]> git.ipfire.org Git - thirdparty/squid.git/log
thirdparty/squid.git
17 years agoBootstrapped SQUID_3_0_PRE4
hno [Wed, 21 Jun 2006 02:58:18 +0000 (02:58 +0000)] 
Bootstrapped

17 years agoFix version number syntax to comply with out release plan
hno [Wed, 21 Jun 2006 02:57:25 +0000 (02:57 +0000)] 
Fix version number syntax to comply with out release plan

17 years agoBootstrapped
hno [Wed, 21 Jun 2006 02:52:42 +0000 (02:52 +0000)] 
Bootstrapped

17 years agoPreparing 3.0.PRE4
hno [Wed, 21 Jun 2006 02:51:11 +0000 (02:51 +0000)] 
Preparing 3.0.PRE4

17 years agoSome overlooked cleanup
hno [Tue, 20 Jun 2006 04:52:58 +0000 (04:52 +0000)] 
Some overlooked cleanup

17 years agoBug #1952: assertion failed: forward.cc:268: "entry->store_status == STORE_PENDING"
hno [Tue, 20 Jun 2006 04:49:59 +0000 (04:49 +0000)] 
Bug #1952: assertion failed: forward.cc:268: "entry->store_status == STORE_PENDING"

was caused by missing handler for aborted requests.

patch by Gonzalo Arana

17 years ago3.0.PRE4 release notes by Doug Dixon (with some cosmetic cleanups)
hno [Tue, 20 Jun 2006 04:45:48 +0000 (04:45 +0000)] 
3.0.PRE4 release notes by Doug Dixon (with some cosmetic cleanups)

17 years agoMS Visual Studio C++ needs to declare global variables
serassio [Sun, 18 Jun 2006 16:05:53 +0000 (16:05 +0000)] 
MS Visual Studio C++ needs to declare global variables
as extern "C".

17 years agoFixed Visual Studio build errors
serassio [Sun, 18 Jun 2006 14:56:32 +0000 (14:56 +0000)] 
Fixed Visual Studio build errors

17 years agoAdded some missing error messages
serassio [Sun, 18 Jun 2006 14:27:50 +0000 (14:27 +0000)] 
Added some missing error messages

17 years agoBootstrapped
hno [Sun, 18 Jun 2006 06:10:52 +0000 (06:10 +0000)] 
Bootstrapped

17 years agoBootstrapped
serassio [Sun, 18 Jun 2006 03:51:31 +0000 (03:51 +0000)] 
Bootstrapped

17 years agoCosmetic update to AC_INIT syntax like 2.6
serassio [Sun, 18 Jun 2006 03:49:48 +0000 (03:49 +0000)] 
Cosmetic update to AC_INIT syntax like 2.6

17 years agoCosmetic: updated some debug messages
serassio [Sat, 17 Jun 2006 22:36:15 +0000 (22:36 +0000)] 
Cosmetic: updated some debug messages

17 years agoBootstrapped
hno [Thu, 15 Jun 2006 06:11:08 +0000 (06:11 +0000)] 
Bootstrapped

17 years agoBug #1593: multiline configs garbled in cache manager
serassio [Thu, 15 Jun 2006 01:52:54 +0000 (01:52 +0000)] 
Bug #1593: multiline configs garbled in cache manager

Patch provided by Gonzalo Arana.

17 years agoBug 1594: ftp-triggered: assertion failed: comm.cc:1438: "fd >= 0"
serassio [Thu, 15 Jun 2006 01:43:28 +0000 (01:43 +0000)] 
Bug 1594: ftp-triggered: assertion failed: comm.cc:1438: "fd >= 0"

Patch provided by Gonzalo Arana.

17 years agoBug #1587: assertion failed on invalid reply
serassio [Thu, 15 Jun 2006 01:39:19 +0000 (01:39 +0000)] 
Bug #1587: assertion failed on invalid reply

Patch provided by Gonzalo Arana.

17 years agoBug #1604: segfault on cachemgr when no redirectors defined
serassio [Thu, 15 Jun 2006 01:29:30 +0000 (01:29 +0000)] 
Bug #1604: segfault on cachemgr when no redirectors defined

Patch provided by Gonzalo Arana.

17 years agoBootstrapped
serassio [Thu, 15 Jun 2006 01:19:23 +0000 (01:19 +0000)] 
Bootstrapped

17 years agoBug #961: Add support for arbitrary headers access controls
serassio [Thu, 15 Jun 2006 01:18:24 +0000 (01:18 +0000)] 
Bug #961: Add support for arbitrary headers access controls

Forward port of 2.5 patch by Gonzalo Arana.

17 years agoBootstrapped
hno [Wed, 14 Jun 2006 06:10:54 +0000 (06:10 +0000)] 
Bootstrapped

17 years agoBootstrapped
serassio [Wed, 14 Jun 2006 02:56:26 +0000 (02:56 +0000)] 
Bootstrapped

17 years agoPartial Windows native port merge:
serassio [Wed, 14 Jun 2006 02:52:05 +0000 (02:52 +0000)] 
Partial Windows native port merge:

- Native Windows comm_select() implementation

17 years agoFix --disable-http-violations again.
hno [Wed, 14 Jun 2006 00:36:52 +0000 (00:36 +0000)] 
Fix --disable-http-violations again.

17 years agoBootstrapped
serassio [Tue, 13 Jun 2006 16:29:28 +0000 (16:29 +0000)] 
Bootstrapped

17 years agoGroup all comm_*.* sources into squid_COMMSOURCES
serassio [Tue, 13 Jun 2006 16:23:57 +0000 (16:23 +0000)] 
Group all comm_*.* sources into squid_COMMSOURCES

17 years agoMissing comm_*.h from make dist
serassio [Tue, 13 Jun 2006 16:11:31 +0000 (16:11 +0000)] 
Missing comm_*.h from make dist

17 years agoTypos in HEAD.
robertc [Mon, 12 Jun 2006 02:48:45 +0000 (02:48 +0000)] 
Typos in HEAD.

17 years agoBug #1490: GPL license issues with RSA MD5 implementation
hno [Sun, 11 Jun 2006 19:29:00 +0000 (19:29 +0000)] 
Bug #1490: GPL license issues with RSA MD5 implementation

replaced the Squid provided MD5 with another public domain copy not
encumbered by the RSA license.

17 years agoAccept (and ignore) the NTLM negotiate packet
hno [Sun, 11 Jun 2006 18:53:36 +0000 (18:53 +0000)] 
Accept (and ignore) the NTLM negotiate packet

17 years agoRename redirect_ to url_rewrite_ to match 2.6.
hno [Sun, 11 Jun 2006 16:13:12 +0000 (16:13 +0000)] 
Rename redirect_ to url_rewrite_ to match 2.6.

17 years agoEnable 2.5 snapshots still.. we are not quite done with 2.5 yet..
hno [Sun, 11 Jun 2006 04:31:14 +0000 (04:31 +0000)] 
Enable 2.5 snapshots still.. we are not quite done with 2.5 yet..

17 years agoBootstrapped
hno [Sat, 10 Jun 2006 06:10:50 +0000 (06:10 +0000)] 
Bootstrapped

17 years agoNeed to dist the comm_*.h files
hno [Fri, 9 Jun 2006 06:33:13 +0000 (06:33 +0000)] 
Need to dist the comm_*.h files

17 years agoBug #1548: Cleanup of 304 processing
hno [Thu, 8 Jun 2006 04:39:33 +0000 (04:39 +0000)] 
Bug #1548: Cleanup of 304 processing
by Doug Dixon <doug.dixon@gmail.com>

The main change is to the logging codes. Since all refreshes hit the origin, it
does not make any sense to refer to them as HIT or MISS. Instead, they have
been renamed to describe the result from the origin server:

* TCP_REFRESH_UNMODIFIED - The requested object was cached but STALE. The IMS
query for the object resulted in "304 not modified".

* TCP_REFRESH_FAIL - The requested object was cached but STALE. The IMS query
failed and the stale object was delivered.

* TCP_REFRESH_MODIFIED - The requested object was cached but STALE. The IMS
query returned the new content.

These actually fit the FAQ now (if you change the names):
http://www.squid-cache.org/Doc/FAQ/FAQ-6.html#ss6.7

Apart from that:

1. I've corrected the set of headers that should NOT be updated in an existing
cache entry on receipt of a 304 from the origin (i.e. hop-by-hop headers)
2. I've removed a couple more functions and put the code inline. It's easier to
follow.

only allowing the entity-headers specified as being cool for 304
(section 10.3.5) to update an entry:

Content-Location
Expires

All the other entity-headers are either a SHOULD NOT or MUST NOT for including
with a 304 anyway, so always ignore these when updating an entry:

Allow
Content-Encoding
Content-Language
Content-Length
Content-MD5
Content-Range
Content-Type
Last-Modified

17 years agoPrzemek Czerkas wrote many refresh_pattern override options
hno [Wed, 7 Jun 2006 19:00:24 +0000 (19:00 +0000)] 
Przemek Czerkas wrote many refresh_pattern override options
(Bug #1057)

17 years agoCorrect Cache-Control header checks and simplify the associated overrides
hno [Wed, 7 Jun 2006 18:32:56 +0000 (18:32 +0000)] 
Correct Cache-Control header checks and simplify the associated overrides

17 years agoBug #624: Various assertion errors on Range requests
hno [Wed, 7 Jun 2006 01:22:13 +0000 (01:22 +0000)] 
Bug #624: Various assertion errors on Range requests

use reply range spec if does not match request range spec (forward as-is)

This assertion is triggered if request range is not satisified by reply.  In
this case, we simply forward the reply as-is.
To forward as-is, we override client_side reply range spec with server_side
reply range spec.

if reply has ranges but request is not: we forward as-is.

Another situation where this assertion is triggered:  if reply is partial
content (status 206) but request is not.  This patch forward as-is the
offending response received on server_side.

17 years agoAdded some missing CONTRIBUTOR names
hno [Tue, 6 Jun 2006 14:07:12 +0000 (14:07 +0000)] 
Added some missing CONTRIBUTOR names

17 years agoBootstrapped
hno [Tue, 6 Jun 2006 06:10:49 +0000 (06:10 +0000)] 
Bootstrapped

17 years agoCheck if LOG_AUTH is defined before try to define LOG_AUTHPRIV
serassio [Tue, 6 Jun 2006 03:08:15 +0000 (03:08 +0000)] 
Check if LOG_AUTH is defined before try to define LOG_AUTHPRIV

17 years agoDefine LOG_AUTHPRIV as LOG_AUTH on systems still using the old deprecated
serassio [Tue, 6 Jun 2006 02:49:45 +0000 (02:49 +0000)] 
Define LOG_AUTHPRIV as LOG_AUTH on systems still using the old deprecated
LOG_AUTH. One is Irix.

17 years agoBug #1468: Crash on HttpHdrRange.cc line 568: assertion failed on "valid"
serassio [Tue, 6 Jun 2006 00:57:08 +0000 (00:57 +0000)] 
Bug #1468: Crash on HttpHdrRange.cc line 568: assertion failed on "valid"

Patch provided by Doug Dixon

17 years agoBug #1559: asnAddNet() might leak memory
serassio [Tue, 6 Jun 2006 00:35:02 +0000 (00:35 +0000)] 
Bug #1559: asnAddNet() might leak memory

Forward port of 2.6 patch.

17 years agoBootstrapped
serassio [Mon, 5 Jun 2006 22:07:57 +0000 (22:07 +0000)] 
Bootstrapped

17 years agoForgotten to fix CXXFLAGS on Irix 6.x
serassio [Mon, 5 Jun 2006 22:04:22 +0000 (22:04 +0000)] 
Forgotten to fix CXXFLAGS on Irix 6.x

17 years agoBootstrapped
hno [Mon, 5 Jun 2006 06:10:50 +0000 (06:10 +0000)] 
Bootstrapped

17 years agoBootstrapped
serassio [Mon, 5 Jun 2006 02:06:28 +0000 (02:06 +0000)] 
Bootstrapped

17 years agoBug #1536: Authentication unconditionally built into Squid
serassio [Mon, 5 Jun 2006 02:04:33 +0000 (02:04 +0000)] 
Bug #1536: Authentication unconditionally built into Squid

Only added an info message

17 years agoUpdated .cvsignore
serassio [Mon, 5 Jun 2006 00:54:56 +0000 (00:54 +0000)] 
Updated .cvsignore

17 years agoUpdated .cvsignore
serassio [Mon, 5 Jun 2006 00:48:03 +0000 (00:48 +0000)] 
Updated .cvsignore

17 years agoBootstrapped
serassio [Mon, 5 Jun 2006 00:47:02 +0000 (00:47 +0000)] 
Bootstrapped

17 years agotestURL needs referer.cc and useragent.cc
serassio [Mon, 5 Jun 2006 00:45:47 +0000 (00:45 +0000)] 
testURL needs referer.cc and useragent.cc

17 years agoForward port of Irix related changes from 2.6
serassio [Mon, 5 Jun 2006 00:43:34 +0000 (00:43 +0000)] 
Forward port of Irix related changes from 2.6

17 years agoFixed typo
serassio [Mon, 5 Jun 2006 00:36:34 +0000 (00:36 +0000)] 
Fixed typo

17 years agoAdded .cvsignore
serassio [Mon, 5 Jun 2006 00:35:06 +0000 (00:35 +0000)] 
Added .cvsignore

17 years agoBootstrapped
serassio [Mon, 5 Jun 2006 00:32:39 +0000 (00:32 +0000)] 
Bootstrapped

17 years agotest_http_range needs referer.cc and useragent.cc
serassio [Mon, 5 Jun 2006 00:31:37 +0000 (00:31 +0000)] 
test_http_range needs referer.cc and useragent.cc

17 years agoBootstrapped
hno [Sat, 3 Jun 2006 06:10:50 +0000 (06:10 +0000)] 
Bootstrapped

17 years agoBootstrapped
serassio [Fri, 2 Jun 2006 22:44:26 +0000 (22:44 +0000)] 
Bootstrapped

17 years agoAdded reference to squid_session man page
serassio [Fri, 2 Jun 2006 22:37:00 +0000 (22:37 +0000)] 
Added reference to squid_session man page

17 years agoAdded Windows info to squid man page
serassio [Fri, 2 Jun 2006 22:35:05 +0000 (22:35 +0000)] 
Added Windows info to squid man page

17 years agoDerive package version automatically from configure
serassio [Fri, 2 Jun 2006 22:31:45 +0000 (22:31 +0000)] 
Derive package version automatically from configure

17 years agoDon't make snapshots of Squid-HEAD. These are already made as 3.0
hno [Fri, 2 Jun 2006 14:16:03 +0000 (14:16 +0000)] 
Don't make snapshots of Squid-HEAD. These are already made as 3.0

17 years agoBug #1395: Add extra tracing to squid_ldap_auth in debug/verbose mode.
hno [Fri, 2 Jun 2006 06:38:23 +0000 (06:38 +0000)] 
Bug #1395: Add extra tracing to squid_ldap_auth in debug/verbose mode.
Correct example on how to connect to MS AD.

17 years agobug #1465: assertion failed: mem_node.cc:65: "n->write_pending"
wessels [Wed, 31 May 2006 23:47:53 +0000 (23:47 +0000)] 
bug #1465: assertion failed: mem_node.cc:65: "n->write_pending"

This assertion was traced to bugs in UFSStoreState code when using aufs
storage.  The primary problem was that UFSStoreState was aggressive
in closing swapout files.  They could be closed before any I/Os had
time to complete.  The assertion seemed to be caused by a "double
callback" whereby memNodeWriteComplete would get called twice.

Additionally, writes were executed out-of-order in cases where a
write was placed on the pending queue because open hadn't completed
yet.  A subsequent write would be given priority even though
the first write was in the pending queue.

Now all writes go into the pending queue.  This sucks a little bit,
especially for synchronous storage shcemes such as UFS and cases
where UFSStoreState gets one write request at a time and queueing
is unnecessary.

See additional comments in fs/ufs/store_io_ufs.cc dated 2006-05-24

17 years agomake Squid compile again with --disable-internal-dns after recent
wessels [Wed, 31 May 2006 23:29:00 +0000 (23:29 +0000)] 
make Squid compile again with --disable-internal-dns after recent
CacheManager changes.

17 years agocosmetic: a few updates/changes to debugging messages and assertions
wessels [Wed, 31 May 2006 23:25:41 +0000 (23:25 +0000)] 
cosmetic: a few updates/changes to debugging messages and assertions

17 years agominor: now that we're passing StoreIOState to these callbacks again,
wessels [Wed, 31 May 2006 23:24:23 +0000 (23:24 +0000)] 
minor: now that we're passing StoreIOState to these callbacks again,
we can replace the assertion that "mem->swapout.sio == sio".

17 years agocosmetic: don't need to use getRaw() when checking for NULL pointer
wessels [Wed, 31 May 2006 23:23:02 +0000 (23:23 +0000)] 
cosmetic: don't need to use getRaw() when checking for NULL pointer

17 years agoMake access_log.cc compile again with --enable-forw-via-db after
wessels [Wed, 31 May 2006 22:34:36 +0000 (22:34 +0000)] 
Make access_log.cc compile again with --enable-forw-via-db after
CacheManager changes.

17 years agoRemoved comm_accept_check_event() and related code since we have
wessels [Wed, 31 May 2006 03:15:58 +0000 (03:15 +0000)] 
Removed comm_accept_check_event() and related code since we have
the AcceptLimiter class which does a better job of deferring new
connections when we start to run out of FDs.

17 years agoBug #1554: HTCP RFC compliance
hno [Tue, 30 May 2006 23:31:23 +0000 (23:31 +0000)] 
Bug #1554: HTCP RFC compliance

Fix field order to comply with the RFC. Bump minor version to 1 to
make migration easier. New cache-peer option to support old Squid HTCP
implementaitons.

17 years agoBootstrapped
hno [Tue, 30 May 2006 06:10:53 +0000 (06:10 +0000)] 
Bootstrapped

17 years agoSuspend 2.5 snapshot releases, hopefully for good.
hno [Tue, 30 May 2006 05:08:25 +0000 (05:08 +0000)] 
Suspend 2.5 snapshot releases, hopefully for good.

17 years agosnapshot 2.6 releases
hno [Tue, 30 May 2006 05:07:26 +0000 (05:07 +0000)] 
snapshot 2.6 releases

17 years agoMove clientCachable logic onto HttpRequest, as all the used fields in the
robertc [Tue, 30 May 2006 03:44:18 +0000 (03:44 +0000)] 
Move clientCachable logic onto HttpRequest, as all the used fields in the
function where on the HttpRequest object.

Apply Gonzalo Aranas bugfix for the replacement to cachemgrRegister,
as having some options disabled/changed did not compile.

17 years agoRemove obsolete stub file.
robertc [Tue, 30 May 2006 03:40:10 +0000 (03:40 +0000)] 
Remove obsolete stub file.

17 years agoBug #1536: Finding out what word in acl list is blocked
serassio [Tue, 30 May 2006 01:05:26 +0000 (01:05 +0000)] 
Bug #1536: Finding out what word in acl list is blocked

17 years agoAdd a CacheManager class which provides the cachemanager menu registration facility...
robertc [Mon, 29 May 2006 06:14:59 +0000 (06:14 +0000)] 
Add a CacheManager class which provides the cachemanager menu registration facility, and remove all compile time dependencies upon cachemgrRegister from squid outside of main.cc. This allows much simpler tests, as modules exporting cache mgr information no longer drag in all of squid via cache_manager.cc.

17 years agoForward port of latest 2.6 changes
serassio [Sun, 28 May 2006 22:11:38 +0000 (22:11 +0000)] 
Forward port of latest 2.6 changes

17 years agoInclude getopt.h
serassio [Sun, 28 May 2006 22:04:59 +0000 (22:04 +0000)] 
Include getopt.h

17 years agoBootstrapped
hno [Sun, 28 May 2006 06:10:51 +0000 (06:10 +0000)] 
Bootstrapped

17 years agoUse ccpunit-config to get information about the system installed ccpunit
hno [Sat, 27 May 2006 06:58:15 +0000 (06:58 +0000)] 
Use ccpunit-config to get information about the system installed ccpunit
instead of hardwired probes

17 years agounbreak testStore testcase.
robertc [Sat, 27 May 2006 06:35:05 +0000 (06:35 +0000)] 
unbreak testStore testcase.

17 years agoFixed TestSwapDir test after I changed StoreIOState.
wessels [Sat, 27 May 2006 05:52:14 +0000 (05:52 +0000)] 
Fixed TestSwapDir test after I changed StoreIOState.

17 years agoFix that strange looking "Notice: Short Strings is 36 bytes instead of requested...
hno [Sat, 27 May 2006 05:43:07 +0000 (05:43 +0000)] 
Fix that strange looking "Notice: Short Strings is 36 bytes instead of requested 36 bytes"
message to say what it meant to say..

17 years agoThere is still code relying on fd_set so we must keep the FD_SETSIZE
hno [Sat, 27 May 2006 05:35:44 +0000 (05:35 +0000)] 
There is still code relying on fd_set so we must keep the FD_SETSIZE
filedescriptor limit for a while still...

17 years agoRename no_cache to cache to better reflect the logics of the directive
hno [Sat, 27 May 2006 04:33:24 +0000 (04:33 +0000)] 
Rename no_cache to cache to better reflect the logics of the directive
(no_cache kept as an alias)

17 years agocosmetic: The debug message "WARNING! Your cache is running out of
wessels [Sat, 27 May 2006 01:58:37 +0000 (01:58 +0000)] 
cosmetic: The debug message "WARNING! Your cache is running out of
filedescriptors" appears in two places.  I changed it to debugs()
macro with HERE argument to differentiate the two messages.

17 years agonitpick: This "Fc = &fdc_table[fd]" optimization seems like a waste here.
wessels [Sat, 27 May 2006 01:53:18 +0000 (01:53 +0000)] 
nitpick: This "Fc = &fdc_table[fd]" optimization seems like a waste here.

17 years agocosmetic change: rename doCallback to doCloseCallback to clarify
wessels [Wed, 24 May 2006 21:25:04 +0000 (21:25 +0000)] 
cosmetic change: rename doCallback to doCloseCallback to clarify
what this function does.

17 years agoRemove redundant newline in cache.log
hno [Wed, 24 May 2006 08:13:27 +0000 (08:13 +0000)] 
Remove redundant newline in cache.log

patch by Pawel Worach

17 years agoBootstrapped
hno [Wed, 24 May 2006 06:10:48 +0000 (06:10 +0000)] 
Bootstrapped

17 years agoBlock copy constructor. Not safe.
hno [Wed, 24 May 2006 02:30:48 +0000 (02:30 +0000)] 
Block copy constructor. Not safe.

Should also block the assignment constructor, but it is currently in use
(privately) as part of the reset method.

17 years agocleanup of how the Vary marker object is created
hno [Wed, 24 May 2006 02:29:04 +0000 (02:29 +0000)] 
cleanup of how the Vary marker object is created
to look more like a normal object creation, not violating layers too much

17 years agoFix NULL pointer access with asyncio when Squid starts up over the disk
wessels [Wed, 24 May 2006 00:24:41 +0000 (00:24 +0000)] 
Fix NULL pointer access with asyncio when Squid starts up over the disk
space limit.

We used to call squidaio_init() from DiskThreadsDiskFile::operator new.
But this only gets called when opening a file.  If the first transaction
is to unlink a file, squidaio will be uninitialized.

I didn't see any other good places to call squidaio_init().  For now
I am calling it from within the squidaio module itself at the
start of each open, unlink/truncate, and stat call.

17 years ago-mc command line option to disable memory pools
hno [Tue, 23 May 2006 22:24:55 +0000 (22:24 +0000)] 
-mc command line option to disable memory pools

17 years agoFix a memory corruption when building the magic Vary marker object headers
hno [Tue, 23 May 2006 22:22:34 +0000 (22:22 +0000)] 
Fix a memory corruption when building the magic Vary marker object headers

somehow this cast made things go very bad. Luckily it isn't needed as
we already hold a reference to the reply a few lines earlier.