]> git.ipfire.org Git - thirdparty/squid.git/log
thirdparty/squid.git
27 years agoadd snprintf.h
wessels [Mon, 5 Jan 1998 07:57:53 +0000 (07:57 +0000)] 
add snprintf.h

27 years agoupdate
wessels [Mon, 5 Jan 1998 07:46:59 +0000 (07:46 +0000)] 
update

27 years ago- Changed algorithm for determining alive/dead state of peers.
wessels [Mon, 5 Jan 1998 07:45:43 +0000 (07:45 +0000)] 
- Changed algorithm for determining alive/dead state of peers.
  Instead of using a fixed number of unacknowledged ICP replies,
  it is now based on timeouts.  If there are no ICP replies
  received from a peer within 'dead_peer_timeout' seconds, then
  we call it dead.
- Added calls to getCurrentTime() in comm_{select,poll}_incoming().
- Fixed shutdown bug when the incoming and outgoing ICP socket is
  the same file descriptor.

27 years agoFix memmove() arg bug. should be conn->in.offset, not in.size
wessels [Sun, 4 Jan 1998 13:57:27 +0000 (13:57 +0000)] 
Fix memmove() arg bug.   should be conn->in.offset, not in.size

27 years agomath precedence bugfix
wessels [Sun, 4 Jan 1998 12:43:48 +0000 (12:43 +0000)] 
math precedence bugfix

27 years agowhy is poll() not being skipped on OSF?
wessels [Sun, 4 Jan 1998 07:17:52 +0000 (07:17 +0000)] 
why is poll() not being skipped on OSF?

27 years agoWe were erroneously assuming that non-200 replies would have a content-length
wessels [Sat, 3 Jan 1998 12:27:17 +0000 (12:27 +0000)] 
We were erroneously assuming that non-200 replies would have a content-length
if there were a reply body.  RFC 2068 does not require this.  If there is
no content-length, then we can't have a persistent connection.

27 years agomake some cbdata function args const
wessels [Sat, 3 Jan 1998 06:41:02 +0000 (06:41 +0000)] 
make some cbdata function args const

27 years agoUse cbdataLock etc on the checklist->access_list structures so we
wessels [Sat, 3 Jan 1998 06:39:54 +0000 (06:39 +0000)] 
Use cbdataLock etc on the checklist->access_list structures so we
don't dump core during reconfigure.

27 years agono change, just a note that we do not need snprintf here
kostas [Sat, 3 Jan 1998 05:45:03 +0000 (05:45 +0000)] 
no change, just a note that we do not need snprintf here

27 years agocode cleanup+security (sprintf->snprintf)
kostas [Sat, 3 Jan 1998 05:39:14 +0000 (05:39 +0000)] 
code cleanup+security (sprintf->snprintf)

27 years agoadd ident tag
wessels [Sat, 3 Jan 1998 05:04:36 +0000 (05:04 +0000)] 
add ident tag

27 years ago - Fixed some dnsserver-related reconfigure bugs. Need to
wessels [Sat, 3 Jan 1998 05:03:40 +0000 (05:03 +0000)] 
    - Fixed some dnsserver-related reconfigure bugs. Need to
          use cbdataLock, etc in fqdncache.c.  Also don't want to
          use ipcacheQueueDrain() and fqdncacheQueueDrain().

27 years agoWe must make the ICP reply callback for neighbor MISSes!
wessels [Sat, 3 Jan 1998 04:59:47 +0000 (04:59 +0000)] 
We must make the ICP reply callback for neighbor MISSes!

27 years agoremove static netdbGetFirst() and netdbGetNext()
wessels [Sat, 3 Jan 1998 04:55:05 +0000 (04:55 +0000)] 
remove static netdbGetFirst() and netdbGetNext()

27 years agoasnFreeMemory calls asnCleanup for purify.
kostas [Sat, 3 Jan 1998 04:52:58 +0000 (04:52 +0000)] 
asnFreeMemory calls asnCleanup for purify.

27 years agouse PAGESIZE value if defined
wessels [Sat, 3 Jan 1998 04:52:40 +0000 (04:52 +0000)] 
use PAGESIZE value if defined

27 years agotest for struct rusage needs to include <sys/time.h
wessels [Sat, 3 Jan 1998 04:49:40 +0000 (04:49 +0000)] 
test for struct rusage needs to include <sys/time.h

27 years agouse strchr(), not index()
wessels [Sat, 3 Jan 1998 04:35:37 +0000 (04:35 +0000)] 
use strchr(), not index()

27 years agocheck for getpagesize()
wessels [Sat, 3 Jan 1998 04:35:13 +0000 (04:35 +0000)] 
check for getpagesize()

27 years agoupdate
wessels [Sat, 3 Jan 1998 02:56:38 +0000 (02:56 +0000)] 
update

27 years agoReset Config.npeers to zero in free_peer()
wessels [Sat, 3 Jan 1998 02:55:28 +0000 (02:55 +0000)] 
Reset Config.npeers to zero in free_peer()

27 years agoFix pointer bug when freeing ASN ACLs
wessels [Sat, 3 Jan 1998 02:52:23 +0000 (02:52 +0000)] 
Fix pointer bug when freeing ASN ACLs

27 years agofake rusage needs ru_utime
wessels [Sat, 3 Jan 1998 02:27:01 +0000 (02:27 +0000)] 
fake rusage needs ru_utime

27 years agoupdate
wessels [Sat, 3 Jan 1998 01:23:42 +0000 (01:23 +0000)] 
update

27 years agoupdate
wessels [Sat, 3 Jan 1998 01:17:50 +0000 (01:17 +0000)] 
update

27 years agoFrom: "Michael O'Reilly" <michael@metal.iinet.net.au>
wessels [Sat, 3 Jan 1998 01:15:32 +0000 (01:15 +0000)] 
From:    "Michael O'Reilly" <michael@metal.iinet.net.au>
here's a patch to fix it. (The threads library is closing FD 0 for
some reason on the fork(). It's a bug in the threads library, but this
is a work around.. )

27 years agoupdate
wessels [Sat, 3 Jan 1998 01:09:55 +0000 (01:09 +0000)] 
update

27 years agoFrom: "Michael O'Reilly" <michael@metal.iinet.net.au>
wessels [Sat, 3 Jan 1998 01:09:20 +0000 (01:09 +0000)] 
From:    "Michael O'Reilly" <michael@metal.iinet.net.au>
In clientRedirectDone(), xstrdup request->headers instead of copy by
reference.

27 years agoTail-chasing because struct rusage not defined on Solaris. The real
wessels [Sat, 3 Jan 1998 01:04:21 +0000 (01:04 +0000)] 
Tail-chasing because struct rusage not defined on Solaris.  The real
problem was 'typedef void struct' instead of 'typedef struct'.  However,
we also now have a configure test for 'struct rusage' instead of
just using HAVE_GETRUSAGE.

27 years agoupdate
wessels [Fri, 2 Jan 1998 23:38:45 +0000 (23:38 +0000)] 
update

27 years agoUse the short 50ms timeout in comm_poll() only when USE_ASYNC_IO
wessels [Fri, 2 Jan 1998 23:30:14 +0000 (23:30 +0000)] 
Use the short 50ms timeout in comm_poll() only when USE_ASYNC_IO
is defined

27 years agoWe must disable read handlers for OUTGOING UDP sockets during shutdown.
wessels [Fri, 2 Jan 1998 13:56:51 +0000 (13:56 +0000)] 
We must disable read handlers for OUTGOING UDP sockets during shutdown.

27 years agogindent mangles wierd #ifdefs
wessels [Fri, 2 Jan 1998 11:53:34 +0000 (11:53 +0000)] 
gindent mangles wierd #ifdefs

27 years agotypo
wessels [Fri, 2 Jan 1998 10:12:02 +0000 (10:12 +0000)] 
typo

27 years agoupdate
wessels [Fri, 2 Jan 1998 10:00:20 +0000 (10:00 +0000)] 
update

27 years agogindent
wessels [Fri, 2 Jan 1998 09:05:59 +0000 (09:05 +0000)] 
gindent

27 years agoSQUID_MIB not needed
wessels [Fri, 2 Jan 1998 09:05:19 +0000 (09:05 +0000)] 
SQUID_MIB not needed

27 years agomemory leaks
wessels [Fri, 2 Jan 1998 02:35:13 +0000 (02:35 +0000)] 
memory leaks

27 years ago - Fixed 4k page leak in icpHandleIMSReply();
wessels [Fri, 2 Jan 1998 02:29:17 +0000 (02:29 +0000)] 
    - Fixed 4k page leak in icpHandleIMSReply();
        - Renamed 'icp*' functions to 'client*' names in client_side.c.

27 years agopreprocessor
wessels [Fri, 2 Jan 1998 01:45:58 +0000 (01:45 +0000)] 
preprocessor

27 years agoalways memset struct rusage in squid_getrusage()
wessels [Fri, 2 Jan 1998 01:42:57 +0000 (01:42 +0000)] 
always memset struct rusage in squid_getrusage()

27 years agoFor unlinkd FD, set the timeout AFTER fd_open().
wessels [Thu, 1 Jan 1998 12:57:17 +0000 (12:57 +0000)] 
For unlinkd FD, set the timeout AFTER fd_open().

27 years ago- Fix bug from Henrik's patch. We can't set the FD timeout after
wessels [Thu, 1 Jan 1998 12:48:38 +0000 (12:48 +0000)] 
- Fix bug from Henrik's patch.  We can't set the FD timeout after
  calling clientReadRequest() because the FD might be closed by then.
- Add an assertion in commSetTimeout() that the FD be open.

27 years agoupdate
wessels [Thu, 1 Jan 1998 07:07:06 +0000 (07:07 +0000)] 
update

27 years agoFrom: Henrik Nordstrom <hno@hem.passagen.se>
wessels [Thu, 1 Jan 1998 07:05:52 +0000 (07:05 +0000)] 
From:    Henrik Nordstrom <hno@hem.passagen.se>
The client request processing (with this patch) is as follows:
1. Read request data from the client into a buffer
2. If the queue of pending objects is less than 2, then process the
buffer and queue objects for delivery.
3. Immediately restart request processing when a object is delivered,
since we might have read additional requests into the buffer.

27 years agounused var tmp
wessels [Thu, 1 Jan 1998 06:57:16 +0000 (06:57 +0000)] 
unused var tmp

27 years agoFrom: Henrik Nordstrom <hno@hem.passagen.se>
wessels [Thu, 1 Jan 1998 06:55:12 +0000 (06:55 +0000)] 
From:    Henrik Nordstrom <hno@hem.passagen.se>
strdup on the request buffer is not a great idea
* The buffer can contain NULL bytes.
* Unneded overhead
memmove does the job much better, and without corrupting the data on the
way.

27 years agoupdate
wessels [Thu, 1 Jan 1998 05:25:39 +0000 (05:25 +0000)] 
update

27 years agochanged meaning of ERR_CANNOT_FORWARD
wessels [Thu, 1 Jan 1998 05:13:53 +0000 (05:13 +0000)] 
changed meaning of ERR_CANNOT_FORWARD

27 years agoupdate
wessels [Thu, 1 Jan 1998 05:13:28 +0000 (05:13 +0000)] 
update

27 years agoin comm_poll(), the code was #ifdef'd out which closed the sever
wessels [Thu, 1 Jan 1998 05:13:20 +0000 (05:13 +0000)] 
in comm_poll(), the code was #ifdef'd out which closed the sever
sockets upon shutdown.

27 years agoThe ERR_CANNOT_FORWARD error was overloaded. The text of the message
wessels [Thu, 1 Jan 1998 05:11:56 +0000 (05:11 +0000)] 
The ERR_CANNOT_FORWARD error was overloaded.  The text of the message
implies that a 'miss_access' rule is the cause.  This error has
been renamed to ERR_FORWARDING_DENIED.  The meaning of ERR_CANNOT_FORWARD
has been changed to mean that we could not find a next-hop source
for this request, presumably because 'never_direct' is in use and all
parent caches are down.

27 years agoFrom: Peter Wemm <peter@netplex.com.au>
wessels [Thu, 1 Jan 1998 03:45:42 +0000 (03:45 +0000)] 
From:    Peter Wemm <peter@netplex.com.au>
Dont use libresolv on FreeBSD.  If its on the system, chances are
that its a bogus one from some other distribution and will not work
properly.

27 years agoFrom: Henrik Nordstrom <hno@hem.passagen.se>
wessels [Thu, 1 Jan 1998 03:41:54 +0000 (03:41 +0000)] 
From:    Henrik Nordstrom <hno@hem.passagen.se>
Here are some additional autoconf options..
--enable-xmalloc-debug-count
--async-io
And a very small spelling correction to the previous patch

27 years agoFrom: Henrik Nordstrom <hno@hem.passagen.se>
wessels [Thu, 1 Jan 1998 03:38:24 +0000 (03:38 +0000)] 
From:    Henrik Nordstrom <hno@hem.passagen.se>
* Enable caching for extended tests (where appropriate)
* --enable-store-key=sha|md5
* --enable-xmalloc-statistics
* --enable-xmalloc-debug
* The autoconf patches I sent earlier (setresuid, gnumalloc)
* A fix to the setresuid patch (HAVE_SETRESUID was not updated)

27 years agoFrom: Henrik Nordstrom <hno@hem.passagen.se>
wessels [Thu, 1 Jan 1998 03:32:32 +0000 (03:32 +0000)] 
From:    Henrik Nordstrom <hno@hem.passagen.se>
Here is a patch that reuses the already opened data channel when we
trigger the different "hacks" (directory without trailing /, or the
slashhack).

27 years agoDon't set a close handler for the data FD. If we have both ctrl and
wessels [Thu, 1 Jan 1998 02:28:08 +0000 (02:28 +0000)] 
Don't set a close handler for the data FD.  If we have both ctrl and
data FD's call ftpStateFree() upon close, then we have to delete the
close handler which did NOT get called to prevent ftpStateFree()
getting called twice.  Instead we'll always call comm_close() on the
ctrl FD.

27 years agoAdd comm_add_close_handler() for ftpState->data.fd. If the data socket
wessels [Thu, 1 Jan 1998 00:52:04 +0000 (00:52 +0000)] 
Add comm_add_close_handler() for ftpState->data.fd.  If the data socket
connection failed, then a comm_close() call was closing the data socket,
but not freeing up the ftpState.

27 years agoupdate
wessels [Wed, 31 Dec 1997 13:01:23 +0000 (13:01 +0000)] 
update

27 years agotypo
wessels [Wed, 31 Dec 1997 12:58:30 +0000 (12:58 +0000)] 
typo

27 years agoupdate
wessels [Wed, 31 Dec 1997 12:56:11 +0000 (12:56 +0000)] 
update

27 years agoWrite the clean swap.state log in the LRU list order instead of the
wessels [Wed, 31 Dec 1997 12:53:19 +0000 (12:53 +0000)] 
Write the clean swap.state log in the LRU list order instead of the
random hash table order.

27 years agorequire poll.h before using poll()
wessels [Wed, 31 Dec 1997 12:28:33 +0000 (12:28 +0000)] 
require poll.h before using poll()

27 years agopthread update
wessels [Wed, 31 Dec 1997 12:23:17 +0000 (12:23 +0000)] 
pthread update

27 years agoFrom: Stewart Forster <slf@connect.com.au>
wessels [Wed, 31 Dec 1997 12:06:22 +0000 (12:06 +0000)] 
From:    Stewart Forster <slf@connect.com.au>
In the default makefile, there is no mention of needing to include
-lpthread for Solaris (I don't know about other OS's).  This WILL
cause squid to fail under Solaris, and the standard C library has
non-reentrant routines that will die unless the appropriate thread
library is included.

27 years agoFrom: Stewart Forster <slf@connect.com.au>
wessels [Wed, 31 Dec 1997 12:01:46 +0000 (12:01 +0000)] 
From:    Stewart Forster <slf@connect.com.au>
In comm.c, for VERY low load if you're using async I/O it will go
VERY slow because it will only timeout of the poll every second if
there's no traffic to squid.  The fix is to reduce the timeout.
This will now poll for asyncio 20 times a second, rather than once.
This will give a certain 40K/sec transfer rate on a Solaris box.

27 years agoupdate
wessels [Wed, 31 Dec 1997 11:59:55 +0000 (11:59 +0000)] 
update

27 years agoFrom: Stewart Forster <slf@connect.com.au>
wessels [Wed, 31 Dec 1997 11:58:04 +0000 (11:58 +0000)] 
From:    Stewart Forster <slf@connect.com.au>
The patch given by Oskar Pearson was too Linux-centric and ignored
all other OS's desires to use poll().  The corrected business is
below.

27 years ago1.2.beta10
wessels [Wed, 31 Dec 1997 11:13:53 +0000 (11:13 +0000)] 
1.2.beta10

27 years agoadd options.h to distclean
wessels [Wed, 31 Dec 1997 11:11:56 +0000 (11:11 +0000)] 
add options.h to distclean

27 years agoupdate
wessels [Wed, 31 Dec 1997 06:58:43 +0000 (06:58 +0000)] 
update

27 years agochange warning-generating-cast in printf to an assertion
wessels [Wed, 31 Dec 1997 06:18:47 +0000 (06:18 +0000)] 
change warning-generating-cast in printf to an assertion

27 years ago- Fixed bogus call to storeDirUpdateSwapSize() for "aborted" objects
wessels [Wed, 31 Dec 1997 06:09:45 +0000 (06:09 +0000)] 
- Fixed bogus call to storeDirUpdateSwapSize() for "aborted" objects
  or those which never reach SWAPOUT_DONE state.
- use e->object_len in every storeDirUpdateSwapSize() call.
- assert swap fileno is free in getUnusedFileno()
- in storeRelease(), call storeLog() before clearing swap_file_number.

27 years agoNeed to lock entry during whois fetch
wessels [Wed, 31 Dec 1997 06:07:11 +0000 (06:07 +0000)] 
Need to lock entry during whois fetch

27 years agoFixed bad pointer parameter to asnMatchIp(), asn code works fine now.
kostas [Wed, 31 Dec 1997 04:58:05 +0000 (04:58 +0000)] 
Fixed bad pointer parameter to asnMatchIp(), asn code works fine now.

27 years agoASN cleanup
wessels [Tue, 30 Dec 1997 11:22:05 +0000 (11:22 +0000)] 
ASN cleanup

27 years agomore thorough check for setresuid(); esp needed on linux
wessels [Tue, 30 Dec 1997 09:49:40 +0000 (09:49 +0000)] 
more thorough check for setresuid(); esp needed on linux

27 years agogindent
wessels [Tue, 30 Dec 1997 09:47:38 +0000 (09:47 +0000)] 
gindent

27 years agoTRACE method fixes
wessels [Tue, 30 Dec 1997 09:47:03 +0000 (09:47 +0000)] 
TRACE method fixes

27 years agoFix for proxy auth username/passwd bug from BoB Miorelli.
wessels [Tue, 30 Dec 1997 09:46:35 +0000 (09:46 +0000)] 
Fix for proxy auth username/passwd bug from BoB Miorelli.
gindent

27 years ago*** empty log message ***
kostas [Mon, 29 Dec 1997 22:26:29 +0000 (22:26 +0000)] 
*** empty log message ***

27 years ago*** empty log message ***
kostas [Sun, 28 Dec 1997 03:23:05 +0000 (03:23 +0000)] 
*** empty log message ***

27 years ago*** empty log message ***
kostas [Sun, 28 Dec 1997 01:24:29 +0000 (01:24 +0000)] 
*** empty log message ***

27 years ago*** empty log message ***
kostas [Sun, 28 Dec 1997 01:16:07 +0000 (01:16 +0000)] 
*** empty log message ***

27 years agoAligned to squid coding practice (xfree/xmalloc etc)
kostas [Sun, 28 Dec 1997 01:15:29 +0000 (01:15 +0000)] 
Aligned to squid coding practice (xfree/xmalloc etc)

27 years agoSquid now caches whois replies for efficiency in retrieving as<->network
kostas [Sun, 28 Dec 1997 01:14:59 +0000 (01:14 +0000)] 
Squid now caches whois replies for efficiency in retrieving as<->network
number info. Error pages are reloaded on SIGHUP. Minor snmp cleanup.

27 years agoAdded radix stuff.
kostas [Sun, 21 Dec 1997 18:21:11 +0000 (18:21 +0000)] 
Added radix stuff.

27 years agoAdded asn protos
kostas [Sun, 21 Dec 1997 18:19:27 +0000 (18:19 +0000)] 
Added asn protos

27 years agoAdded description for as# acls and example for as# routing.
kostas [Sun, 21 Dec 1997 18:10:02 +0000 (18:10 +0000)] 
Added description for as# acls and example for as# routing.

27 years agoHeader file for radix.c
kostas [Sun, 21 Dec 1997 18:03:28 +0000 (18:03 +0000)] 
Header file for radix.c

27 years agoRadix-tree algorithm, from NetBSD-4.4
kostas [Sun, 21 Dec 1997 18:02:49 +0000 (18:02 +0000)] 
Radix-tree algorithm, from NetBSD-4.4

27 years agoAdded radixtree algorithm, derived from NetBSD-4.4
kostas [Sun, 21 Dec 1997 18:01:51 +0000 (18:01 +0000)] 
Added radixtree algorithm, derived from NetBSD-4.4

27 years agoModified for AS support. We need to know the ip address before going
kostas [Sun, 21 Dec 1997 18:00:19 +0000 (18:00 +0000)] 
Modified for AS support. We need to know the ip address before going
to PeerSelectFoo() and neighborCount() because AclMatchAcl() can't
block (and it would be more complicated to make a n-b version of it)

27 years agoInitial as number acl checking for clients plus hack for server as number
kostas [Sun, 21 Dec 1997 17:57:52 +0000 (17:57 +0000)] 
Initial as number acl checking for clients plus hack for server as number
routing

27 years agoAesthetically improved urn:menu.* using tables
kostas [Thu, 11 Dec 1997 04:53:15 +0000 (04:53 +0000)] 
Aesthetically improved urn:menu.* using tables

27 years agoAdded menu with rtt plus [cached] indication
kostas [Thu, 11 Dec 1997 04:05:02 +0000 (04:05 +0000)] 
Added menu with rtt plus [cached] indication

27 years agomenu hack fix
wessels [Tue, 9 Dec 1997 11:13:04 +0000 (11:13 +0000)] 
menu hack fix

27 years agoforce menu option
wessels [Tue, 9 Dec 1997 11:02:08 +0000 (11:02 +0000)] 
force menu option

27 years ago1.2.beta9
wessels [Mon, 8 Dec 1997 22:15:36 +0000 (22:15 +0000)] 
1.2.beta9