moparisthebest/curl
1
0
Fork 0
mirror of https://github.com/moparisthebest/curl synced 2024-12-22 16:18:48 -05:00
Code Issues Releases Wiki Activity
2,229 Commits 10 Branches 144 Tags 113 MiB
e21926f7f0
Commit Graph

924 Commits

Author SHA1 Message Date
Daniel Stenberg
e177f14595 SSL writes passed back a silly length... 2002-01-16 23:28:58 +00:00
Daniel Stenberg
5c6eddcadd fixed time-out returned without error text set 2002-01-16 22:26:01 +00:00
Daniel Stenberg
b3b4786990 Kevin Roth's SSLeay() patch, slight edited by me. Works with OpenSSL 0.9.5 
now.
 2002-01-16 17:45:08 +00:00
Daniel Stenberg
8d97792dbc - shrunk the BUFSIZE define from 50K to 20K 
- made a separate buffer for uploads (due to the non-blocking stuff)
- added two connectdata struct fields for non-blocking uploads
 2002-01-16 14:53:19 +00:00
Daniel Stenberg
8d07c87be7 modified to deal with the new non-blocking versions of Curl_read() and 
Curl_write().
 2002-01-16 14:50:53 +00:00
Daniel Stenberg
ed21701df3 Curl_write's 5th argument now is signed 2002-01-16 14:49:51 +00:00
Daniel Stenberg
df01507582 Curl_read() and Curl_write() are both now adjusted to return properly in 
cases where EWOULDBLOCK or equivalent is returned. We must not block.
 2002-01-16 14:49:08 +00:00
Daniel Stenberg
f2bda5fd5b Curl_write() now takes a different 5th argument 2002-01-16 14:47:50 +00:00
Daniel Stenberg
cba9838e8f Somewhat ugly fix to deal with non-blocking sockets. We just loop and try 
again. THIS IS NOT A NICE FIX.
 2002-01-16 14:47:00 +00:00
Daniel Stenberg
b6dba9f5dd Somewhat ugly fix to deal with non-blocking sockets. We just loop and try 
again. THIS IS NOT A NICE FIX. We should/must make a select() then and only
retry when we can write to the socket again.
 2002-01-16 14:46:00 +00:00
Daniel Stenberg
4931fbce49 Curl_read() now returns a negative return code if EWOULDBLOCK or similar 2002-01-14 23:14:59 +00:00
Daniel Stenberg
fefc7ea600 a memory leak when name lookup failed is now removed 2002-01-14 23:14:24 +00:00
Sterling Hughes
a1f910c159 Remove erreaneous include, setup.h is included one line above 2002-01-14 05:36:28 +00:00
Daniel Stenberg
d84a0c51e0 Cris Bailiff found out that when the SSL session cache was filled, libcurl 
would crash. This corrects the problem.
 2002-01-09 09:38:37 +00:00
Daniel Stenberg
d57e09889a added a missing failf() before returning an error code 2002-01-08 23:23:24 +00:00
Daniel Stenberg
eecb86bfb0 this seems to correct the SSL reading problem introduced when switching 
over to non-blocking sockets, but this loops very nastily. We should return
back to the select() and wait there until more data arrives, not just blindly
attempt again and again...
 2002-01-08 23:19:32 +00:00
Daniel Stenberg
75bba0da92 added two typecasts to prevent compiler (gcc3) warnings 2002-01-08 07:06:07 +00:00
Sterling Hughes
c0bfe7be15 1) the dns_cache_timeout should be an integer, not a bool 
2) in the curl_dns_cache_entry structure, timestamp should be
a time_t instead of an integer (although I doubt it matters).
 2002-01-08 04:30:59 +00:00
Sterling Hughes
22ac08e06d Add support for DNS cache timeouts via the CURLOPT_DNS_CACHE_TIMEOUT option. 
The default cache timeout for this is 60 seconds, which is arbitrary and
completely subject to change :)
 2002-01-08 04:26:47 +00:00
Daniel Stenberg
87037136ef As identified in bug report #495290, the last "name=value" pair in a 
Set-Cookie: line was ignored if they didn't end with a trailing
semicolon. This is indeed wrong syntax, but there are high-profile web sites
out there sending cookies like that so we must make a best-effort to parse
them.
 2002-01-07 23:05:36 +00:00
Daniel Stenberg
2182e37433 the bool typedef is now made unsigned, to make sure it stays that on all 
platforms, unrelated to what they might prefer by default
 2002-01-07 22:47:21 +00:00
Daniel Stenberg
1de82b220d removed silly check for >=0 of a supposedly unsigned value! 2002-01-07 22:46:38 +00:00
Sterling Hughes
bd878756fc Probably not necessary, but good practice. 2002-01-07 20:55:35 +00:00
Sterling Hughes
8d7f402efb Make cach'ing work with threads now, there are now three cases: 
- Use a global dns cache (via setting the tentatively named,
    CURLOPT_DNS_USE_GLOBAL_CACHE option to true)
    - Use a per-handle dns cache, by default
    - Use a pooled dns cache when in the "multi" interface
 2002-01-07 20:52:32 +00:00
Daniel Stenberg
d3299beec7 Modified to use non-blocking sockets all the time. 2002-01-07 18:38:01 +00:00
Daniel Stenberg
c69c0c0446 added proper breaks in the switch() 2002-01-07 15:24:52 +00:00
Daniel Stenberg
e31a306a38 HTTP response 204 should be treated similar to 304, that is we must not 
expect (nor read) any response-body
 2002-01-07 14:57:18 +00:00
Daniel Stenberg
d9a7773011 added precautions to not go insane when two matching cookies end up in the 
cookie list, even though they're not supposed to do that...
 2002-01-07 14:56:15 +00:00
sm
2b14916813 Add hash and llist to VC dsp file 2002-01-04 23:48:28 +00:00
sm
1d1530e14c Add hash and llist to VC makefile 2002-01-04 23:47:07 +00:00
Daniel Stenberg
5a0f0023cf replaced printf() => Curl_sendf() 2002-01-04 09:53:39 +00:00
Daniel Stenberg
6dcdb8b821 removed a commented line 2002-01-04 09:53:10 +00:00
Daniel Stenberg
781f52a287 fixed an inet_ntoa() occurance to use inet_ntoa_r() if it is available. 
I also replaced all printf() calls with calls to Curl_failf()
 2002-01-04 09:52:44 +00:00
Daniel Stenberg
f75ff58b4b an unconditional occurance of inet_ntoa() now uses inet_ntoa_r() on all 
platforms that have such a function.
This affects multi-thread running libcurls on IPv4 systems that have VERBOSE
switched on. The previous version was risking that another thread overwrote
the data before it was read out in this thread. There could possibly also
be a slight risk that the data isn't zero terminated for a short while and
thus could cause the thread to crash...
 2002-01-04 09:38:52 +00:00
Daniel Stenberg
ae9bf16dee #include the local "inet_ntoa_r.h" file if no proto was found in the global 
header directory but the function *is* present!
 2002-01-04 09:35:23 +00:00
Daniel Stenberg
17a8bf212f The buffer in ftp_pasv_verbose(), used for gethostbyaddr_r(), is now defined 
to become properly 8-byte aligned on 64-bit archs. Philip Gladstone reported.
 2002-01-04 09:17:52 +00:00
Daniel Stenberg
4fc76afef4 The FTP response lines are now passed to the function callback registered for 
headers.
 2002-01-04 09:03:11 +00:00
Daniel Stenberg
8b6314ccfb merged the multi-dev branch back into MAIN again 2002-01-03 15:01:22 +00:00
Daniel Stenberg
6de7dc5879 Sterling Hughes' provided initial DNS cache source code. 2002-01-03 10:22:59 +00:00
Daniel Stenberg
6aaee5f23b minor changes 2002-01-03 09:43:17 +00:00
Daniel Stenberg
0077b9c0a2 pass an 'int' as the third argument to bind() 2002-01-03 00:51:33 +00:00
Daniel Stenberg
fe37fb5921 Philip Gladstone's 64-bit sparc native compiler compatibility issues fixed. 2002-01-02 10:06:47 +00:00
Daniel Stenberg
8a9098a36c *cool* fix by Bjrn Stenberg, makes proxy transfers work better...! :-) 2001-12-20 15:58:22 +00:00
Daniel Stenberg
28027c2aa2 If nobody is set we won't download any FTP file. If include_header is set, 
we return a set of headers not more. This enables FTP operations that don't
transfer any data, only perform FTP commands.
 2001-12-20 11:22:01 +00:00
Daniel Stenberg
af6c394785 Gtz Babin-Ebell's OpenSSL ENGINE patch 2001-12-17 23:01:39 +00:00
Daniel Stenberg
2eb355733f Marcus Webster's newly added CURLFORM_CONTENTHEADER 2001-12-14 12:59:16 +00:00
Daniel Stenberg
e66cdacb93 minor changes 2001-12-13 07:16:27 +00:00
Daniel Stenberg
c67f2da283 solaris 2.5.1 needs the sys/types.h file before the sys/socket.h 2001-12-11 15:08:27 +00:00
Daniel Stenberg
e192261788 failf() calls should not have newlines in the message string! 2001-12-11 13:13:01 +00:00
Daniel Stenberg
47e67eab26 corrected the comment above gmtime_r 2001-12-07 15:56:57 +00:00
Daniel Stenberg
650b95045d added gmtime_r check 2001-12-07 15:51:59 +00:00
Daniel Stenberg
7b832e1745 Jon Travis suggested fix. when CURLOPT_HTTPGET is used we must assign 
set.upload to FALSE or else we might still get an upload if the previous
operation was an upload!
 2001-12-05 06:47:01 +00:00
Daniel Stenberg
436d147925 Eric's #include fixes for better macos compiles 2001-12-04 13:03:08 +00:00
Daniel Stenberg
7ee6a9dc25 i'm soooo funny 2001-12-04 09:14:41 +00:00
Daniel Stenberg
d52c0b6f05 more comments 2001-12-04 07:47:21 +00:00
Daniel Stenberg
779043f7a3 As Eric Lavigne pointed out, the ftp response reader MUST cache data that 
is not dealt with when we find an end-of-response line, as there might be
important stuff even after the correct line. So on subsequent invokes, the
cached data must be used!
 2001-12-03 13:48:59 +00:00
Daniel Stenberg
7493db2338 Eric nailed a but in strnequal() for macintosh 2001-12-03 12:57:45 +00:00
Daniel Stenberg
c16c017f8b more careful re-use of connections when SSL is used over proxies 2001-12-02 14:16:34 +00:00
Daniel Stenberg
2f03ef39d1 SM renamed the debug DLL 2001-12-02 12:09:00 +00:00
Daniel Stenberg
533c24a471 disabling EPSV is now possible 2001-11-29 12:49:10 +00:00
Daniel Stenberg
6a9697387a stdin is file descriptor 0 2001-11-29 12:48:08 +00:00
Daniel Stenberg
332eb7651a CURLOPT_FTP_USE_EPSV can now be set to FALSE to prevent libcurl from 
attempting to use EPSV before the standard PASV.
 2001-11-28 23:20:14 +00:00
Daniel Stenberg
cfdcf5c933 fill memory with junk on malloc() 2001-11-28 23:19:17 +00:00
Daniel Stenberg
a32cd520bd more more more MORE 2001-11-28 16:00:18 +00:00
Daniel Stenberg
b93a60daf9 the perform "state machine" is more explained now 2001-11-28 15:46:25 +00:00
Daniel Stenberg
e2844f5e04 mods 2001-11-28 15:25:01 +00:00
Daniel Stenberg
f0d3fccd4b Added EPSV which is now unconditionally always tried before PASV, which 
makes it work reaaaaly nicely on IPv6-enabled hosts!
Added SIZE before RETR is made, always done on downloads. It makes us know
the size prior to download much more frequently.
Unfortunately, this breaks all the FTP test cases. *fixfixfix*
 2001-11-28 13:05:39 +00:00
Daniel Stenberg
aff19f64b5 use in_addr_t for inet_addr() return code. Now, now portable is this *REALLY*? 
We should add some configure tests for this!
 2001-11-28 12:16:52 +00:00
Daniel Stenberg
15a56b42d6 used in the new multi interface, not yet actually part of libcurl but 
added to CVS to make them available to others
 2001-11-28 11:09:18 +00:00
Daniel Stenberg
dccc77a325 Eric Lavigne updates 2001-11-27 07:27:32 +00:00
Daniel Stenberg
bbf80d0f93 commented out the EPSV support 2001-11-27 00:50:52 +00:00
Daniel Stenberg
6003f24f78 initial code added to support EPSV (IPv6-style PASV) 2001-11-27 00:48:45 +00:00
Daniel Stenberg
f0ee7115d3 Andrs Garca's minor fix to make it compile on win32 2001-11-23 09:04:56 +00:00
Daniel Stenberg
0e7203be89 this fix seems to make the connect fail properly even on IPv4-only Linux 
machines!
 2001-11-22 13:57:00 +00:00
Daniel Stenberg
1c8da21083 Eric fixed a wild write 2001-11-22 09:40:34 +00:00
Daniel Stenberg
8f304d8167 Eric found a missing comma!! 2001-11-22 09:39:03 +00:00
sm
30a0bd9cf5 Fixed release-ssl build 2001-11-22 00:12:48 +00:00
Daniel Stenberg
d1ea596f88 SM added connect.obj 2001-11-21 23:10:55 +00:00
Daniel Stenberg
17b18bca3c added error text for a failed connect case 2001-11-21 22:57:42 +00:00
Daniel Stenberg
ca0fd33d2d Georg Horn's STARTTRANSFER_TIME patch 2001-11-20 15:00:50 +00:00
Daniel Stenberg
b0130e6b3b use the ws2_32.lib now (Miklos Nemeth reporteD) 2001-11-19 20:09:02 +00:00
Daniel Stenberg
d0c1f3e25b long port => int port, as the c source uses! (Miklos Nemeth found this) 2001-11-19 20:08:01 +00:00
Daniel Stenberg
b244710ddb Miklos Nemeth pointed out the missing connect.obj 2001-11-19 20:06:29 +00:00
Daniel Stenberg
e6dd4a6456 Klevtsov Vadim's time condition fix 2001-11-16 11:21:50 +00:00
Sterling Hughes
8d62e21072 looks better on one line (testing the cvs diffing via mail, but I also think 
this looks a bit better ;)
 2001-11-15 14:16:13 +00:00
Daniel Stenberg
b8ff21124a Samuel Listopad's fix to allow global_init => global_cleanup => global_init 
for ssl
 2001-11-14 07:11:39 +00:00
Daniel Stenberg
6aafc2dfd2 corrected the ftp_getsize() usage, as the HPUX compiler warned on them 2001-11-13 12:46:29 +00:00
Daniel Stenberg
65b22480f4 uninitialized variable 2001-11-13 12:09:05 +00:00
Daniel Stenberg
c2479ccb7a my proxytunnel fix accidentally ruined the normal https connects 2001-11-13 08:34:24 +00:00
Daniel Stenberg
c7cdb0f266 make sure to "read out" the server reply even if we didn't get any data from 
the server when that's the only error
 2001-11-12 22:27:05 +00:00
Daniel Stenberg
92aedf850e made Curl_tvdiff round the diff better and make the subtraction before 
the multiply to not wrap-around
 2001-11-12 22:10:09 +00:00
Daniel Stenberg
05f3ca880f made CURLOPT_HTTPPROXYTUNNEL work for plain HTTP as well 2001-11-12 14:08:41 +00:00
Daniel Stenberg
a18d41a463 include setup.h 2001-11-12 10:19:36 +00:00
Daniel Stenberg
1affbff8f9 new Curl_ConnectHTTPProxyTunnel() function, needs a **lot** of testing!!! 2001-11-12 09:47:09 +00:00
Daniel Stenberg
c55d0bb804 We need at least one millisecond to calculate current speed with! I also 
made the getinfo() stuff divide with 1000.0 now to enforce floating point
since Paul Harrington claims the 7.9.1 still uses even second resolution
in the timers there
 2001-11-12 08:50:59 +00:00
Daniel Stenberg
0ffec712e1 Marcus Webster reported and fixed this read-one-byte-too-many problem... 2001-11-08 15:06:58 +00:00
Daniel Stenberg
fe3a78ab19 we use signal() to ignore signals only as long as we have to, and we now 
restore the previous (if any) signal handler properly on return.
 2001-11-07 14:13:29 +00:00
Daniel Stenberg
1a984ea847 get the previous struct keep_sigact 2001-11-07 12:56:13 +00:00
Daniel Stenberg
2a0cde3041 adjusted after Ramana Mokkapati's comments 2001-11-07 09:39:49 +00:00
Daniel Stenberg
2e32d415c0 myalarm() is history, we now use HAVE_ALARM and we now do our very best to 
1 - restore the previous sigaction struct as soon as we are about to shut
off our timeout
2 - restore the previous alarm() timeout, in case an application or similar
had it running before we "borrowed" it for a while.

No, this does not fix the multi-thread problem you get with alarm(). This
patch should correct bug report #478780:
//sourceforge.net/tracker/?func=detail&atid=100976&aid=478780&group_id=976

If not, please post details!
 2001-11-06 19:33:13 +00:00