strerrorDaniel Stenberg
2004-10-06Added tld_check_name().Gisle Vanem
2004-10-06Fixed tld_check_name(). idna_to_unicode_lzlz() should never fail,Gisle Vanem
but return FALSE if 'uc_name == NULL' just in case.
2004-10-06USE_LIBIDN: Added Top-level-domain (TLD) check for host->name.Gisle Vanem
Only print a warning if check fails.
2004-10-06test resume and redirectDaniel Stenberg
2004-10-06Chih-Chung Chang reported that if you use CURLOPT_RESUME_FROM and enabledDaniel Stenberg
CURLOPT_FOLLOWLOCATION, libcurl reported error if a redirect happened even if the new URL would provide the resumed file. Test case 188 added to verify the fix (together with existing test 99).
2004-10-06*** empty log message ***Gisle Vanem
2004-10-06updates of todayDaniel Stenberg
2004-10-06avoid warnings on systems with this member set constDaniel Stenberg
2004-10-06removed tabs and trailing whitespace from sourceDaniel Stenberg
2004-10-06Dan Fandrich fix for hosts that need both -lnsl and -lsocketDaniel Stenberg
2004-10-06untabifyDaniel Stenberg
2004-10-05djgpp has basename()Gisle Vanem
2004-10-05minor edit to re-use a variable and to hopefully avoid a (moot) warningDaniel Stenberg
about code that won't be reached
2004-10-05recent fixesDaniel Stenberg
2004-10-05make the given path to --with-libidn override any other installationDaniel Stenberg
2004-10-05avoid warning for unused variableDaniel Stenberg
2004-10-05use idna_strerror() if it is available (only in libidn 0.5.6 or later)Daniel Stenberg
2004-10-05Only include libgen.h if we have a basename as well.Daniel Stenberg
Mainly meant to deal with the IRIX case which seems to requrie a "-lgen" lib to find the basename function and thus without the gen lib, it finds the header but not the function and our replacement function has a prototype that doesn't match the IRIX one. A different approach would be to make configure detect and use -lgen for the systems that require it.
2004-10-05let our basename() be staticDaniel Stenberg
2004-10-04name mix fixDaniel Stenberg
2004-10-04closing in on releaseDaniel Stenberg
2004-10-04Made the dns entry remain locked while a connection to the host remains toDaniel Stenberg
allow verbose output during this period. Bertrand Demiddelaer reported and helped fixing.
2004-10-03set async.done to TRUE last in the addrinfo callback to prevent the risk thatDaniel Stenberg
the multi-threaded resolver does wrong
2004-10-03Replaced the use of isspace() with our own version instead since we have mostDaniel Stenberg
data as 'char *' and that makes us pass in negative values if there is 8bit data in the string. Changing to unsigned causes too much warnings or too many required typecasts to the normal string functions.
2004-10-03when building with libidn support, check for idna_strerror() which isDaniel Stenberg
included in very recent versions
2004-10-03added info about how users get info (like the CURLcode return code) fromDaniel Stenberg
individual transfers