squid/icap-2_5 changes

These changesets represents the changes on the squid/icap-2_5 development branch of Squid. If there is any questions please contact squid-dev@squid-cache.org.

Browse repository

2006/07/21 00:29:02dwsquid+2 -2minor fix: typo in debugging statement
2006/06/25 10:39:57chtsanti+31 -21Graeme Bissot reported that when icap server responds with 204 the download
2006/05/25 16:04:55chtsanti+1 -3"Allow: 204" icap header refers to the 204 responses outside of previews
2006/02/17 12:50:52oliv3+2 -2bugfix
2006/02/17 12:45:21oliv3+7 -2Applied the "squidgrowthfix" from Graeme Bisset
2006/02/17 12:43:35oliv3+9 -6Applied the "http0.9fix" patch from Graeme Bisset
2006/02/17 12:41:16oliv3+11 -1Applied the "closebeforesendfix" from Graeme Bisset
2006/02/16 17:21:45oliv3+80 -19Added X-Server-IP support.
2005/12/06 21:53:44dwsquid+11 -17This patch eliminates a double call to icapService() for REQMOD
2005/12/01 00:37:09dwsquid+7 -4After more testing, the assertions in clientReadRequest are a bad idea
2005/11/30 23:39:53dwsquid+25 -6Changed icapReqModParseHttpRequest() a little so that it copies the
2005/11/30 22:07:58dwsquid+2 -2Changed "#ifdef ICAP_PROXY_KEEP_ALIVE" to "#if ICAP_PROXY_KEEP_ALIVE"
2005/11/30 21:52:14dwsquid+16 -21Put proxy_keepalive flag back into request_t. It almost worked
2005/11/23 20:36:35dwsquid+4 -3gindent trashed this ugly #ifdef
2005/11/23 20:34:34dwsquid+10 -11ran gindent
2005/11/23 20:33:06dwsquid+50 -27Change the way we do pconn accounting.
2005/11/22 23:42:07dwsquid+21 -16moved client-side proxy_keepalive flag from request_t to ConnStateData.
2005/11/22 22:41:48dwsquid+193 -205ran 'gindent -br -ce -i4 -ci4 -l80 -nlp -npcs -npsl -d0 -sc -di0 -psl'
2005/11/22 22:36:51dwsquid+9 -3A hack to avoid counting pconns for faked-up ICAP connStateData
2005/11/06 19:48:32chtsanti+21 -5A small proxy keep-alive related hack
2005/11/02 18:21:26chtsanti+4 -1If the service was unreachable and the next options request to the
2005/11/01 19:33:00chtsanti+19 -5The proxy-keepalive patch was completelly wrong.
2005/10/30 07:30:15chtsanti+7 -5applying Mateus Groess patch in order to compile using gcc 2.95.x
2005/09/26 17:26:24dwsquid+7 -8bugfix: icapOptReadReply() was marking a service as unavailable if
2005/09/24 10:31:00chtsanti+31 -0Adding errors/Greek/ERR_ICAP_FAILURE file
2005/09/24 07:28:50chtsanti+1 -10Remove the ICAP_PROXY_KEEP_ALIVE define.
2005/09/22 19:59:48chtsanti+14 -1A try to make proxy-keepalive works using icap client.....
2005/09/12 17:15:35chtsanti+4 -4201 responces is not bad idea and exists at least one antivirus icap server
2005/05/16 20:52:40hno+10 -5 - strnstr prototype in util.h. needed for compiling strnstr.c
2005/05/15 20:10:33chtsanti+108 -39http 0.9 support (no responce headers)
2005/05/14 07:55:58chtsanti+7 -3Memory leaks corrections
2005/04/26 16:09:59chtsanti+14 -10When the upload of an object aborted by the web client the
2005/04/24 15:16:49chtsanti+37 -9When a request aborted the icap module stopped the proccesing
2005/04/05 17:57:01chtsanti+60 -52- Rearrangement of icap code in http.c
2005/04/01 17:17:41chtsanti+36 -24ICAP request modifications corrections: In some cases
2005/03/31 17:31:01chtsanti+9 -9lfs related changes ....
2005/03/30 18:51:05chtsanti+9 -2If the remote http server closed the connection
2005/03/30 15:18:52chtsanti+3 -3Corrections related with large file support ...
2005/03/28 17:09:38chtsanti+13 -20some cleanup's ...
2005/03/24 18:03:55chtsanti+3 -2If the squid did not have all the http headers readed
2005/03/20 18:11:47chtsanti+341 -223- KeepAlive/close connection problems
2005/03/13 17:57:49chtsanti+922 -120Some pieces of code not updated during merging with s2_5 tree.
2005/03/07 09:41:50oliv3+2 -2Changed debug level
2005/03/07 09:26:36oliv3+40 -42Several RESPMOD + preview related patches.
2005/02/28 11:47:19oliv3+6 -1Better handling of 400 response
2005/02/28 11:01:26oliv3+13 -2fix for unreachable services
2005/02/22 13:13:04oliv3+4 -3added a define for the 100 response support (still experimental)
2005/02/22 12:21:57oliv3+11 -2Bugfix for RESPMOD when a request was aborted
2004/12/10 08:51:02oliv3+11 -3bug 1143 fix
2004/11/19 09:47:02oliv3+53 -16High-availability patch from <stephane.davy@fr.alcove.com>
2004/11/19 09:38:13oliv3+11 -2High-availability patch from <stephane.davy@fr.alcove.com>
2004/10/27 08:20:46oliv3+5 -3update
2004/10/20 13:27:23oliv3+2 -2Patch from Graeme Bisset <gbisset@packetdynamics.com>
2004/10/19 15:41:02oliv3+26 -14Added X-Client-IP/X-Authenticated-User detection from OPTIONS response
2004/10/19 14:18:05oliv3+63 -44Applied patch from Graeme <gbisset@packetdynamics.com>
2004/10/19 08:51:13oliv3+103 -216preparing new merge
2004/09/21 10:30:57oliv3+6 -3applied patch from Dmitry Adamushko <da@vba.com.by>
2004/09/15 23:21:55dwsquid+9 -3A patch from Olivier Girondel that figures out of we need to
2004/08/20 21:55:48dwsquid+3 -3Undo one of Olivier's recent changes with req-body and null-body
2004/08/04 17:24:36dwsquid+1 -2oops, I was supposed to remove ICAP_CHUNKED since its not used anywhere.
2004/08/04 17:21:58dwsquid+5 -4Patch From olivier <olivier@dolphian.com>
2004/08/04 17:20:24dwsquid+10 -6Patch From olivier <olivier@dolphian.com>
2004/08/04 17:18:11dwsquid+2 -2Patch From olivier <olivier@dolphian.com>
2004/08/04 17:16:23dwsquid+2 -7Patch From olivier <olivier@dolphian.com>
2004/08/04 17:14:46dwsquid+2 -2typo
2004/08/04 17:13:00dwsquid+14 -8Patch From olivier <olivier@dolphian.com>
2004/08/04 17:02:25dwsquid+2 -2Patch from olivier <olivier@dolphian.com>
2004/08/04 16:59:58dwsquid+2 -1Patch from olivier <olivier@dolphian.com>
2004/08/04 16:55:42dwsquid+3 -3Patch from olivier <olivier@dolphian.com>
2004/08/04 16:52:25dwsquid+16 -16Patch from olivier <olivier@dolphian.com>
2004/07/02 10:38:14rhorstmann+2 -3fixed calculation of the encapsulated header as proposed by
2004/05/13 16:26:19dwsquid+41 -13A user noticed that Squid's byte hit ratio numbers are really screwed
2004/05/03 12:24:09rhorstmann+55 -11applied icap timeout patch from Stephane Davy.
2004/04/30 23:33:59dwsquid+16 -2The ICAP URI parser in cache_cf.c was too complex and had at least one
2004/04/06 13:37:16rhorstmann+6 -6removed some compiler warnings
2004/04/06 13:04:37rhorstmann+138 -1added configure/Makefile check for strcasestr function.
2004/04/06 10:19:49rhorstmann+2 -1icap persistent connections were broken during update to stable5.
2004/04/03 21:12:55dwsquid+27 -5use case insensitive string comparison when searching for
2004/03/23 08:20:05rhorstmann+13 -11added german translation from Florian Effenberger.
2004/01/14 08:23:21rhorstmann+31 -15some icap servers seem to ignore the "Connection: close" header for options
2004/01/13 15:24:10rhorstmann+10 -8adjusted some debug statements to make it easier to track down
2003/12/08 12:30:56rhorstmann+841 -4copied error template ERR_ICAP_FAILURE to all language subdirectories
2003/12/08 11:55:56rhorstmann+1 -6removed the -i option which was commented out anyway
2003/12/05 18:00:52rhorstmann+2 -2fixed compiler error.
2003/12/02 14:18:51rhorstmann+2 -47use urlCanonical and remove icapUrlCanonical again
2003/12/01 14:23:44rhorstmann+98 -37Added new error template for ICAP. Implemented bypass for
2003/11/28 14:34:39rhorstmann+94 -9Added X-Authenticated-User patch from GaŽl Roualland <gael.roualland@oleane.net>
2003/11/28 13:47:10rhorstmann+104 -41Implemented bypass for reqmod.
2003/11/26 08:42:45rhorstmann+53 -2urlCanonicalClean may strip away everything after a question mark.
2003/11/24 15:51:10rhorstmann+3 -2When calling httpStateFree from icap, fd might be -1. With delay pools
2003/11/18 16:50:04dwsquid+516 -315Synchronized with cvs.squid-cache.org branch s2_5
2003/11/14 20:10:48dwsquid+6 -3bugfix for the case when we don't get the entire HTTP response in
2003/11/14 17:28:46dwsquid+8 -8fix a memory leak
2003/11/14 02:09:21dwsquid+16 -15Need to make sure that icapReadReply3() is called for ICAP 204
2003/11/14 01:34:36dwsquid+4 -5icap->flags.icap_server_eof is unused. changing them to http_server_eof
2003/11/14 00:37:35dwsquid+17 -14fixed a bug with ICAP 204 responses and HTTP replies with content-length 0.
2003/11/13 23:48:54dwsquid+3 -1handle a case with 'res-hdrs' and 'null-body' when calculating the expected
2003/11/13 23:39:39dwsquid+21 -5fixed some bugs with ICAP 204 replies. Need to zero buffer size after
2003/11/11 23:29:25dwsquid+19 -3Better handling of the case when Squid fails to connect to the ICAP server
2003/11/07 05:52:53dwsquid+6 -6change buildRespModHeader()'s local 'bs' (bytes sent) into a
2003/11/06 22:28:45dwsquid+41 -19more fixes to buildRespModHeader(). The recently committed version
2003/11/06 22:26:28dwsquid+6 -5Change the way that 'chunk_buf' and memBufRead() work. Instead of using
2003/11/06 22:24:58dwsquid+7 -4protect calls to icap RESPMOD with cbdataValid checks
2003/11/06 18:18:53dwsquid+3 -4It will be better to initialize the chunk_buf MemBuf as soon as the
2003/11/06 07:32:39dwsquid+2 -1need to initialize 'chunk_buf' when REQMOD does a handoff to RESPMOD.
2003/11/06 07:31:51dwsquid+2 -2when reading from ICAP server, also handle the "connection closed"
2003/11/06 06:41:12dwsquid+23 -12handle the case when http.c doesn't get the full HTTP response headers
2003/11/04 18:22:33dwsquid+124 -59Numerous fixes to handle cases where the buffer we are parsing
2003/10/31 17:24:57dwsquid+9 -1bugfix: make icapPconnTransferDone() properly detect null-body responses.
2003/10/31 08:20:19rhorstmann+0 -2removed old icap source entries
2003/10/30 21:22:49dwsquid+27 -58Removed do_icap_* flags from struct _request_flags. They are essentially
2003/10/30 18:01:36dwsquid+87 -21Support for ICAP persistent connections in REQMOD transactions.
2003/10/30 17:58:02dwsquid+4 -2add fd_bytes() calls to account for data read from ICAP server.
2003/10/30 13:03:26rhorstmann+8 -3set all the do_icap* flags in the request struct to 1 when an icap acl
2003/10/30 12:41:32rhorstmann+6 -1copy the icap specific parts of the request struct when creating a new request
2003/10/29 22:08:47dwsquid+2 -2oops, increase debugging level for this message
2003/10/29 22:05:30dwsquid+249 -62Support for sending HTTP request bodies to the ICAP server for REQMOD.
2003/10/29 20:31:36dwsquid+6 -6For successful ICAP REQMOD replies, link the fake clientHttpRequest
2003/10/29 17:40:23dwsquid+23 -9Need to pass start time and log_addr to icapReqModStart() for proper
2003/10/29 17:24:57dwsquid+30 -47combine replicated code into icapReqModCreateClientState().
2003/10/29 16:29:58dwsquid+137 -139Move icapReadChunkSize() and icapReadChunkedBody() from icap_respmod.c
2003/10/29 16:18:05dwsquid+21 -14Need to split request_flags->do_icap into four new flags for
2003/10/27 18:06:44dwsquid+7 -7bug from merging getICAPReqModString()
2003/10/27 18:04:10dwsquid+75 -31added comments for each function, also removed getICAPReqModString()
2003/10/27 17:33:37dwsquid+7 -3Change assertion into a soft error when Squid does not find end
2003/10/23 19:30:11uid131750+3 -3support an ICAP REQMOD response that contains an HTTP REPLY.
2003/10/23 18:54:42uid131750+46 -4support an ICAP REQMOD response that contains an HTTP REPLY.
2003/10/23 18:39:42uid131750+9 -4in icapConnect(), configure timeout and close handler before calling
2003/10/23 15:18:53dwsquid+8 -8correct typo: icapReadModReadHttpPart -> icapReqModReadHttpPart
2003/10/17 20:51:26dwsquid+4 -1Memory leak casued by forgetting to unlock httpState->icap_writer.
2003/10/17 19:53:09dwsquid+73 -26moved our own strnstr() from src/icap_common.c to lib/strnstr.c
2003/10/17 15:59:41rhorstmann+23 -1added missing strnstr (copied from gnutls)
2003/10/17 15:13:21dwsquid+209 -216Fixes to the MSG_PEEK code.
2003/10/17 15:01:47dwsquid+36 -36gindent-1.9.1 -br -ce -i4 -ci4 -l80 -nlp -npcs -npsl -d0 -sc -di0 -psl
2003/10/15 22:21:28dwsquid+2 -1need to call fd_note() after reusing a persistent connection so that
2003/10/15 22:20:48dwsquid+3 -3make 'icap_preview_enable' disabled by default because of bugs in
2003/10/15 21:19:23dwsquid+8 -1Must parse the HTTP reply into the 'reply' structure so we can get
2003/10/15 20:01:23dwsquid+73 -21Some support for persistent connections to the ICAP server
2003/09/25 16:24:39dwsquid+8 -7The storeAbort() call from icapStateFree() is necessary (e.g., in
2003/09/25 08:29:43rhorstmann+17 -18prevent rescheduling of options requests on "-k rotate"
2003/09/24 19:35:01dwsquid+18 -5Added 'icap_preview_enable' directive to allow Squid the option
2003/09/23 23:48:55dwsquid+3 -3run 'gindent' and change debugging level on recent 100 Continue code.
2003/09/23 23:47:34dwsquid+6 -1handle ICAP "100 continue" responses (outside of preview mode)
2003/09/23 22:41:16dwsquid+114 -23Support for ICAP 204 (no content) responses.
2003/09/23 12:43:00rhorstmann+5 -1reset opt pointers after call to icapOptDataFree
2003/09/23 12:33:04rhorstmann+4 -2don't start option request when icap is disabled
2003/09/23 12:28:02rhorstmann+3 -1fix memory leak in icapCheckAcl
2003/09/22 21:30:28dwsquid+14 -16ran gindent-1.9.1 -br -ce -i4 -ci4 -l80 -nlp -npcs -npsl -d0 -sc -di0 -psl
2003/09/22 19:50:56dwsquid+19 -5patch from webwasher:
2003/09/22 19:19:40dwsquid+0 -1494this file now replaced with icap_{common,reqmod,respmod}.c
2003/09/22 19:17:14dwsquid+2049 -190Significant changes to the ICAP implementation. In particular,
2003/09/14 02:17:36dwsquid+3 -10make icap.c compile again. With merge from s2_5, MemObject->fd went
2003/09/14 00:54:23dwsquid+2 -2typo in if-block, missing 'else'
2003/09/14 00:53:21dwsquid+9 -5*** empty log message ***
2003/08/27 13:08:54dwsquid+2 -2Use getOutgoingAddr(NULL) instead of 'no_addr' because commResetFD()
2003/08/21 23:45:35dwsquid+2 -1add debugging in icapSendRespMod()
2003/08/21 23:41:50dwsquid+4 -10cleanup icapReadResponse() -- don't need a MemObject, etc
2003/08/21 23:40:45dwsquid+28 -3add icapRespModGobble() -- reads bytes from ICAP server up to
2003/08/21 23:35:12dwsquid+10 -11minor changes to debugging and comments
2003/08/21 23:27:42dwsquid+10 -1add 'Encapsulated' header parsing to icapRespModReadReply()
2003/08/21 23:14:08dwsquid+5 -2passing 'no_addr' to comm_open() causes commBind() to fail in
2003/08/21 23:11:37dwsquid+7 -7gindent -br -ce -i4 -ci4 -l80 -nlp -npcs -npsl -d0 -sc -di0 -psl
2003/08/21 22:08:48dwsquid+110 -55clean up icapReqModReadReply and icapRespModReadReply functions.
2003/08/21 20:59:44dwsquid+6 -6rename scheduleReadReply -> icapScheduleReadReply and make it static
2003/08/21 20:52:39dwsquid+19 -3make icapParseEncapsulated() understand and parse: req-hdr, req-body, opt-body
2003/08/21 20:47:13dwsquid+14 -9some better (more verbose) debugging for httpPconnTransferDone()
2003/08/21 20:45:39dwsquid+10 -8change icap->reqmod to icap->flags.reqmod
2003/08/21 20:43:09dwsquid+78 -79changed debugging section to avoid overlapping other modules
2003/07/09 21:25:23hno+5 -6Further cleanup of unneeded non-static changes
2003/07/09 21:19:36hno+3 -3Removed unneeded changes to make certain functions non-static
2003/06/27 01:15:18hno+639 -652Indented according to Squid-2 coding style
2003/06/27 01:05:34hno+43 -34Rewrite of how icap replaces the original request in REQMOD to now
2003/06/27 01:03:08hno+3 -1Note: reqmod_precache and respmod_postcache is not yet implemented
2003/06/10 12:57:41hno+16 -30icap squid -k reconfigure bugfix.
2003/06/03 16:10:57hno+9 -4Build a fake request in REQMOD to have the request forwarded as a proxy
2003/05/27 12:55:19hno+16 -17hook infrastructure used by ICAP REQMOD when reading the modified
2003/05/09 17:00:03hno+59 -34Make the REQMOD readin of the modified request async.
2003/05/09 14:14:31hno+21 -23removed the ICAP_FIELD macro to make the code look more like normal
2003/04/07 13:33:44hno+26 -10Oops.. did not mean to back out those Squid-2.5 changes..
2003/04/07 13:28:50hno+69 -95Prototype cleanups to make GCC happy. No code changes.
2003/04/04 16:57:23hno+1864 -0Imported ICAP client 1.2.1
2003/04/04 16:53:51hno+1127 -56Imported ICAP client 1.2.1 with update for Squid-2.5.STABLE2

Release statistics

chtsanti23+1695 -571
dwsquid96+4986 -3558
hno15+3910 -970
oliv320+465 -385
rhorstmann25+1521 -235
uid1317503+58 -11
total182+12635 -5730