]> git.ipfire.org Git - thirdparty/squid.git/log
thirdparty/squid.git
22 years agoCustom log formats, and selective access logging. See logformat
hno [Mon, 7 Jul 2003 03:50:55 +0000 (03:50 +0000)] 
Custom log formats, and selective access logging. See logformat
and cache_access_log directives

22 years agoExternal acl extensions:
hno [Mon, 7 Jul 2003 03:43:36 +0000 (03:43 +0000)] 
External acl extensions:

%SRCPORT, %MYPORT, %MYADDR
New format tags for matching client TCP connection details

grace=N
defines a grace period on external acl cache entries. If the entry has less
then the grace period left when being refercenced then a new lookup is
initiated but without waiting for the reply.

22 years agoBootstrapped
hno [Sun, 6 Jul 2003 21:42:53 +0000 (21:42 +0000)] 
Bootstrapped

22 years agoBugfix: Speed up processing of queued events significantly, to not cause
hno [Sun, 6 Jul 2003 21:30:42 +0000 (21:30 +0000)] 
Bugfix: Speed up processing of queued events significantly, to not cause
large delays when under low load.

22 years agoBugfix: Speed up processing of queued events significantly, to not cause
hno [Sun, 6 Jul 2003 21:29:35 +0000 (21:29 +0000)] 
Bugfix: Speed up processing of queued events significantly, to not cause
large delays when under low load.

22 years ago--disable-optimizations option to force compiler optimizations (including
hno [Sun, 6 Jul 2003 20:56:12 +0000 (20:56 +0000)] 
--disable-optimizations option to force compiler optimizations (including
inlining) to be disabled. Very useful for debugging.

22 years agoBugfix: Speed up processing of queued events significantly, to not cause
hno [Sun, 6 Jul 2003 20:37:45 +0000 (20:37 +0000)] 
Bugfix: Speed up processing of queued events significantly, to not cause
large delays when under low load.

22 years agoBugfix: Speed up processing of queued events significantly, to not cause
hno [Sun, 6 Jul 2003 20:16:56 +0000 (20:16 +0000)] 
Bugfix: Speed up processing of queued events significantly, to not cause
large delays when under low load.

22 years agoclass DeferredRead is not a cbdata enabled class. Don't attempt to define
hno [Sun, 6 Jul 2003 19:43:40 +0000 (19:43 +0000)] 
class DeferredRead is not a cbdata enabled class. Don't attempt to define
the cbdata glue as this confuses GCC when optimization is not enabled

22 years agoImprovement: Increased time granularity for cache.log, down to msec
hno [Sun, 6 Jul 2003 18:12:28 +0000 (18:12 +0000)] 
Improvement: Increased time granularity for cache.log, down to msec
precision for debug messages (level > 1).

22 years agoBugfix: Make sure the clock is updated after poll/select to account for
hno [Sun, 6 Jul 2003 18:03:40 +0000 (18:03 +0000)] 
Bugfix: Make sure the clock is updated after poll/select to account for
the time sleeping waiting for I/O.

22 years agoRemoved 'minimum_retry_timeout' and Config.retry.timeout. This
wessels [Wed, 2 Jul 2003 02:42:27 +0000 (02:42 +0000)] 
Removed 'minimum_retry_timeout' and Config.retry.timeout.  This
variable has not been used for some time it seems.

22 years agoBootstrapped
hno [Sat, 28 Jun 2003 06:13:09 +0000 (06:13 +0000)] 
Bootstrapped

22 years ago* ext_user acl type to match user name returned by external acl
hno [Sat, 28 Jun 2003 04:42:08 +0000 (04:42 +0000)] 
* ext_user acl type to match user name returned by external acl

* cleanup of how external acls present a user name to the
  Squid core to make the code more logical and fix a
  minor security issues if there is downstream proxies.

* concept of password returned by external acl type. Integrated
  with login= cache_peer option to have the password forwarded
  to peers (both proxies and origin type peers)

22 years ago* ext_user acl type to match user name returned by external acl
hno [Sat, 28 Jun 2003 04:32:30 +0000 (04:32 +0000)] 
* ext_user acl type to match user name returned by external acl

* cleanup of how external acls present a user name to the
  Squid core to make the code more logical and fix a
  minor security issues if there is downstream proxies.

* concept of password returned by external acl type. Integrated
  with login= cache_peer option to have the password forwarded
  to peers (both proxies and origin type peers)

22 years agoMinor refactoring to C++, and renamed error to message in preparation
hno [Sat, 28 Jun 2003 02:54:45 +0000 (02:54 +0000)] 
Minor refactoring to C++, and renamed error to message in preparation
of later merge.

22 years agoSummary: Convert mem_hdr to the use of a splay.
robertc [Thu, 26 Jun 2003 18:51:57 +0000 (18:51 +0000)] 
Summary: Convert mem_hdr to the use of a splay.
Keywords:

* Add a dataRange() method to mem_node, to allow direct access to the range
  available.
* Add mem_node::deleteSelf().
* Update copyright on stmem.[cc|h].
* Test new mem_node methods.
* Add Splay::end() method and test.
* Add mem_hdr::unlink().
* Remove mem_hdr::unlinkHead(), mem_hdr::getHighest.., mem_hdr::head and mem_hdr::tail.
* Remove mem_node::next.
* Introduce mem_hdr::dump().
* Introduce mem_hdr::size().
* Introduce mem_hdr::start().
* Make mem_hdr::getBlockContainingLocation() public.

22 years agoBootstrapped
hno [Wed, 25 Jun 2003 06:13:03 +0000 (06:13 +0000)] 
Bootstrapped

22 years agochanging type of cacheMesh.cachePeerTable.cachePeerEntry.cachePeerPingsSent
wessels [Wed, 25 Jun 2003 03:08:02 +0000 (03:08 +0000)] 
changing type of cacheMesh.cachePeerTable.cachePeerEntry.cachePeerPingsSent
and cachePeerPingsAcked to match the MIB.  Was ASN_INTEGER, is not
SMI_COUNTER32.

22 years agoIn parseEtcHosts(), using strtok() to strip comments does not work if
wessels [Wed, 25 Jun 2003 02:53:39 +0000 (02:53 +0000)] 
In parseEtcHosts(), using strtok() to strip comments does not work if
the first character is a comment.  The buf[0] check is still necessary.

22 years agoSummary: Add missing profile counter.
robertc [Wed, 25 Jun 2003 01:49:58 +0000 (01:49 +0000)] 
Summary: Add missing profile counter.
Keywords:

Oops, managed to leave out one of the coutner declarations.

22 years agoSummary: Profiling additions.
robertc [Tue, 24 Jun 2003 18:42:25 +0000 (18:42 +0000)] 
Summary: Profiling additions.
Keywords:

* Added an eventRun profiler hook.
* Added a store_client profiler hook, surrounding kick_reads.
  Prevented reentrancy.
* Added a storeDirCallback profiler hook.
  Refactored to reduce variable scope.
* Added a profiler hook to comm_calliocallback.
  Refactored to reduce variable scope.
* Added surrounding hooks to the read,accept,write and fill
  comm callback callers.
* Added profiling hook around HttpStateData::readReply.
* Added profiling hook around HttpStateData::processReplyData.
* Added profiling hook around StoreEntry::write.
* Added profiling hook around storeGetMemSpace.
* Added profiling hook around MemObject::write.
* Added profiling hook around storeWriteComplete.
* Added profiling hook around mem_hdr::write.

22 years agoSummary: mem_hdr testing and corrections.
robertc [Tue, 24 Jun 2003 18:30:59 +0000 (18:30 +0000)] 
Summary: mem_hdr testing and corrections.
Keywords:

* Introduce operator < for mem_nodes, to provide ordering.
* Create a mem_hdr constructor and destructor.
* bugfix: writing before the lowest offset was broken.
* Reuse the Range template for mem_hdr::hasContigousContentRange.
* Prepare splay trees for replacing the list of mem_nodes.

22 years agoSummary: Add TODO for mem_node_test.
robertc [Tue, 24 Jun 2003 02:53:47 +0000 (02:53 +0000)] 
Summary: Add TODO for mem_node_test.
Keywords:

When MemPools are off, mem_node always returns 0 for the InUseCount. This leads to store.cc never freeing resources.

22 years agoRegarding store_check_cachable_stats, put check for 'wrong_content_length'
wessels [Mon, 23 Jun 2003 22:49:13 +0000 (22:49 +0000)] 
Regarding store_check_cachable_stats, put check for 'wrong_content_length'
before 'not_entry_cachable' because objects with ENTRY_BAD_LEGNTH
also always have ENTRY_CACHABLE cleared and nothing ever gets counted
as 'wrong_content_length'.

22 years agoSummary: comm cleanups.
robertc [Mon, 23 Jun 2003 20:13:03 +0000 (20:13 +0000)] 
Summary: comm cleanups.
Keywords:

* Add a FIXME comment in disabled client_side code where fd abstraction is broken.
* fde.rwstate is actually specific to writes - rename to wstate.
* The warning about rwstate should be a fatal, as it leads to incorrect callbacks occuring.
  disabling persistent connections will disable this bug - so a fatal is no impediment to testers.

22 years agoSummary: Refactoring: Extract method in comm.cc
robertc [Mon, 23 Jun 2003 18:27:12 +0000 (18:27 +0000)] 
Summary: Refactoring: Extract method in comm.cc
Keywords:

Extract common asserts to ensureOpenAndActive.

22 years agoSummary: strcpy->strncpy in pconn.cc.
robertc [Mon, 23 Jun 2003 18:11:45 +0000 (18:11 +0000)] 
Summary: strcpy->strncpy in pconn.cc.
Keywords:

Whilst debugging, I noticed we don't use strncpy - converted for safety.

22 years agoSummary: CBData reference callback pointers in comm_read and related..
robertc [Mon, 23 Jun 2003 18:03:07 +0000 (18:03 +0000)] 
Summary: CBData reference callback pointers in comm_read and related..
Keywords:

We were not cbdata referencing user supplied pointers. This allows user supplied data to be freed before we call back. Still TODO: check for cbdataReferenceValid upon callback.

22 years agoSummary: Fix segv in RefCount operator=.
robertc [Mon, 23 Jun 2003 17:14:52 +0000 (17:14 +0000)] 
Summary: Fix segv in RefCount operator=.
Keywords:

When a cycle is broken for refcounted objects, if the breaking smart pointer is in one of the cycled objects, and is freed as a result, we could end up assigned to freed RAM.

22 years agoSummary: Callback event loop was busted.
robertc [Mon, 23 Jun 2003 16:39:52 +0000 (16:39 +0000)] 
Summary: Callback event loop was busted.
Keywords:

The comms callback event loop was not dereferencing the dlink_node correctly, leading to comparing the sequence number to a pointer. Once again, typecasts leads to bugs :[.
* Increment sequence number on each loop.
* Compare to the event's record sequence number.
* Add a virtual destructor to CommCallbackData to remove some (spurious) warnings that apply to the synthetic constructor under -some- gcc's.

22 years agoSummary: Fix race condition in storeUnlockObject.
robertc [Mon, 23 Jun 2003 16:07:36 +0000 (16:07 +0000)] 
Summary: Fix race condition in storeUnlockObject.
Keywords:

storeUnlockObject would, under rare circumstances free the object it is unlocking, and -then- attempt to reference it.

22 years agoFrom Reuben Farrelly:
hno [Mon, 23 Jun 2003 14:38:39 +0000 (14:38 +0000)] 
From Reuben Farrelly:

A while ago we introduced the information line in squid that stated the
type of connections being accepted, like this:

2003/06/23 19:14:49| Accepting  transparently proxiedHTTP connections at
192.168.0.2, port 3128, FD 12.

Here's a patch (attached) to squid-3 CVS as of now, which fixes up a minor
problem with whitespaces - as illustrated above.

Reuben

22 years agoparseEtcHosts() does not properly handle comments. It looks for comments
wessels [Mon, 23 Jun 2003 06:41:31 +0000 (06:41 +0000)] 
parseEtcHosts() does not properly handle comments.  It looks for comments
(#) only at the beginning of the line, but they can appear anywhere.
This fix uses strtok() to terminate the buffer at the first comment
character.

22 years agoBootstrapped
hno [Mon, 23 Jun 2003 06:13:14 +0000 (06:13 +0000)] 
Bootstrapped

22 years agouse sbrk() for high_memory_warning check if mallinfo() and mstats()
wessels [Mon, 23 Jun 2003 03:31:24 +0000 (03:31 +0000)] 
use sbrk() for high_memory_warning check if mallinfo() and mstats()
are not available.

22 years agoSummary: Fix more client_side races with store client callbacks.
robertc [Sun, 22 Jun 2003 15:30:00 +0000 (15:30 +0000)] 
Summary: Fix more client_side races with store client callbacks.
Keywords:

Fix more client_side races with store client callbacks.

22 years agoSummary: Add a mempools testcase.
robertc [Sun, 22 Jun 2003 14:04:28 +0000 (14:04 +0000)] 
Summary: Add a mempools testcase.
Keywords:

Add a mempools testcase.

22 years agoSummary: Ensure that pconns start with no recorded fd's.
robertc [Sun, 22 Jun 2003 13:47:43 +0000 (13:47 +0000)] 
Summary: Ensure that pconns start with no recorded fd's.
Keywords:

Ensure that pconns start with no recorded fd's. Technically, this isn't needed, as cbdata is meant to return cleared memory, but it's good practice to explicitly initialise required values. (And, as we get more C++ syntax, the compiler will optimise that as part of operator new).

22 years agoBootstrapped
hno [Sat, 21 Jun 2003 06:13:07 +0000 (06:13 +0000)] 
Bootstrapped

22 years agoSummary: Some build options led to failures..
robertc [Fri, 20 Jun 2003 18:36:35 +0000 (18:36 +0000)] 
Summary: Some build options led to failures..
Keywords:

* client_side_request.h - include the needed headers for client_side_request.cci if inline is on.

22 years agoSummary: Encapsulate ClientHttpRequest::entry.
robertc [Fri, 20 Jun 2003 07:00:57 +0000 (07:00 +0000)] 
Summary: Encapsulate ClientHttpRequest::entry.
Keywords:

* Add accessor and mutator methods storeEntry().
* Find all mutations, and convert to using the mutator.
* Bugfix libTrie for --disable-inline builds.
* Make ClientHttpRequest::memObject() const inline.
* Introduce removeClientStoreReference to clientReplyContext to
  manage the encapsulation.
* Add needed MemObject.h header includes to gopher, ftp and wais.cc.

22 years agoBootstrapped
hno [Fri, 20 Jun 2003 06:13:08 +0000 (06:13 +0000)] 
Bootstrapped

22 years agoReshuffling of prototypes to make GCC even more happy
hno [Fri, 20 Jun 2003 06:05:11 +0000 (06:05 +0000)] 
Reshuffling of prototypes to make GCC even more happy

22 years agoSummary: Fixup GNUregex.c to meet current gcc warning checks.
robertc [Fri, 20 Jun 2003 04:25:18 +0000 (04:25 +0000)] 
Summary: Fixup GNUregex.c to meet current gcc warning checks.
Keywords:

Fixup GNUregex.c to meet current gcc warning checks. Mostly K&R->ANSI protoypes, as well as a const correctness issue exposed by the correct prototypes;

22 years agoneighbor_type_domain usage incorrect; missing neighbor hostname
wessels [Fri, 20 Jun 2003 03:28:50 +0000 (03:28 +0000)] 
neighbor_type_domain usage incorrect; missing neighbor hostname

22 years agoMerge conflict
hno [Fri, 20 Jun 2003 01:56:11 +0000 (01:56 +0000)] 
Merge conflict

22 years agoBug: MemPool.c is C source, not C++. C does not allow variables to be
hno [Fri, 20 Jun 2003 01:53:15 +0000 (01:53 +0000)] 
Bug: MemPool.c is C source, not C++. C does not allow variables to be
declared in the middle of a block.

22 years ago2003-06-17 09:18 hno
hno [Fri, 20 Jun 2003 01:07:21 +0000 (01:07 +0000)] 
2003-06-17 09:18  hno

Bug #683: external_acl does not wait for ident lookups to complete

22 years ago2003-05-21 01:51 hno
hno [Fri, 20 Jun 2003 01:02:33 +0000 (01:02 +0000)] 
2003-05-21 01:51  hno

Bug #644: strListGetItem() failed to account for quoting, causing digest
authentication to fail on URLs with comma and possibly other odd HTTP
parsing errors.

22 years ago2003-05-18 23:49 hno
hno [Fri, 20 Jun 2003 00:56:59 +0000 (00:56 +0000)] 
2003-05-18 23:49  hno
2003-05-24 13:08  robertc

Bug #630: digest authentication and buggy browsers

This patch does four things:

* correct signalling of stale digest nonces

* auth_param digest check-nonce-count off option to completely
disable the nonce count check. Needed to work around Konqueror and
certain other broken digest qop implementations.

* auth_param digest post_workaround on to work around certain
broken browsers sending POST requests with a GET digest response

* changes the default of nonce_strictness to off. It is to early to
use strict nonce count increase by one checks, and even in the long
run this is probably not wanted. Enforcing increase by one does not
really add to security, only complexity.

22 years ago2003-05-13 10:18 hno
hno [Fri, 20 Jun 2003 00:21:59 +0000 (00:21 +0000)] 
2003-05-13 10:18  hno

Correct handling of Squid-2.5 escaped usernames and passwords

22 years ago2003-05-11 23:44 hno
hno [Fri, 20 Jun 2003 00:20:51 +0000 (00:20 +0000)] 
2003-05-11 23:44  hno

Bug #591: segfault in authenticateAuthenticate() if debugging enabled

22 years ago2003-05-11 19:30 hno
hno [Fri, 20 Jun 2003 00:18:46 +0000 (00:18 +0000)] 
2003-05-11 19:30  hno

Adds a cbdata fence to clientReadBody() to make sure there is no
instabilities in PUT/POST due to aborted or failed requests.

(the same Squid-2.5 commit also includes cosmetic signed/unsigned cleanup
not committed here)

22 years ago2003-02-26 22:03 hno
hno [Fri, 20 Jun 2003 00:05:11 +0000 (00:05 +0000)] 
2003-02-26 22:03  hno

Bug #552: using username from external_acl for http_access and delay_access

22 years ago2003-05-29 04:59 wessels
hno [Thu, 19 Jun 2003 23:34:10 +0000 (23:34 +0000)] 
2003-05-29 04:59  wessels

Remove confusing "cache_dir diskd" line in cache_dir documentation.
It is apparently left over from a version that is quite old by now.
The correct usage for diskd is given later in the section.

22 years ago2003-05-25 15:47 hno
hno [Thu, 19 Jun 2003 23:32:53 +0000 (23:32 +0000)] 
2003-05-25 15:47  hno

Dist Lithuanian error pages

22 years ago2003-04-02 06:39 wessels
hno [Thu, 19 Jun 2003 23:25:53 +0000 (23:25 +0000)] 
2003-04-02 06:39  wessels

document 'htcp' and 'carp-load-factor' options to cache_peer.

(carp-load-factor does not exists in Squid-3)

22 years ago2003-01-09 06:04 robertc
hno [Thu, 19 Jun 2003 22:42:40 +0000 (22:42 +0000)] 
2003-01-09 06:04  robertc

Require _REENTRANT when compiing aufs

22 years agoBugfix for accel mode: Should not reuse persistent connections for
hno [Thu, 19 Jun 2003 19:47:25 +0000 (19:47 +0000)] 
Bugfix for accel mode: Should not reuse persistent connections for
different domains on the same origin server peer.

22 years agoBugfix in memAllocString to not crash and burn if alignment requirements
hno [Thu, 19 Jun 2003 19:27:50 +0000 (19:27 +0000)] 
Bugfix in memAllocString to not crash and burn if alignment requirements
make the actual allocated size different than expected (IA64)

22 years agoSummary: Gcc-3.3 tidyups.
robertc [Thu, 19 Jun 2003 19:11:59 +0000 (19:11 +0000)] 
Summary: Gcc-3.3 tidyups.
Keywords:

* Gave lib/ the same AM CXX and CFLAGS as src/
* In mempool, extract method on the free cache cleanup, and use a more correct type for Free, to avoid a warning on our use of 'type punned' pointers.
* Make squid_rn_inithead typesafe, and adjust usage throughout.
* rfc1738_do_escape needed adjustment for the use of char comparison with 0xFF.
* Make ACL Request/Reply HeaderStrategy singleton instance variable a pointer rather than instance to allow for stricter scoping rules.
* VectorMap<Key,Value>::indexUsed (unsigned char const index) had a tautology assert.

22 years agoSummary: Fix bug #685.
robertc [Wed, 18 Jun 2003 19:10:12 +0000 (19:10 +0000)] 
Summary: Fix bug #685.
Keywords:

Fix bug #685.

22 years agoSummary: Apply Leeann Bent's patch for bug #647.
robertc [Wed, 18 Jun 2003 18:34:51 +0000 (18:34 +0000)] 
Summary: Apply Leeann Bent's patch for bug #647.
Keywords:

Apply Leeann Bent's patch for bug #647.
Added Leeann as a contributor.

22 years agoSummary: Really fix bug #646
robertc [Wed, 18 Jun 2003 18:24:02 +0000 (18:24 +0000)] 
Summary: Really fix bug #646
Keywords:

Really fix it this time.

22 years agosome HTCP-related items
wessels [Mon, 16 Jun 2003 03:52:14 +0000 (03:52 +0000)] 
some HTCP-related items

22 years agoBugzilla #655: icmpRecv: recv: (11) Resource temporarily unavailable
wessels [Fri, 13 Jun 2003 05:56:38 +0000 (05:56 +0000)] 
Bugzilla #655: icmpRecv: recv: (11) Resource temporarily unavailable

22 years agoBugzilla #655: icmpRecv: recv: (11) Resource temporarily unavailable
wessels [Fri, 13 Jun 2003 05:54:17 +0000 (05:54 +0000)] 
Bugzilla #655: icmpRecv: recv: (11) Resource temporarily unavailable

Handle the case when recv() returns EAGAIN and do not treat it
like an error.

22 years agocorrection to comments. RFC 2396 (not 2616) talks about dealing with
wessels [Wed, 11 Jun 2003 02:45:29 +0000 (02:45 +0000)] 
correction to comments.  RFC 2396 (not 2616) talks about dealing with
whitespace in URIs.

22 years agonote my recent changes
wessels [Tue, 10 Jun 2003 04:06:24 +0000 (04:06 +0000)] 
note my recent changes

22 years agolog_quote() and username_quote() should always quote the '%' character
wessels [Tue, 10 Jun 2003 03:52:21 +0000 (03:52 +0000)] 
log_quote() and username_quote() should always quote the '%' character
since it is used to encode other characters.

22 years agoSummary: Update TODO.
robertc [Mon, 9 Jun 2003 13:23:53 +0000 (13:23 +0000)] 
Summary: Update TODO.
Keywords:

Update TODO.

22 years agoboostrapped
robertc [Mon, 9 Jun 2003 13:22:48 +0000 (13:22 +0000)] 
boostrapped

22 years agoSummary: fix 'not an lvalue' in HttpHdrSc.cc.
robertc [Mon, 9 Jun 2003 11:27:03 +0000 (11:27 +0000)] 
Summary: fix 'not an lvalue' in HttpHdrSc.cc.
Keywords:

fix 'not an lvalue' in HttpHdrSc.cc.

22 years agoSummary: const correctness in ESIExpression:getsymbol().
robertc [Mon, 9 Jun 2003 11:22:33 +0000 (11:22 +0000)] 
Summary: const correctness in ESIExpression:getsymbol().
Keywords:

const correctness in ESIExpression:getsymbol().

22 years agoSummary: Remove trailing , in ESICustomerParser.h enums.
robertc [Mon, 9 Jun 2003 11:12:04 +0000 (11:12 +0000)] 
Summary: Remove trailing , in ESICustomerParser.h enums.
Keywords:

Remove trailing , in ESICustomerParser.h enums.

22 years agoSummary: parseConfigFile had a const correctness flaw.
robertc [Mon, 9 Jun 2003 11:09:34 +0000 (11:09 +0000)] 
Summary: parseConfigFile had a const correctness flaw.
Keywords:

Summary: parseConfigFile had a const correctness flaw. It used a char * pointer to handle the result of a strrchr, which on some platforms leads to compiler warnings. Refactored to avoid this.

22 years agoSummary: const correctness in ftpListParseParts.
robertc [Mon, 9 Jun 2003 10:51:10 +0000 (10:51 +0000)] 
Summary: const correctness in ftpListParseParts.
Keywords:

ftpListParseParts abused a char * pointer to index into a const char * string - fixed.

22 years agoSummary: Refactor urlHostname.
robertc [Mon, 9 Jun 2003 10:41:36 +0000 (10:41 +0000)] 
Summary: Refactor urlHostname.
Keywords:

urlHostname reused the same variable for both constant and non-constant operations. This leads to warnings on some compilers. Refactored urlHostname to be a class, allowing greater flexability, and to use appropriate const corrent variables for accessing strchr() results.

22 years agoSummary: Make the test suite use the same flags as the main source.
robertc [Mon, 9 Jun 2003 10:02:16 +0000 (10:02 +0000)] 
Summary: Make the test suite use the same flags as the main source.
Keywords:

Make the test suite use the same flags as the main source.

22 years agoSummary: Cleanup splay template methods.
robertc [Mon, 9 Jun 2003 09:39:08 +0000 (09:39 +0000)] 
Summary: Cleanup splay template methods.
Keywords:

include/splay.h used the same name in parameters as the nodes do - which potentially could lead to confusion on the programmers part.

22 years agobootstrapped
robertc [Mon, 9 Jun 2003 09:24:03 +0000 (09:24 +0000)] 
bootstrapped

22 years agoSummary: Range used a not-quite standard syntax.
robertc [Mon, 9 Jun 2003 09:10:00 +0000 (09:10 +0000)] 
Summary: Range used a not-quite standard syntax.
Keywords:

Fixup Range.h for better portability.

22 years agoSummary: Fix bug #668
robertc [Mon, 9 Jun 2003 08:52:25 +0000 (08:52 +0000)] 
Summary: Fix bug #668
Keywords:

Fix bug #668 - remove -Wmissing-declarations from CXXFlags.

22 years agocheck open("/dev/null") return value for errors. Especially important
wessels [Mon, 9 Jun 2003 05:27:50 +0000 (05:27 +0000)] 
check open("/dev/null") return value for errors.  Especially important
later in ipcCreate() where Squid stays in a do loop if it cannot
open /dev/null.

22 years agoBootstrapped
hno [Mon, 2 Jun 2003 06:13:02 +0000 (06:13 +0000)] 
Bootstrapped

22 years agoThe proper type for fd_mask may not be detected by configure if sys/select.h
hno [Sun, 1 Jun 2003 15:01:41 +0000 (15:01 +0000)] 
The proper type for fd_mask may not be detected by configure if sys/select.h
is not listed in SQUID_DEFAULT_INCLUDES. (POSIX)

22 years ago.cvsignore updates to keep cvs silent
hno [Sun, 1 Jun 2003 14:51:15 +0000 (14:51 +0000)] 
.cvsignore updates to keep cvs silent

22 years agoBootstrapped
hno [Sat, 31 May 2003 17:34:59 +0000 (17:34 +0000)] 
Bootstrapped

22 years agoSpelling error in crypt_md5.h filename prevending make dist from working
hno [Sat, 31 May 2003 17:31:46 +0000 (17:31 +0000)] 
Spelling error in crypt_md5.h filename prevending make dist from working

22 years agoOverlapping helper requests for standard helpers (not stateful)
hno [Thu, 29 May 2003 21:54:07 +0000 (21:54 +0000)] 
Overlapping helper requests for standard helpers (not stateful)

 - basic auth
 - redirectors
 - external acls

22 years agoBootstrapped
hno [Thu, 29 May 2003 06:13:12 +0000 (06:13 +0000)] 
Bootstrapped

22 years agoDocumented the MD5 support in ncsa_auth
hno [Thu, 29 May 2003 04:45:12 +0000 (04:45 +0000)] 
Documented the MD5 support in ncsa_auth

22 years agoImported MD5 hash support for NCSA auth, supporting both shadow and
hno [Thu, 29 May 2003 04:42:32 +0000 (04:42 +0000)] 
Imported MD5 hash support for NCSA auth, supporting both shadow and
Apache htpasswd MD5 hases (same algorithms, different permutation strings)

22 years agoUnified the bootstrap.sh scripts to make sure there is no mismatches
hno [Thu, 29 May 2003 04:36:46 +0000 (04:36 +0000)] 
Unified the bootstrap.sh scripts to make sure there is no mismatches
in autotool versions between the main sources and libTrie. As it was
the version used in libTrie was very old and failed to find compatible
versions if not the system default..

22 years agoClean up if "Setting maxBodySize to.." debug message to show correct context
hno [Mon, 26 May 2003 02:29:24 +0000 (02:29 +0000)] 
Clean up if "Setting maxBodySize to.." debug message to show correct context

22 years agoSummary: Fix PURGE.
robertc [Sun, 25 May 2003 08:24:05 +0000 (08:24 +0000)] 
Summary: Fix PURGE.
Keywords:

Purging a nonexistant object would attempt to release a null store entry.

22 years agoSummary: Fix bug #646
robertc [Sat, 24 May 2003 18:43:30 +0000 (18:43 +0000)] 
Summary: Fix bug #646
Keywords:

Tail requests when there is any range offset limit should not result in full object retrieval.

22 years agoSummary: Fix bug #632 (Range reply Status code)
robertc [Sat, 24 May 2003 18:09:03 +0000 (18:09 +0000)] 
Summary: Fix bug #632 (Range reply Status code)
Keywords:

We need to set the reply status code to Partial content if we are generating a range reply..

22 years agoSummary: Bugfix: #651
robertc [Sat, 24 May 2003 17:32:34 +0000 (17:32 +0000)] 
Summary: Bugfix: #651
Keywords:

Quote the Surrogate-Capability capability statement as per the draft spec.