-Revision 1.522 2005/03/04 21:48:12 cheshire
-<rdar://problem/4037283> Fractional time rounded down instead of up on platforms with coarse clock granularity
-
-Revision 1.521 2005/02/25 04:21:00 cheshire
-<rdar://problem/4015377> mDNS -F returns the same domain multiple times with different casing
-
-Revision 1.520 2005/02/16 01:14:11 cheshire
-Convert RR Cache LogOperation() calls to debugf()
-
-Revision 1.519 2005/02/15 01:57:20 cheshire
-When setting "q->LastQTxTime = m->timenow", must also clear q->RecentAnswerPkts to zero
-
-Revision 1.518 2005/02/10 22:35:17 cheshire
-<rdar://problem/3727944> Update name
-
-Revision 1.517 2005/02/03 00:21:21 cheshire
-Update comments about BIND named and zero-length TXT records
-
-Revision 1.516 2005/01/28 06:06:32 cheshire
-Update comment
-
-Revision 1.515 2005/01/27 00:21:49 cheshire
-<rdar://problem/3973798> Remove mDNSResponder sleep/wake syslog message
-
-Revision 1.514 2005/01/21 01:33:45 cheshire
-<rdar://problem/3962979> Shutdown time regression: mDNSResponder not responding to SIGTERM
-
-Revision 1.513 2005/01/21 00:07:54 cheshire
-<rdar://problem/3962717> Infinite loop when the same service is registered twice, and then suffers a name conflict
-
-Revision 1.512 2005/01/20 00:37:45 cheshire
-<rdar://problem/3941448> mDNSResponder crashed in mDNSCoreReceiveResponse
-Take care not to recycle records while they are on the CacheFlushRecords list
-
-Revision 1.511 2005/01/19 22:48:53 cheshire
-<rdar://problem/3955355> Handle services with subtypes correctly when doing mDNS_RenameAndReregisterService()
-
-Revision 1.510 2005/01/19 03:12:45 cheshire
-Move LocalRecordReady() macro from mDNS.c to DNSCommon.h
-
-Revision 1.509 2005/01/19 03:08:49 cheshire
-<rdar://problem/3961051> CPU Spin in mDNSResponder
-Log messages to help catch and report CPU spins
-
-Revision 1.508 2005/01/18 18:56:32 cheshire
-<rdar://problem/3934245> QU responses not promoted to multicast responses when appropriate
-
-Revision 1.507 2005/01/18 01:12:07 cheshire
-<rdar://problem/3956258> Logging into VPN causes mDNSResponder to reissue multicast probes
-
-Revision 1.506 2005/01/17 23:28:53 cheshire
-Fix compile error
-
-Revision 1.505 2005/01/11 02:02:56 shersche
-Move variable declaration to the beginning of statement block
-
-Revision 1.504 2004/12/20 20:24:35 cheshire
-<rdar://problem/3928456> Network efficiency: Don't keep polling if we have at least one unique-type answer
-
-Revision 1.503 2004/12/20 18:41:47 cheshire
-<rdar://problem/3591622> Low memory support: Provide answers even when we don't have cache space
-
-Revision 1.502 2004/12/20 18:04:08 cheshire
-<rdar://problem/3923098> For now, don't put standard wide-area unicast responses in our main cache
-
-Revision 1.501 2004/12/19 23:50:18 cheshire
-<rdar://problem/3751638> kDNSServiceInterfaceIndexLocalOnly should return all local records
-Don't show "No active interface to send" messages for kDNSServiceInterfaceIndexLocalOnly services
-
-Revision 1.500 2004/12/18 03:13:46 cheshire
-<rdar://problem/3751638> kDNSServiceInterfaceIndexLocalOnly should return all local records
-
-Revision 1.499 2004/12/17 23:37:45 cheshire
-<rdar://problem/3485365> Guard against repeating wireless dissociation/re-association
-(and other repetitive configuration changes)
-
-Revision 1.498 2004/12/17 05:25:46 cheshire
-<rdar://problem/3925163> Shorten DNS-SD queries to avoid NAT bugs
-
-Revision 1.497 2004/12/17 03:20:58 cheshire
-<rdar://problem/3925168> Don't send unicast replies we know will be ignored
-
-Revision 1.496 2004/12/16 22:18:26 cheshire
-Make AddressIsLocalSubnet() a little more selective -- ignore point-to-point interfaces
-
-Revision 1.495 2004/12/16 21:27:37 ksekar
-Fixed build failures when compiled with verbose debugging messages
-
-Revision 1.494 2004/12/16 20:46:56 cheshire
-Fix compiler warnings
-
-Revision 1.493 2004/12/16 20:13:00 cheshire
-<rdar://problem/3324626> Cache memory management improvements
-
-Revision 1.492 2004/12/16 08:03:24 shersche
-Fix compilation error when UNICAST_DISABLED is set
-
-Revision 1.491 2004/12/11 01:52:11 cheshire
-<rdar://problem/3785820> Support kDNSServiceFlagsAllowRemoteQuery for registering services too
-
-Revision 1.490 2004/12/10 20:06:25 cheshire
-<rdar://problem/3915074> Reduce egregious stack space usage
-Reduced SendDelayedUnicastResponse() stack frame from 9K to 112 bytes
-
-Revision 1.489 2004/12/10 20:03:43 cheshire
-<rdar://problem/3915074> Reduce egregious stack space usage
-Reduced mDNSCoreReceiveQuery() stack frame from 9K to 144 bytes
-
-Revision 1.488 2004/12/10 19:50:41 cheshire
-<rdar://problem/3915074> Reduce egregious stack space usage
-Reduced SendResponses() stack frame from 9K to 176 bytes
-
-Revision 1.487 2004/12/10 19:39:13 cheshire
-<rdar://problem/3915074> Reduce egregious stack space usage
-Reduced SendQueries() stack frame from 18K to 112 bytes
-
-Revision 1.486 2004/12/10 14:16:17 cheshire
-<rdar://problem/3889788> Relax update rate limiting
-We now allow an average rate of ten updates per minute.
-Updates in excess of that are rate limited, but more gently than before.
-
-Revision 1.485 2004/12/10 02:09:24 cheshire
-<rdar://problem/3898376> Modify default TTLs
-
-Revision 1.484 2004/12/09 03:15:40 ksekar
-<rdar://problem/3806610> use _legacy instead of _default to find "empty string" browse domains
-
-Revision 1.483 2004/12/07 23:00:14 ksekar
-<rdar://problem/3908336> DNSServiceRegisterRecord() can crash on deregistration:
-Call RecordProbeFailure even if there is no record callback
-
-Revision 1.482 2004/12/07 22:49:06 cheshire
-<rdar://problem/3908850> BIND doesn't allow zero-length TXT records
-
-Revision 1.481 2004/12/07 21:26:04 ksekar
-<rdar://problem/3908336> DNSServiceRegisterRecord() can crash on deregistration
-
-Revision 1.480 2004/12/07 20:42:33 cheshire
-Add explicit context parameter to mDNS_RemoveRecordFromService()
-
-Revision 1.479 2004/12/07 17:50:49 ksekar
-<rdar://problem/3908850> BIND doesn't allow zero-length TXT records
-
-Revision 1.478 2004/12/06 21:15:22 ksekar
-<rdar://problem/3884386> mDNSResponder crashed in CheckServiceRegistrations
-
-Revision 1.477 2004/12/04 02:12:45 cheshire
-<rdar://problem/3517236> mDNSResponder puts LargeCacheRecord on the stack
-
-Revision 1.476 2004/11/29 23:34:31 cheshire
-On platforms with coarse time resolutions, ORing time values with one to ensure they are non-zero
-is crude, and effectively halves the time resolution. The more selective NonZeroTime() function
-only nudges the time value to 1 if the interval calculation happens to result in the value zero.
-
-Revision 1.475 2004/11/29 23:13:31 cheshire
-<rdar://problem/3484552> All unique records in a set should have the cache flush bit set
-Additional check: Make sure we don't unnecessarily send packets containing only additionals.
-(This could occur with multi-packet KA lists, if the answer and additionals were marked
-by the query packet, and then the answer were later suppressed in a subsequent KA packet.)
-
-Revision 1.474 2004/11/29 17:18:12 cheshire
-Remove "Unknown DNS packet type" message for update responses
-
-Revision 1.473 2004/11/25 01:57:52 cheshire
-<rdar://problem/3484552> All unique records in a set should have the cache flush bit set
-
-Revision 1.472 2004/11/25 01:28:09 cheshire
-<rdar://problem/3557050> Need to implement random delay for 'QU' unicast replies (and set cache flush bit too)
-
-Revision 1.471 2004/11/25 01:10:13 cheshire
-Move code to add additional records to a subroutine called AddAdditionalsToResponseList()
-
-Revision 1.470 2004/11/24 21:54:44 cheshire
-<rdar://problem/3894475> mDNSCore not receiving unicast responses properly
-
-Revision 1.469 2004/11/24 04:50:39 cheshire
-Minor tidying
-
-Revision 1.468 2004/11/24 01:47:07 cheshire
-<rdar://problem/3780207> DNSServiceRegisterRecord should call CallBack on success.
-
-Revision 1.467 2004/11/24 01:41:28 cheshire
-Rename CompleteProbing() to AcknowledgeRecord()
-
-Revision 1.466 2004/11/23 21:08:07 ksekar
-Don't use ID to demux multicast/unicast now that unicast uses random IDs
-
-Revision 1.465 2004/11/15 20:09:21 ksekar
-<rdar://problem/3719050> Wide Area support for Add/Remove record
-
-Revision 1.464 2004/11/03 01:44:36 cheshire
-Update debugging messages
-
-Revision 1.463 2004/10/29 02:38:48 cheshire
-Fix Windows compile errors
-
-Revision 1.462 2004/10/28 19:21:07 cheshire
-Guard against registering interface with zero InterfaceID
-
-Revision 1.461 2004/10/28 19:02:16 cheshire
-Remove \n from LogMsg() call
-
-Revision 1.460 2004/10/28 03:24:40 cheshire
-Rename m->CanReceiveUnicastOn as m->CanReceiveUnicastOn5353
-
-Revision 1.459 2004/10/26 22:34:37 cheshire
-<rdar://problem/3468995> Need to protect mDNSResponder from unbounded packet flooding
-
-Revision 1.458 2004/10/26 20:45:28 cheshire
-Show mask in "invalid mask" message
-
-Revision 1.457 2004/10/26 06:28:36 cheshire
-Now that we don't check IP TTL any more, remove associated log message
-
-Revision 1.456 2004/10/26 06:21:42 cheshire
-Adjust mask validity check to allow an all-ones mask (for IPv6 ::1 loopback address)
-
-Revision 1.455 2004/10/26 06:11:40 cheshire
-Add improved logging to aid in diagnosis of <rdar://problem/3842714> mDNSResponder crashed
-
-Revision 1.454 2004/10/23 01:16:00 cheshire
-<rdar://problem/3851677> uDNS operations not always reliable on multi-homed hosts
-
-Revision 1.453 2004/10/22 20:52:06 ksekar
-<rdar://problem/3799260> Create NAT port mappings for Long Lived Queries
-
-Revision 1.452 2004/10/20 01:50:40 cheshire
-<rdar://problem/3844991> Cannot resolve non-local registrations using the mach API
-Implemented ForceMCast mode for AuthRecords as well as for Questions
-
-Revision 1.451 2004/10/19 21:33:15 cheshire
-<rdar://problem/3844991> Cannot resolve non-local registrations using the mach API
-Added flag 'kDNSServiceFlagsForceMulticast'. Passing through an interface id for a unicast name
-doesn't force multicast unless you set this flag to indicate explicitly that this is what you want
-
-Revision 1.450 2004/10/19 17:42:59 ksekar
-Fixed compiler warnings for non-debug builds.
-
-Revision 1.449 2004/10/18 22:57:07 cheshire
-<rdar://problem/3711302> Seen in console: Ignored apparent spoof mDNS Response with TTL 1
-
-Revision 1.448 2004/10/16 00:16:59 cheshire
-<rdar://problem/3770558> Replace IP TTL 255 check with local subnet source address check
-
-Revision 1.447 2004/10/15 00:51:21 cheshire
-<rdar://problem/3711302> Seen in console: Ignored apparent spoof mDNS Response with TTL 1
-
-Revision 1.446 2004/10/14 00:43:34 cheshire
-<rdar://problem/3815984> Services continue to announce SRV and HINFO
-
-Revision 1.445 2004/10/12 21:07:09 cheshire
-Set up m->p in mDNS_Init() before calling mDNSPlatformTimeInit()
-
-Revision 1.444 2004/10/11 17:54:16 ksekar
-Changed hashtable pointer output from debugf to verbosedebugf.
-
-Revision 1.443 2004/10/10 07:05:45 cheshire
-For consistency, use symbol "localdomain" instead of literal string
-
-Revision 1.442 2004/10/08 20:25:10 cheshire
-Change of plan for <rdar://problem/3831716> -- we're not going to do that at this time
-
-Revision 1.441 2004/10/08 03:25:01 ksekar
-<rdar://problem/3831716> domain enumeration should use LLQs
-
-Revision 1.440 2004/10/06 01:44:19 cheshire
-<rdar://problem/3813936> Resolving too quickly sometimes returns stale TXT record
-
-Revision 1.439 2004/10/03 23:14:11 cheshire
-Add "mDNSEthAddr" type and "zeroEthAddr" constant
-
-Revision 1.438 2004/09/29 23:07:04 cheshire
-Patch from Pavel Repin to fix compile error on Windows
-
-Revision 1.437 2004/09/28 02:23:50 cheshire
-<rdar://problem/3637266> Deliver near-pending "remove" events before new "add" events
-Don't need to search the entire cache for nearly-expired records -- just the appropriate hash slot
-For records with the cache flush bit set, defer the decision until the end of the packet
-
-Revision 1.436 2004/09/28 01:27:04 cheshire
-Update incorrect log message
-
-Revision 1.435 2004/09/25 02:41:39 cheshire
-<rdar://problem/3637266> Deliver near-pending "remove" events before new "add" events
-
-Revision 1.434 2004/09/25 02:32:06 cheshire
-Update comments
-
-Revision 1.433 2004/09/25 02:24:27 cheshire
-Removed unused rr->UseCount
-
-Revision 1.432 2004/09/24 21:35:17 cheshire
-<rdar://problem/3561220> Browses are no longer piggybacking on other browses
-TargetPort and TargetQID are allowed to be undefined if no question->Target is set
-
-Revision 1.431 2004/09/24 21:33:12 cheshire
-Adjust comment
-
-Revision 1.430 2004/09/24 02:15:49 cheshire
-<rdar://problem/3680865> Late conflicts don't send goodbye packets on other interfaces
-
-Revision 1.429 2004/09/24 00:20:21 cheshire
-<rdar://problem/3483349> Any rrtype is a conflict for unique records
-
-Revision 1.428 2004/09/24 00:12:25 cheshire
-Get rid of unused RRUniqueOrKnownUnique(RR)
-
-Revision 1.427 2004/09/23 20:44:11 cheshire
-<rdar://problem/3813148> Reduce timeout before expiring records on failure
-
-Revision 1.426 2004/09/23 20:21:07 cheshire
-<rdar://problem/3426876> Refine "immediate answer burst; restarting exponential backoff sequence" logic
-Associate a unique sequence number with each received packet, and only increment the count of recent answer
-packets if the packet sequence number for this answer record is not one we've already seen and counted.
-
-Revision 1.425 2004/09/23 20:14:38 cheshire
-Rename "question->RecentAnswers" to "question->RecentAnswerPkts"
-
-Revision 1.424 2004/09/23 00:58:36 cheshire
-<rdar://problem/3781269> Rate limiting interferes with updating TXT records
-
-Revision 1.423 2004/09/23 00:50:53 cheshire
-<rdar://problem/3419452> Don't send a (DE) if a service is unregistered after wake from sleep
-
-Revision 1.422 2004/09/22 02:34:46 cheshire
-Move definitions of default TTL times from mDNS.c to mDNSEmbeddedAPI.h
-
-Revision 1.421 2004/09/21 23:29:49 cheshire
-<rdar://problem/3680045> DNSServiceResolve should delay sending packets
-
-Revision 1.420 2004/09/21 23:01:42 cheshire
-Update debugf messages
-
-Revision 1.419 2004/09/21 19:51:14 cheshire
-Move "Starting time value" message from mDNS.c to mDNSMacOSX/daemon.c
-
-Revision 1.418 2004/09/21 18:40:17 cheshire
-<rdar://problem/3376752> Adjust default record TTLs
-
-Revision 1.417 2004/09/21 17:32:16 cheshire
-<rdar://problem/3809484> Rate limiting imposed too soon
-
-Revision 1.416 2004/09/20 23:52:01 cheshire
-CFSocket{Puma}.c renamed to mDNSMacOSX{Puma}.c
-
-Revision 1.415 2004/09/18 01:14:09 cheshire
-<rdar://problem/3485375> Resolve() should not bother doing AAAA queries on machines with no IPv6 interfaces
-
-Revision 1.414 2004/09/18 01:06:48 cheshire
-Add comments
-
-Revision 1.413 2004/09/17 01:08:48 cheshire
-Renamed mDNSClientAPI.h to mDNSEmbeddedAPI.h
- The name "mDNSClientAPI.h" is misleading to new developers looking at this code. The interfaces
- declared in that file are ONLY appropriate to single-address-space embedded applications.
- For clients on general-purpose computers, the interfaces defined in dns_sd.h should be used.
-
-Revision 1.412 2004/09/17 00:46:33 cheshire
-mDNS_TimeNow should take const mDNS parameter
-
-Revision 1.411 2004/09/17 00:31:51 cheshire
-For consistency with ipv6, renamed rdata field 'ip' to 'ipv4'
-
-Revision 1.410 2004/09/17 00:19:10 cheshire
-For consistency with AllDNSLinkGroupv6, rename AllDNSLinkGroup to AllDNSLinkGroupv4
-
-Revision 1.409 2004/09/16 21:59:15 cheshire
-For consistency with zerov6Addr, rename zeroIPAddr to zerov4Addr
-
-Revision 1.408 2004/09/16 21:36:36 cheshire
-<rdar://problem/3803162> Fix unsafe use of mDNSPlatformTimeNow()
-Changes to add necessary locking calls around unicast DNS operations
-
-Revision 1.407 2004/09/16 02:29:39 cheshire
-Moved mDNS_Lock/mDNS_Unlock to DNSCommon.c; Added necessary locking around
-uDNS_ReceiveMsg, uDNS_StartQuery, uDNS_UpdateRecord, uDNS_RegisterService
-
-Revision 1.406 2004/09/16 01:58:14 cheshire
-Fix compiler warnings
-
-Revision 1.405 2004/09/16 00:24:48 cheshire
-<rdar://problem/3803162> Fix unsafe use of mDNSPlatformTimeNow()
-
-Revision 1.404 2004/09/15 21:44:11 cheshire
-<rdar://problem/3681031> Randomize initial timenow_adjust value in mDNS_Init
-Show time value in log to help diagnose errors
-
-Revision 1.403 2004/09/15 00:46:32 ksekar
-Changed debugf to verbosedebugf in CheckCacheExpiration
-
-Revision 1.402 2004/09/14 23:59:55 cheshire
-<rdar://problem/3681031> Randomize initial timenow_adjust value in mDNS_Init
-
-Revision 1.401 2004/09/14 23:27:46 cheshire
-Fix compile errors
-
-Revision 1.400 2004/09/02 03:48:47 cheshire
-<rdar://problem/3709039> Disable targeted unicast query support by default
-1. New flag kDNSServiceFlagsAllowRemoteQuery to indicate we want to allow remote queries for this record
-2. New field AllowRemoteQuery in AuthRecord structure
-3. uds_daemon.c sets AllowRemoteQuery if kDNSServiceFlagsAllowRemoteQuery is set
-4. mDNS.c only answers remote queries if AllowRemoteQuery is set
-
-Revision 1.399 2004/09/02 01:39:40 cheshire
-For better readability, follow consistent convention that QR bit comes first, followed by OP bits
-
-Revision 1.398 2004/09/01 03:59:29 ksekar
-<rdar://problem/3783453>: Conditionally compile out uDNS code on Windows
-
-Revision 1.397 2004/08/25 22:04:25 rpantos
-Fix the standard Windows compile error.
-
-Revision 1.396 2004/08/25 00:37:27 ksekar
-<rdar://problem/3774635>: Cleanup DynDNS hostname registration code
-
-Revision 1.395 2004/08/18 17:21:18 ksekar
-Removed double-call of uDNS_AdvertiseInterface from mDNS_SetFQDNs()
-
-Revision 1.394 2004/08/14 03:22:41 cheshire
-<rdar://problem/3762579> Dynamic DNS UI <-> mDNSResponder glue
-Add GetUserSpecifiedDDNSName() routine
-Convert ServiceRegDomain to domainname instead of C string
-Replace mDNS_GenerateFQDN/mDNS_GenerateGlobalFQDN with mDNS_SetFQDNs
-
-Revision 1.393 2004/08/13 23:42:52 cheshire
-Removed unused "zeroDomainNamePtr"
-
-Revision 1.392 2004/08/13 23:37:02 cheshire
-Now that we do both uDNS and mDNS, global replace "uDNS_info.hostname" with
-"uDNS_info.UnicastHostname" for clarity
-
-Revision 1.391 2004/08/13 23:25:00 cheshire
-Now that we do both uDNS and mDNS, global replace "m->hostname" with
-"m->MulticastHostname" for clarity
-
-Revision 1.390 2004/08/11 02:17:01 cheshire
-<rdar://problem/3514236> Registering service with port number 0 should create a "No Such Service" record
-
-Revision 1.389 2004/08/10 23:19:14 ksekar
-<rdar://problem/3722542>: DNS Extension daemon for Wide Area Service Discovery
-Moved routines/constants to allow extern access for garbage collection daemon
-
-Revision 1.388 2004/07/30 17:40:06 ksekar
-<rdar://problem/3739115>: TXT Record updates not available for wide-area services
-
-Revision 1.387 2004/07/26 22:49:30 ksekar
-<rdar://problem/3651409>: Feature #9516: Need support for NATPMP in client
-
-Revision 1.386 2004/07/13 21:24:24 rpantos
-Fix for <rdar://problem/3701120>.
-
-Revision 1.385 2004/06/18 19:09:59 cheshire
-<rdar://problem/3588761> Current method of doing subtypes causes name collisions
-
-Revision 1.384 2004/06/15 04:31:23 cheshire
-Make sure to clear m->CurrentRecord at the end of AnswerNewLocalOnlyQuestion()
-
-Revision 1.383 2004/06/11 00:04:59 cheshire
-<rdar://problem/3595602> TTL must be greater than zero for DNSServiceRegisterRecord
-
-Revision 1.382 2004/06/08 04:59:40 cheshire
-Tidy up wording -- log messages are already prefixed with "mDNSResponder", so don't need to repeat it
-
-Revision 1.381 2004/06/05 00:57:30 cheshire
-Remove incorrect LogMsg()
-
-Revision 1.380 2004/06/05 00:04:26 cheshire
-<rdar://problem/3668639>: wide-area domains should be returned in reg. domain enumeration
-
-Revision 1.379 2004/05/28 23:42:36 ksekar
-<rdar://problem/3258021>: Feature: DNS server->client notification on record changes (#7805)
-
-Revision 1.378 2004/05/25 17:25:25 cheshire
-Remove extraneous blank lines and white space
-
-Revision 1.377 2004/05/18 23:51:25 cheshire
-Tidy up all checkin comments to use consistent "<rdar://problem/xxxxxxx>" format for bug numbers
-
-Revision 1.376 2004/05/05 18:30:44 ksekar
-Restored surpressed Cache Tail debug messages.
-
-Revision 1.375 2004/04/26 21:36:25 cheshire
-Only send IPv4 (or v6) multicast when IPv4 (or v6) multicast send/receive
-is indicated as being available on that interface
-
-Revision 1.374 2004/04/21 02:53:26 cheshire
-Typo in debugf statement
-
-Revision 1.373 2004/04/21 02:49:11 cheshire
-To reduce future confusion, renamed 'TxAndRx' to 'McastTxRx'
-
-Revision 1.372 2004/04/21 02:38:51 cheshire
-Add debugging checks
-
-Revision 1.371 2004/04/14 23:09:28 ksekar
-Support for TSIG signed dynamic updates.
-
-Revision 1.370 2004/04/09 17:40:26 cheshire
-Remove unnecessary "Multicast" field -- it duplicates the semantics of the existing McastTxRx field
-
-Revision 1.369 2004/04/09 16:34:00 cheshire
-Debugging code for later; currently unused
-
-Revision 1.368 2004/04/02 19:19:48 cheshire
-Add code to do optional logging of multi-packet KA list time intervals
-
-Revision 1.367 2004/03/20 03:16:10 cheshire
-Minor refinement to "Excessive update rate" message
-
-Revision 1.366 2004/03/20 03:12:57 cheshire
-<rdar://problem/3587619>: UpdateCredits not granted promptly enough
-
-Revision 1.365 2004/03/19 23:51:22 cheshire
-Change to use symbolic constant kUpdateCreditRefreshInterval instead of (mDNSPlatformOneSecond * 60)
-
-Revision 1.364 2004/03/13 01:57:33 ksekar
-<rdar://problem/3192546>: DynDNS: Dynamic update of service records
-
-Revision 1.363 2004/03/12 21:00:51 cheshire
-Also show port numbers when logging "apparent spoof mDNS Response" messages
-
-Revision 1.362 2004/03/12 08:58:18 cheshire
-Guard against empty TXT records
-
-Revision 1.361 2004/03/09 03:00:46 cheshire
-<rdar://problem/3581961> Don't take lock until after mDNS_Update() has validated that the data is good.
-
-Revision 1.360 2004/03/08 02:52:41 cheshire
-Minor debugging fix: Make sure 'target' is initialized so we don't crash writing debugging log messages
-
-Revision 1.359 2004/03/02 03:21:56 cheshire
-<rdar://problem/3549576> Properly support "_services._dns-sd._udp" meta-queries
-
-Revision 1.358 2004/02/20 08:18:34 cheshire
-<rdar://problem/3564799>: mDNSResponder sometimes announces AAAA records unnecessarily
-
-Revision 1.357 2004/02/18 01:47:41 cheshire
-<rdar://problem/3553472>: Insufficient delay waiting for multi-packet KA lists causes AirPort traffic storms
-
-Revision 1.356 2004/02/06 23:04:19 ksekar
-Basic Dynamic Update support via mDNS_Register (dissabled via
-UNICAST_REGISTRATION #define)
-
-Revision 1.355 2004/02/05 09:32:33 cheshire
-Fix from Bob Bradley: When using the "%.*s" string form,
-guard against truncating in the middle of a multi-byte UTF-8 character.
-
-Revision 1.354 2004/02/05 09:30:22 cheshire
-Update comments
-
-Revision 1.353 2004/01/28 03:41:00 cheshire
-<rdar://problem/3541946>: Need ability to do targeted queries as well as multicast queries
-
-Revision 1.352 2004/01/28 02:30:07 ksekar
-Added default Search Domains to unicast browsing, controlled via
-Networking sharing prefs pane. Stopped sending unicast messages on
-every interface. Fixed unicast resolving via mach-port API.
-
-Revision 1.351 2004/01/27 20:15:22 cheshire
-<rdar://problem/3541288>: Time to prune obsolete code for listening on port 53
-
-Revision 1.350 2004/01/24 23:38:16 cheshire
-Use mDNSVal16() instead of shifting and ORing operations
-
-Revision 1.349 2004/01/23 23:23:14 ksekar
-Added TCP support for truncated unicast messages.
-
-Revision 1.348 2004/01/22 03:54:11 cheshire
-Create special meta-interface 'mDNSInterface_ForceMCast' (-2),
-which means "do this query via multicast, even if it's apparently a unicast domain"
-
-Revision 1.347 2004/01/22 03:50:49 cheshire
-If the client has specified an explicit InterfaceID, then do query by multicast, not unicast
-
-Revision 1.346 2004/01/22 03:48:41 cheshire
-Make sure uDNS client doesn't accidentally use query ID zero
-
-Revision 1.345 2004/01/22 03:43:08 cheshire
-Export constants like mDNSInterface_LocalOnly so that the client layers can use them
-
-Revision 1.344 2004/01/21 21:53:18 cheshire
-<rdar://problem/3448144>: Don't try to receive unicast responses if we're not the first to bind to the UDP port
-
-Revision 1.343 2003/12/23 00:07:47 cheshire
-Make port number in debug message be five-character field, left justified
-
-Revision 1.342 2003/12/20 01:34:28 cheshire
-<rdar://problem/3515876>: Error putting additional records into packets
-Another fix from Rampi: responseptr needs to be updated inside the "for" loop,
-after every record, not once at the end.
-
-Revision 1.341 2003/12/18 22:56:12 cheshire
-<rdar://problem/3510798>: Reduce syslog messages about ignored spoof packets
-
-Revision 1.340 2003/12/16 02:31:37 cheshire
-Minor update to comments
-
-Revision 1.339 2003/12/13 05:50:33 bradley
-Fixed crash with mDNS_Lock/Unlock being called for the initial GrowCache before the platform
-layer has been initialized. Protect mDNS_reentrancy when completing the core initialization to
-fix a race condition during async initialization. Fixed buffer overrun for 1 byte mDNS_snprintf.
-
-Revision 1.338 2003/12/13 03:05:27 ksekar
-<rdar://problem/3192548>: DynDNS: Unicast query of service records
-
-Revision 1.337 2003/12/01 21:46:05 cheshire
-mDNS_StartQuery returns mStatus_BadInterfaceErr if the specified interface does not exist
-
-Revision 1.336 2003/12/01 21:26:19 cheshire
-Guard against zero-length sbuffer in mDNS_vsnprintf()
-
-Revision 1.335 2003/12/01 20:27:48 cheshire
-Display IPv6 addresses correctly (e.g. in log messages) on little-endian processors
-
-Revision 1.334 2003/11/20 22:59:53 cheshire
-Changed runtime checks in mDNS.c to be compile-time checks in mDNSEmbeddedAPI.h
-Thanks to Bob Bradley for suggesting the ingenious compiler trick to make this work.
-
-Revision 1.333 2003/11/20 20:49:53 cheshire
-Another fix from HP: Use packedstruct macro to ensure proper packing for on-the-wire packet structures
-
-Revision 1.332 2003/11/20 05:47:37 cheshire
-<rdar://problem/3490355>: Don't exclude known answers whose expiry time is before the next query
-Now that we only include answers in the known answer list if they are less than
-halfway to expiry, the check to also see if we have another query scheduled
-before the record expires is no longer necessary (and in fact, not correct).
-
-Revision 1.331 2003/11/19 22:31:48 cheshire
-When automatically adding A records to SRVs, add them as additionals, not answers
-
-Revision 1.330 2003/11/19 22:28:50 cheshire
-Increment/Decrement mDNS_reentrancy around calls to m->MainCallback()
-to allow client to make mDNS calls (specifically the call to mDNS_GrowCache())
-
-Revision 1.329 2003/11/19 22:19:24 cheshire
-Show log message when ignoring packets with bad TTL.
-This is to help diagnose problems on Linux versions that may not report the TTL reliably.
-
-Revision 1.328 2003/11/19 22:06:38 cheshire
-Show log messages when a service or hostname is renamed
-
-Revision 1.327 2003/11/19 22:03:44 cheshire
-Move common "m->NextScheduledResponse = m->timenow" to before "if" statement
-
-Revision 1.326 2003/11/17 22:27:02 cheshire
-Another fix from ramaprasad.kr@hp.com: Improve reply delay computation
-on platforms that have native clock rates below fifty ticks per second.
-
-Revision 1.325 2003/11/17 20:41:44 cheshire
-Fix some missing mDNS_Lock(m)/mDNS_Unlock(m) calls.
-
-Revision 1.324 2003/11/17 20:36:32 cheshire
-Function rename: Remove "mDNS_" prefix from AdvertiseInterface() and
-DeadvertiseInterface() -- they're internal private routines, not API routines.
-
-Revision 1.323 2003/11/14 20:59:08 cheshire
-Clients can't use AssignDomainName macro because mDNSPlatformMemCopy is defined in mDNSPlatformFunctions.h.
-Best solution is just to combine mDNSEmbeddedAPI.h and mDNSPlatformFunctions.h into a single file.
-
-Revision 1.322 2003/11/14 19:47:52 cheshire
-Define symbol MAX_ESCAPED_DOMAIN_NAME to indicate recommended buffer size for ConvertDomainNameToCString
-
-Revision 1.321 2003/11/14 19:18:34 cheshire
-Move AssignDomainName macro to mDNSEmbeddedAPI.h to that client layers can use it too
-
-Revision 1.320 2003/11/13 06:45:04 cheshire
-Fix compiler warning on certain compilers
-
-Revision 1.319 2003/11/13 00:47:40 cheshire
-<rdar://problem/3437556> We should delay AAAA record query if A record already in cache.
-
-Revision 1.318 2003/11/13 00:33:26 cheshire
-Change macro "RRIsAddressType" to "RRTypeIsAddressType"
-
-Revision 1.317 2003/11/13 00:10:49 cheshire
-<rdar://problem/3436412>: Verify that rr data is different before updating.
-
-Revision 1.316 2003/11/08 23:37:54 cheshire
-Give explicit zero initializers to blank static structure, required by certain compilers.
-(Thanks to ramaprasad.kr@hp.com for reporting this.)
-
-Revision 1.315 2003/11/07 03:32:56 cheshire
-<rdar://problem/3472153> mDNSResponder delivers answers in inconsistent order
-This is the real fix. Checkin 1.312 was overly simplistic; Calling GetFreeCacheRR() can sometimes
-purge records from the cache, causing tail pointer *rp to be stale on return. The correct fix is
-to maintain a system-wide tail pointer for each cache slot, and then if neccesary GetFreeCacheRR()
-can update this pointer, so that mDNSCoreReceiveResponse() appends records in the right place.
-
-Revision 1.314 2003/11/07 03:19:49 cheshire
-Minor variable renaming for clarity
-
-Revision 1.313 2003/11/07 03:14:49 cheshire
-Previous checkin proved to be overly simplistic; reversing
-
-Revision 1.312 2003/11/03 23:45:15 cheshire
-<rdar://problem/3472153> mDNSResponder delivers answers in inconsistent order
-Build cache lists in FIFO order, not customary C LIFO order
-(Append new elements to tail of cache list, instead of prepending at the head.)
-
-Revision 1.311 2003/10/09 18:00:11 cheshire
-Another compiler warning fix.
-
-Revision 1.310 2003/10/07 20:27:05 cheshire
-Patch from Bob Bradley, to fix warning and compile error on Windows
-
-Revision 1.309 2003/09/26 01:06:36 cheshire
-<rdar://problem/3427923> Set kDNSClass_UniqueRRSet bit for updates too
-Made new routine HaveSentEntireRRSet() to check if flag should be set
-
-Revision 1.308 2003/09/23 01:05:01 cheshire
-Minor changes to comments and debugf() message
-
-Revision 1.307 2003/09/09 20:13:30 cheshire
-<rdar://problem/3411105> Don't send a Goodbye record if we never announced it
-Ammend checkin 1.304: Off-by-one error: By this place in the function we've already decremented
-rr->AnnounceCount, so the check needs to be for InitialAnnounceCount-1, not InitialAnnounceCount
-
-Revision 1.306 2003/09/09 03:00:03 cheshire
-<rdar://problem/3413099> Services take a long time to disappear when switching networks.
-Added two constants: kDefaultReconfirmTimeForNoAnswer and kDefaultReconfirmTimeForCableDisconnect
-
-Revision 1.305 2003/09/09 02:49:31 cheshire
-<rdar://problem/3413975> Initial probes and queries not grouped on wake-from-sleep
-
-Revision 1.304 2003/09/09 02:41:19 cheshire
-<rdar://problem/3411105> Don't send a Goodbye record if we never announced it
-
-Revision 1.303 2003/09/05 19:55:02 cheshire
-<rdar://problem/3409533> Include address records when announcing SRV records
-
-Revision 1.302 2003/09/05 00:01:36 cheshire
-<rdar://problem/3407549> Don't accelerate queries that have large KA lists
-
-Revision 1.301 2003/09/04 22:51:13 cheshire
-<rdar://problem/3398213> Group probes and goodbyes better
-
-Revision 1.300 2003/09/03 02:40:37 cheshire
-<rdar://problem/3404842> mDNSResponder complains about '_'s
-Underscores are not supposed to be legal in standard DNS names, but IANA appears
-to have allowed them in previous service name registrations, so we should too.
-
-Revision 1.299 2003/09/03 02:33:09 cheshire
-<rdar://problem/3404795> CacheRecordRmv ERROR
-Don't update m->NewQuestions until *after* CheckCacheExpiration();
-
-Revision 1.298 2003/09/03 01:47:01 cheshire
-<rdar://problem/3319418> Services always in a state of flux
-Change mDNS_Reconfirm_internal() minimum timeout from 5 seconds to 45-60 seconds
-
-Revision 1.297 2003/08/29 19:44:15 cheshire
-<rdar://problem/3400967> Traffic reduction: Eliminate synchronized QUs when a new service appears
-1. Use m->RandomQueryDelay to impose a random delay in the range 0-500ms on queries
- that already have at least one unique answer in the cache
-2. For these queries, go straight to QM, skipping QU
-
-Revision 1.296 2003/08/29 19:08:21 cheshire
-<rdar://problem/3400986> Traffic reduction: Eliminate huge KA lists after wake from sleep
-Known answers are no longer eligible to go in the KA list if they are more than half-way to their expiry time.
-
-Revision 1.295 2003/08/28 01:10:59 cheshire
-<rdar://problem/3396034> Add syslog message to report when query is reset because of immediate answer burst
-
-Revision 1.294 2003/08/27 02:30:22 cheshire
-<rdar://problem/3395909> Traffic Reduction: Inefficiencies in DNSServiceResolverResolve()
-One more change: "query->GotTXT" is now a straightforward bi-state boolean again
-
-Revision 1.293 2003/08/27 02:25:31 cheshire
-<rdar://problem/3395909> Traffic Reduction: Inefficiencies in DNSServiceResolverResolve()