summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2001-11-21Include the protocol variables in the conn_s structure.Robert James Kaes2-2/+8
2001-11-21Extract the version from the request line and store it in the conn_sRobert James Kaes1-9/+44
structure. Don't send a Connection header to the remote connection. Handle the Via header correctly.
2001-11-12Changed the read() call into a recv() call with a flag of NOSIGNAL since IRobert James Kaes1-5/+9
don't want signals messing up my calling conventions.
2001-11-05Added profiling flags to the DEBUG version of CFLAGS.Robert James Kaes1-1/+2
2001-11-05Made is_anonymous_enabled() an inline function.Robert James Kaes1-2/+2
2001-11-05Added a clean rule to remove the gprof files.Robert James Kaes1-1/+5
2001-11-05Replaced all calls to buffer_size() with the macro BUFFER_SIZE().Robert James Kaes1-7/+7
2001-11-05Removed the buffer_size() function and turned it into BUFFER_SIZE() macro.Robert James Kaes3-60/+81
Moved the struct buffer_s into the header file. Added more assert() calls to better document the assumptions the functions make. Removed incorrect code in remove_from_buffer() which was never actually called anyway.
2001-11-03ACL fixes. Use a better differentiation between string and numericRobert James Kaes1-0/+5
addresses.
2001-11-03Tightened the string/numeric ACL checks.Robert James Kaes1-6/+13
2001-11-02Informing the user of the connection to the tunnel is a INFO type message,Robert James Kaes1-2/+2
not a CONN level message.
2001-10-26Removed all the aclocal.m4, acconfig.h, etc. files since they are includedRobert James Kaes1-1/+1
automatically by the autoconf scripts.
2001-10-26No longer needs the file. When somebody helps with tinyproxy they will beRobert James Kaes1-45/+0
thanked in the ChangeLog entry relating to the problem solved.
2001-10-25Changes reflecting the new conns.* files and the organization of theRobert James Kaes1-0/+8
header files.
2001-10-25Header reorganization. Basically all system headers are now included inRobert James Kaes10-66/+28
tinyproxy.h and all the other files include the tinyproxy.h header. This moves all the dependancy issues into one file.
2001-10-25Changed the included headers to be only the "tinyproxy.h" since itRobert James Kaes2-7/+4
contains all the code for properly bringing in the system headers (in the correct order and with proper dependancies.)
2001-10-25Some code clean up due to the creation of the conns.{c,h} files which nowRobert James Kaes6-47/+13
store the creation/destruction and definition of the connection structure.
2001-10-25Moved the connection creation/destruction and definition into it's ownRobert James Kaes2-0/+99
files.
2001-10-25Updated the news file.Robert James Kaes1-15/+17
2001-10-25Fixed a few spelling problems and added additional information regardingRobert James Kaes1-11/+30
the new configure arguments and information about the mailing lists.
2001-10-25Moved the code to include the TODO file into the top level Makefile.amRobert James Kaes2-2/+2
2001-10-25Moving the TODO list from the doc directory to the root directory so thatRobert James Kaes1-0/+0
it's more visible.
2001-10-25Bumped up the pre-version.Robert James Kaes1-2/+2
2001-10-25Changes to reflect the various fixes to have CONNECT upstream proxyingRobert James Kaes1-0/+10
work.
2001-10-25Fixed the problems with sending/not sending headers so that the CONNECTRobert James Kaes1-15/+9
(SSL) method would work with upstream proxying!
2001-10-25Added the upstream flag to the conn_s structure.Robert James Kaes1-1/+2
2001-10-25Bit of source code clean up.Robert James Kaes1-1/+2
2001-10-25More changes to reqs.c and buffer.cRobert James Kaes1-0/+14
2001-10-25Changed the way relay_connection() works by having the both buffersRobert James Kaes1-22/+42
emptied when either socket is closed. This should be better for the tunnel connections. Change the connect_to_upstream() function to better utilize the establish_http_connection() function. Code re-use is cool. :)
2001-10-25Lowered the size of the read buffer to 2K since Ethernet seems to workRobert James Kaes1-4/+5
with 1400 byte segments.
2001-10-24Error message cleanup to make the information more informative.Robert James Kaes5-39/+35
2001-10-23Fixed a spelling problem in a log message.Robert James Kaes1-2/+2
2001-10-23Update to reflect the current maintainer (Robert) again.Robert James Kaes1-1/+4
2001-10-23New changes to getpeer_string().Robert James Kaes1-0/+3
2001-10-23Moved the zero-length string setting to outside the tests in getpeer_*()Robert James Kaes1-4/+15
functions. Also added a more robust error reporting for DNS errors.
2001-10-23Hopefully cleaned up the source a bit.Robert James Kaes1-1/+1
2001-10-23Cleaned up the file and added the request for a more consistent errorRobert James Kaes1-22/+40
system.
2001-10-22Bumped the version number.Robert James Kaes1-2/+2
2001-10-22New TODO entry for handling potential security problems.Robert James Kaes1-0/+2
2001-10-22Changed the getpeer_*() functions to return the reason why the name or IPRobert James Kaes1-3/+3
address could not be looked up.
2001-10-22Cleaned up the code in handle_connection() and added a NULL pointer testRobert James Kaes2-18/+26
in the free request function.
2001-10-22Fixed a problem where a user's variable would be uninitialized if a DNSRobert James Kaes2-1/+11
lookup failed.
2001-10-19More moving around of the source code. Still can't get CONNECT to workRobert James Kaes2-55/+67
with an upstream proxy.
2001-10-18Fixed a problem where a full buffer on the read side would cause theRobert James Kaes2-2/+8
connection to be closed.
2001-10-18Added netinet/in.h so that tinyproxy will compile on OpenBSD.Robert James Kaes1-1/+2
2001-10-17Trying to allow CONNECT methods through an Upstream proxy.Robert James Kaes1-18/+33
2001-10-01Added the ability to compile tinyproxy with static linking.Robert James Kaes4-7/+26
2001-09-28Did a couple of TODOs and added a few new ones. :)Robert James Kaes1-3/+2
2001-09-16More changes, mostly related to the support for an upstream proxy.Robert James Kaes1-0/+21
2001-09-16send_http_message() doesn't create a memory block and store it in theRobert James Kaes1-11/+5
connection's output_message variable. Instead the error is sent to the client right away. Once we finish processing the client's headers it will automatically accept the error message. So we get the same result, but less memory is used.