aboutsummaryrefslogtreecommitdiff
path: root/lib
AgeCommit message (Collapse)Author
2003-10-30Only compile the ipv6-section for ipv6-enabled libcurls. Should saveDaniel Stenberg
us some trouble.
2003-10-29David Hull made the file: URL parser also accept the somewhat sloppy fileDaniel Stenberg
syntax: file:/path. I added test case 203 to verify this.
2003-10-28Another glibc resolve name fixDaniel Stenberg
2003-10-28overlapping memory chunks with strcpy(), detected by the friendly valgrindDaniel Stenberg
2003-10-27James Bursa's fix to prevent failf() to write outside its buffer boundaryDaniel Stenberg
2003-10-26snprintf() made a single-byte buffer overflow, as it could write a zeroDaniel Stenberg
outside its given buffer. Discovered and reported by James Bursa.
2003-10-24Resuming a download of an already downloaded document, that is trying to getDaniel Stenberg
a range of a document beyond its size, caused libcurl to "hang" until the server closed the connection and then it returned error 18. This is bad. This way, we don't return any error at all, which isn't nice either, as we need to alert the app somehow that the request range was out of size.
2003-10-24Based on David Hull's fix in bug report 804599, we now check for solarisDaniel Stenberg
and gcc, to set the -mimpure-text link flag for linking the lib better.
2003-10-24if malloc fails, we must bail out nicelyDaniel Stenberg
2003-10-24Improved checks for bad memory situations and proper behaviour when weDaniel Stenberg
run out of memory.
2003-10-24better bailing-out cleanup if a malloc fails in the DNS cacheDaniel Stenberg
2003-10-23Georg Horn's fixes to do different CA cert verifications. They can now beDaniel Stenberg
done even if the result is ignored, as some sites seem to require that.
2003-10-22Dylan Ellicott added the release-libcurl-ssl-dll targetDaniel Stenberg
2003-10-22Do the auth stuff at the end-of-headers and not at the start-of-body, asDaniel Stenberg
we might not get a body when we get a 401 with a set of WWW-Authenticate: headers. This fixes the problem Kevin Roth detected in 7.10.8-pre4 and pre5. Verified by test case 91.
2003-10-21updated for mingw32Daniel Stenberg
2003-10-21conn->user and conn->passwd will now always be set to point to something.Daniel Stenberg
If not user or password is set, they will point to a "" string.
2003-10-20allow no user and password for proxy too, BasicDaniel Stenberg
2003-10-20allow no user and no passwd when doing basic authDaniel Stenberg
2003-10-20lock the DNS cache properly before adding an entry when using asynch DNSDaniel Stenberg
2003-10-19better error output on SSL errors when receiving data - Georg Horn patchDaniel Stenberg
2003-10-19fixed the bug my previous change introducedDaniel Stenberg
2003-10-18minor fix to not shadow a variableDaniel Stenberg
2003-10-18builds warning-free with -Wshadow nowDaniel Stenberg
2003-10-18don't shadow 'socket'Daniel Stenberg
2003-10-18fixed gcc -Wshadow warningsDaniel Stenberg
2003-10-18removed gcc -Wshadow warningDaniel Stenberg
2003-10-18memory leakDaniel Stenberg
2003-10-18gcc -Wshadow complaints fixedDaniel Stenberg
2003-10-18compiler warning fixDaniel Stenberg
2003-10-18prevent compiler warnings with relly picky compiler optionsDaniel Stenberg
2003-10-18Dominick Meglio pointed out these files should be removed from hereDaniel Stenberg
2003-10-17Dominick Meglio implemented CURLOPT_MAXFILESIZE and --max-filesize.Daniel Stenberg
2003-10-17typecasts to prevent warningsDaniel Stenberg
2003-10-17make empty names look fine in verbose outputDaniel Stenberg
2003-10-17make no user or no password just mean blank fields, not aborted operationDaniel Stenberg
2003-10-17o the name and password arrays are 256 bytes, so let's accept that lengthyDaniel Stenberg
input o have ->passwd and ->name be NULL if no name/passwd was given o only set default user+password for FTP if no userpwd was given
2003-10-17made the code deal with empty name and passwordDaniel Stenberg
2003-10-17show info text (verbose) about auth type and user name in useDaniel Stenberg
2003-10-16password promting support removed from libcurlDaniel Stenberg
2003-10-16rewritten alternative name checkDaniel Stenberg
2003-10-15the previous fix left this compiler errorDaniel Stenberg
2003-10-15bad license situation for the altname patchDaniel Stenberg
2003-10-15Avoid doing getsockopt() on Windows to verify connects. It seems that thisDaniel Stenberg
hogs Windows machines when libcurl is being used multi-threaded (with > ~50 threads). Andrew Fuller helped us verify and test this. According to a MSDN web page on connect(), it returns 0 when the connect is done and thus we don't need the getsockopt() call anyway on Windows.
2003-10-14Kimmo Kinnunen fixed a crash with duphandle() when CURLDEBUG was setDaniel Stenberg
2003-10-14Gisle Vanem's IPv6-on-Windows patch applied!Daniel Stenberg
2003-10-13added more mirrored versions of the ares 1.1.1 packageDaniel Stenberg
2003-10-13Updated build instruction since I now offer a complete ares-package with allDaniel Stenberg
the patches already applied. No need to apply any patches manually anymore. ares 1.1.1 is still compatible.
2003-10-13Clear the connection pointer after the async resolve has failed.Daniel Stenberg
This cures the problem reported by Giuseppe Attardi on October 12, 2003.
2003-10-13Removed extra LDFLAGS assignment. Pointed out in debian bug report #212086Daniel Stenberg
(http://bugs.debian.org/212086). Patch by Domenico Andreoli.
2003-10-12Dirk Manske made the share-locking around DNS lookups a bit "looser" so thatDaniel Stenberg
multiple DNS lookups can run simultaneously faster. The downside is that resolving the same host name now can be made at once from multiple threads, but the upside is that threads now don't alwys have to wait for the others' resolves. Test case 506 updated accordingly.