From sgruby at markspace.com Wed Nov 2 11:52:16 2005 From: sgruby at markspace.com (Scott Gruby) Date: Wed Nov 2 11:52:27 2005 Subject: [sync-services-developers] New Syncrospector? Message-ID: <22901BDD-27FE-49C9-A947-F9D09D3E3132@markspace.com> Does anyone know if there is a Syncrospector later than 0.5 that works with 10.4.3? It launches, but doesn't work and dumps the following in Console: 2005-11-02 11:50:36.877 Syncrospector[23219] *** -[NSCFArray isd_enumerateObjectsAsArgumentToFunction:]: selector not recognized [self = 0x6079d30] 2005-11-02 11:50:37.125 Syncrospector[23219] [404340] ##### RAISE NSInvalidArgumentException: *** -[NSCFArray isd_enumerateObjectsAsArgumentToFunction:]: selector not recognized [self = 0x6079d30] -- Scott Gruby Lead Engineer Mark/Space, Inc. Please visit for assistance with Mark/Space products. From eseidel at apple.com Wed Nov 2 17:51:44 2005 From: eseidel at apple.com (Eric Seidel) Date: Wed Nov 2 17:51:53 2005 Subject: [sync-services-developers] New Syncrospector? In-Reply-To: <22901BDD-27FE-49C9-A947-F9D09D3E3132@markspace.com> References: <22901BDD-27FE-49C9-A947-F9D09D3E3132@markspace.com> Message-ID: I believe there is a newer one available from ADC. -eric On 02.11.2005, at 11:52, Scott Gruby wrote: > Does anyone know if there is a Syncrospector later than 0.5 that > works with 10.4.3? It launches, but doesn't work and dumps the > following in Console: > > 2005-11-02 11:50:36.877 Syncrospector[23219] *** -[NSCFArray > isd_enumerateObjectsAsArgumentToFunction:]: selector not recognized > [self = 0x6079d30] > 2005-11-02 11:50:37.125 Syncrospector[23219] [404340] ##### RAISE > NSInvalidArgumentException: *** -[NSCFArray > isd_enumerateObjectsAsArgumentToFunction:]: selector not recognized > [self = 0x6079d30] > > > -- > Scott Gruby > Lead Engineer > Mark/Space, Inc. > > > Please visit for assistance > with Mark/Space products. > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers From mildm8nnered at gmail.com Fri Nov 4 09:44:54 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Fri Nov 4 09:44:58 2005 Subject: [sync-services-developers] MySync public beta Message-ID: <9f38cd3c0511040944n501e8e9fs9449cedd33d9b322@mail.gmail.com> Dear sync-services developers, some time ago I posted to this list asking for beta testers for MySync (and a few more questions besides). MySync provides .mac like syncing between macs on a local network, without .mac I released the MySync public beta yesterday. If you're working with sync services, you might find it interesting to take a look at. You can obtain the public beta from http://www.mildmanneredjanitor.com/mysync/ Thanks very much to everyone on the list for all their help. cheers, Martin From mildm8nnered at gmail.com Sat Nov 5 11:25:16 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Sat Nov 5 11:25:18 2005 Subject: [sync-services-developers] Crazily long filename in ~/Library/Application Support/SyncServices/Local/clientdata Message-ID: <9f38cd3c0511051125j724e5f53q8ae9f045b1febc7c@mail.gmail.com> One of my users is reporting the error message below appearing in their logs. Basically, the Sync Framework seems to be trying to create a file with a crazily long filename in ~/Library/Application Support/SyncServices/Local/clientdata The first filename below is 262 chars. Looking in my own ~/Library/Application Support/SyncServices/Local/clientdata I can see a dozen or so files with similar names (0063006f006d002e006d0 ...), running in length from 66 to 202 chars in length. The user is running Tiger Server, but I'm not sure if this is a factor or not. I haven't had any other similar reports from anyone else (I think I'm up to a hundred or so users at this point, but very few of them, if any, will be running OS X Server). Has anyone else seen anything like this? cheers, Martin 2005-11-06 01:31:57.614 MySync[602] exception while performing selector registerClientWithIdentifier:description:descriptionFilePath: on server: NSInvalidArgumentException [NOTE: this exception originated in the server.] can't create directory /Users/SOMEUSER/Library/Application Support/ SyncServices/Local/clientdata/ 0063006f006d002e006d0069006c0064006d0061006e006e0065007200650064006a0061 006e00690074006f0072002e004d007900530079006e0063002e00720065006500630065 004000720065006500630065002d0061006e0064006500720073006f006e0073002d0070 006f0077006500720062006f006f006b002d00670034002d00310035: File name too long 2005-11-06 01:31:57.618 MySync[602] [syncThread] Controller Uncaught Exception: [NOTE: this exception originated in the server.] can't create directory /Users/SOMEUSER/Library/Application Support/ SyncServices/Local/clientdata/ 0063006f006d002e006d0069006c0064006d0061006e006e0065007200650064006a0061 006e00690074006f0072002e004d007900530079006e0063002e00720065006500630065 004000720065006500630065002d0061006e0064006500720073006f006e0073002d0070 006f0077006500720062006f006f006b002d00670034002d00310035: File name too long From michael at synaptech.com Sun Nov 6 08:54:01 2005 From: michael at synaptech.com (Michael Lafferty) Date: Sun Nov 6 08:54:59 2005 Subject: [sync-services-developers] Crazily long filename in ~/Library/Application Support/SyncServices/Local/clientdata In-Reply-To: <9f38cd3c0511051125j724e5f53q8ae9f045b1febc7c@mail.gmail.com> References: <9f38cd3c0511051125j724e5f53q8ae9f045b1febc7c@mail.gmail.com> Message-ID: <55F8EF04-C125-4D58-8B01-B04AA492E197@synaptech.com> On Nov 5, 2005, at 11:25, Martin Redington wrote: > One of my users is reporting the error message below appearing in > their logs. > > Basically, the Sync Framework seems to be trying to create a file with > a crazily long filename in ~/Library/Application > Support/SyncServices/Local/clientdata > > The first filename below is 262 chars. > > 2005-11-06 01:31:57.614 MySync[602] exception while performing > selector > registerClientWithIdentifier:description:descriptionFilePath: on > server: NSInvalidArgumentException [NOTE: this exception originated > in the server.] > can't create directory /Users/SOMEUSER/Library/Application Support/ > SyncServices/Local/clientdata/ > 0063006f006d002e006d0069006c0064006d0061006e006e0065007200650064006a00 > 61 > 006e00690074006f0072002e004d007900530079006e0063002e007200650065006300 > 65 > 004000720065006500630065002d0061006e0064006500720073006f006e0073002d00 > 70 > 006f0077006500720062006f006f006b002d00670034002d00310035: File name > too long > 2005-11-06 01:31:57.618 MySync[602] [syncThread] Controller Uncaught > Exception: [NOTE: this exception originated in the server.] > can't create directory /Users/SOMEUSER/Library/Application Support/ > SyncServices/Local/clientdata/ > 0063006f006d002e006d0069006c0064006d0061006e006e0065007200650064006a00 > 61 > 006e00690074006f0072002e004d007900530079006e0063002e007200650065006300 > 65 > 004000720065006500630065002d0061006e0064006500720073006f006e0073002d00 > 70 > 006f0077006500720062006f006f006b002d00670034002d00310035: File name > too long An issue resulting in an error nearly identical to this was posted in the Apple Discussion forum. This user was using The Missing Sync version 5, but I have not yet identified if Mac OS X 10.4.x Server was involved, or simply Mac OS X 10.4.3 itself. Inexplicably, he was able to eliminate the error by shortening his HotSync name, a change he made by using The Missing Sync. Michael Lafferty Director of Research & Development SynapTECH (541) 342 7865 voice (541) 342 7867 FAX http://www.synaptech.com From mildm8nnered at gmail.com Sun Nov 6 12:00:40 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Sun Nov 6 12:00:47 2005 Subject: [sync-services-developers] Crazily long filename in ~/Library/Application Support/SyncServices/Local/clientdata In-Reply-To: <55F8EF04-C125-4D58-8B01-B04AA492E197@synaptech.com> References: <9f38cd3c0511051125j724e5f53q8ae9f045b1febc7c@mail.gmail.com> <55F8EF04-C125-4D58-8B01-B04AA492E197@synaptech.com> Message-ID: <9f38cd3c0511061200k2a0ae35bn38599b0d35671a7a@mail.gmail.com> Looking at the range of filename length I have in that dir, I suspect there's some random element to the filename, but its not clear that its one-to-one with (for example) the sync client name. Maybe it was just the fact that he was triggering re-registration(and luckily came out under the limit), rather than shortening the client name, per se. On 11/6/05, Michael Lafferty wrote: > > On Nov 5, 2005, at 11:25, Martin Redington wrote: > > > One of my users is reporting the error message below appearing in > > their logs. > > > > Basically, the Sync Framework seems to be trying to create a file with > > a crazily long filename in ~/Library/Application > > Support/SyncServices/Local/clientdata > > > > The first filename below is 262 chars. > > > > 2005-11-06 01:31:57.614 MySync[602] exception while performing > > selector > > registerClientWithIdentifier:description:descriptionFilePath: on > > server: NSInvalidArgumentException [NOTE: this exception originated > > in the server.] > > can't create directory /Users/SOMEUSER/Library/Application Support/ > > SyncServices/Local/clientdata/ > > 0063006f006d002e006d0069006c0064006d0061006e006e0065007200650064006a00 > > 61 > > 006e00690074006f0072002e004d007900530079006e0063002e007200650065006300 > > 65 > > 004000720065006500630065002d0061006e0064006500720073006f006e0073002d00 > > 70 > > 006f0077006500720062006f006f006b002d00670034002d00310035: File name > > too long > > 2005-11-06 01:31:57.618 MySync[602] [syncThread] Controller Uncaught > > Exception: [NOTE: this exception originated in the server.] > > can't create directory /Users/SOMEUSER/Library/Application Support/ > > SyncServices/Local/clientdata/ > > 0063006f006d002e006d0069006c0064006d0061006e006e0065007200650064006a00 > > 61 > > 006e00690074006f0072002e004d007900530079006e0063002e007200650065006300 > > 65 > > 004000720065006500630065002d0061006e0064006500720073006f006e0073002d00 > > 70 > > 006f0077006500720062006f006f006b002d00670034002d00310035: File name > > too long > > An issue resulting in an error nearly identical to this was posted in > the Apple Discussion forum. This user was using The Missing Sync > version 5, but I have not yet identified if Mac OS X 10.4.x Server > was involved, or simply Mac OS X 10.4.3 itself. Inexplicably, he was > able to eliminate the error by shortening his HotSync name, a change > he made by using The Missing Sync. > > > > Michael Lafferty > Director of Research & Development > SynapTECH > > (541) 342 7865 voice > (541) 342 7867 FAX > > http://www.synaptech.com > > > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers > From mildm8nnered at gmail.com Mon Nov 7 03:50:47 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Mon Nov 7 03:50:49 2005 Subject: [sync-services-developers] com.apple.palm.* datatypes Message-ID: <9f38cd3c0511070350h71aa9c26g6feb1eb6e4f4fec5@mail.gmail.com> My application (MySync) uses Apple's sync framework to sync macs. At startup, it checks the registered datatypes, and generates a plist for sync client registration. One of my users is reporting the following problem. On his primary machine, there are some com.apple.palm.* datatypes registered under the com.apple.* datatypes (contacts and calendars I believe). On his secondary machine, these com.apple.palm.* datatypes are not registered. Records pulled from the truth store on his primary cannot be pushed to the truth store on his secondary, because the com.apple.palm.* datatypes are not registered on the secondary. The user has the Missing Sync installed on the primary, but not on the secondary. He may have had old versions of the Palm software installed on both machines. Does anyone know where the com.apple.palm.* datatypes come from? Are they installed by the Missing Sync? How can he get them registered on his secondary? cheers, Martin From bcriscuolo at markspace.com Mon Nov 7 05:04:51 2005 From: bcriscuolo at markspace.com (Brian M. Criscuolo) Date: Mon Nov 7 05:05:05 2005 Subject: [sync-services-developers] com.apple.palm.* datatypes In-Reply-To: <9f38cd3c0511070350h71aa9c26g6feb1eb6e4f4fec5@mail.gmail.com> References: <9f38cd3c0511070350h71aa9c26g6feb1eb6e4f4fec5@mail.gmail.com> Message-ID: Martin, They are definitely *not* coming from Missing Sync. I suspect that they are the iSync identifiers, but since I don't have Apple's Palm conduits installed I can't say for sure. b On Nov 7, 2005, at 6:50 AM, Martin Redington wrote: > My application (MySync) uses Apple's sync framework to sync macs. > > At startup, it checks the registered datatypes, and generates a plist > for sync client registration. > > One of my users is reporting the following problem. > > On his primary machine, there are some com.apple.palm.* datatypes > registered under the com.apple.* datatypes (contacts and calendars I > believe). > > On his secondary machine, these com.apple.palm.* datatypes are not > registered. > > Records pulled from the truth store on his primary cannot be pushed to > the truth store on his secondary, because the com.apple.palm.* > datatypes are not registered on the secondary. > > The user has the Missing Sync installed on the primary, but not on the > secondary. He may have had old versions of the Palm software installed > on both machines. > > Does anyone know where the com.apple.palm.* datatypes come from? > > Are they installed by the Missing Sync? > > How can he get them registered on his secondary? > > cheers, > Martin > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers ******************** Brian M. Criscuolo Senior Software Engineer Mark/Space, Inc. bcriscuolo@markspace.com From bcriscuolo at markspace.com Mon Nov 7 05:55:15 2005 From: bcriscuolo at markspace.com (Brian M. Criscuolo) Date: Mon Nov 7 05:55:28 2005 Subject: [sync-services-developers] MacDevCenter article on Sync Services Message-ID: <4AFF87CC-50EB-4090-B13A-2A01BC400F8D@markspace.com> Hi all, MacDevCenter has an article up about Sync Services in Tiger: Is the author, Mary, on this list? b ******************** Brian M. Criscuolo Senior Software Engineer Mark/Space, Inc. bcriscuolo@markspace.com From mildm8nnered at gmail.com Mon Nov 7 07:34:48 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Mon Nov 7 07:34:50 2005 Subject: [sync-services-developers] com.apple.palm.* datatypes In-Reply-To: References: <9f38cd3c0511070350h71aa9c26g6feb1eb6e4f4fec5@mail.gmail.com> Message-ID: <9f38cd3c0511070734t55c5351p6ac0d319076fb5b0@mail.gmail.com> Thanks Brian, I'll try and see what else the user has installed. cheers, m. On 11/7/05, Brian M. Criscuolo wrote: > Martin, > > They are definitely *not* coming from Missing Sync. I suspect that > they are the iSync identifiers, but since I don't have Apple's Palm > conduits installed I can't say for sure. > > b > > On Nov 7, 2005, at 6:50 AM, Martin Redington wrote: > > > My application (MySync) uses Apple's sync framework to sync macs. > > > > At startup, it checks the registered datatypes, and generates a plist > > for sync client registration. > > > > One of my users is reporting the following problem. > > > > On his primary machine, there are some com.apple.palm.* datatypes > > registered under the com.apple.* datatypes (contacts and calendars I > > believe). > > > > On his secondary machine, these com.apple.palm.* datatypes are not > > registered. > > > > Records pulled from the truth store on his primary cannot be pushed to > > the truth store on his secondary, because the com.apple.palm.* > > datatypes are not registered on the secondary. > > > > The user has the Missing Sync installed on the primary, but not on the > > secondary. He may have had old versions of the Palm software installed > > on both machines. > > > > Does anyone know where the com.apple.palm.* datatypes come from? > > > > Are they installed by the Missing Sync? > > > > How can he get them registered on his secondary? > > > > cheers, > > Martin > > > > _______________________________________________ > > sync-services-developers mailing list > > sync-services-developers@lists.markspace.com > > http://lists.markspace.com/mailman/listinfo/sync-services-developers > > > ******************** > Brian M. Criscuolo > Senior Software Engineer > Mark/Space, Inc. > bcriscuolo@markspace.com > > > > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers > From nick at chronosnet.com Thu Nov 10 08:53:31 2005 From: nick at chronosnet.com (Nick Zitzmann) Date: Thu Nov 10 08:54:04 2005 Subject: [sync-services-developers] Sync Services not taking "no" for an answer? Message-ID: <7FB5271C-168F-48E5-886A-5EFC12A335CE@chronosnet.com> I noticed recently that my sync client lately has been taking an abnormally long amount of time to complete its work, almost as if the sync client was performing a slow sync every single time. So I took a look at the call history in Syncrospector, and it was performing a fast sync, but the sync client was continuously pulling changes that the client refused earlier, and these changes are delete changes that have not been modified between sessions. Two back-to-back call histories show that the sync client rejected the same change twice, but Sync Services keeps pushing it. All that the sync client is doing is: 1. Taking the change obtained from the change enumerator 2. Calling -clientRefusedChangesForRecordWithIdentifier: with the sync client's version of the record identifier (which was taken from the change from #1) 3. After taking in all changes from all of the change enumerators, calling -clientCommittedAcceptedChanges, followed by -finishSyncing According to the documentation, that should be all, but apparently it's not. It doesn't pull committed changes twice, so why is it pulling refused changes twice? Has anyone else seen this before, and if so, what did you do to get Sync Services to take "no" for an answer? Nick Zitzmann From bcriscuolo at markspace.com Fri Nov 11 08:31:21 2005 From: bcriscuolo at markspace.com (Brian M. Criscuolo) Date: Fri Nov 11 08:31:30 2005 Subject: [sync-services-developers] Sync Services not taking "no" for an answer? In-Reply-To: <7FB5271C-168F-48E5-886A-5EFC12A335CE@chronosnet.com> References: <7FB5271C-168F-48E5-886A-5EFC12A335CE@chronosnet.com> Message-ID: Nick, Welcome to the changes in Sync Services brought on with 10.4.3! This didn't happen before the update, and it does now. Records that have been refused are now being pushed every time for me as well - in a shipping product! Its driving me absolutely crazy. In my case, I refused records (events) from a read-only calendar. The calendar has not been refreshed, so the events haven't changed. I guess both you and I should submit this to Apple so that they can get it fixed at some point. For now, I've taken to storing an array of refused records in the client data and checking each one against that. Totally sucks. There are some folks from Apple on this list - would any care to elaborate on the various changes in Sync Services with the 10.4.3 update? We really need to know what has been changed. b On Nov 10, 2005, at 11:53 AM, Nick Zitzmann wrote: > I noticed recently that my sync client lately has been taking an > abnormally long amount of time to complete its work, almost as if > the sync client was performing a slow sync every single time. So I > took a look at the call history in Syncrospector, and it was > performing a fast sync, but the sync client was continuously > pulling changes that the client refused earlier, and these changes > are delete changes that have not been modified between sessions. > Two back-to-back call histories show that the sync client rejected > the same change twice, but Sync Services keeps pushing it. > > All that the sync client is doing is: > 1. Taking the change obtained from the change enumerator > 2. Calling -clientRefusedChangesForRecordWithIdentifier: with the > sync client's version of the record identifier (which was taken > from the change from #1) > 3. After taking in all changes from all of the change enumerators, > calling -clientCommittedAcceptedChanges, followed by -finishSyncing > > According to the documentation, that should be all, but apparently > it's not. It doesn't pull committed changes twice, so why is it > pulling refused changes twice? Has anyone else seen this before, > and if so, what did you do to get Sync Services to take "no" for an > answer? > > Nick Zitzmann > > > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers ******************** Brian M. Criscuolo Senior Software Engineer Mark/Space, Inc. bcriscuolo@markspace.com From nick at chronosnet.com Fri Nov 11 09:20:47 2005 From: nick at chronosnet.com (Nick Zitzmann) Date: Fri Nov 11 09:21:10 2005 Subject: [sync-services-developers] Sync Services not taking "no" for an answer? In-Reply-To: References: <7FB5271C-168F-48E5-886A-5EFC12A335CE@chronosnet.com> Message-ID: <7B07402E-79FB-4DE2-83DD-24174180672F@chronosnet.com> On Nov 11, 2005, at 9:31 AM, Brian M. Criscuolo wrote: > I guess both you and I should submit this to Apple so that they can > get it fixed at some point. For now, I've taken to storing an array > of refused records in the client data and checking each one against > that. Totally sucks. Filed: #4339476. Please file a duplicate report; that way this may get noticed. I'm thinking of changing my client so that it only refuses add & modify changes, while delete changes that would normally be refused are actually marked as accepted. That would work around the problem for now, unless I'm missing something... Nick Zitzmann From juds at letterrip.com Fri Nov 11 09:58:25 2005 From: juds at letterrip.com (Jud Spencer) Date: Fri Nov 11 09:58:20 2005 Subject: [sync-services-developers] Sync Services not taking "no" for an answer? In-Reply-To: <7B07402E-79FB-4DE2-83DD-24174180672F@chronosnet.com> Message-ID: On 11/11/05 9:20:47 AM, "Nick Zitzmann" wrote: > I'm thinking of changing my client so that it only refuses add & > modify changes, while delete changes that would normally be refused > are actually marked as accepted. That would work around the problem > for now, unless I'm missing something... Why would you refuse a delete? From nick at chronosnet.com Fri Nov 11 10:59:48 2005 From: nick at chronosnet.com (Nick Zitzmann) Date: Fri Nov 11 11:00:20 2005 Subject: [sync-services-developers] Sync Services not taking "no" for an answer? In-Reply-To: References: Message-ID: On Nov 11, 2005, at 10:58 AM, Jud Spencer wrote: >> I'm thinking of changing my client so that it only refuses add & >> modify changes, while delete changes that would normally be refused >> are actually marked as accepted. That would work around the problem >> for now, unless I'm missing something... > > Why would you refuse a delete? The sync client description uses many entities with cascading one-to- many relationships, and when the engine pushes a delete change for a parent instance of an entity, previously it was refusing the delete changes on the child entities that were now irrelevant. But since - clientRefusedChangesForRecordWithIdentifier: apparently does nothing, it was getting bombarded with delete changes that were previously refused, which made sync sessions slower (and call histories longer) as more objects were deleted. That's why I originally brought this up... Nick Zitzmann From juds at letterrip.com Fri Nov 11 11:17:11 2005 From: juds at letterrip.com (Jud Spencer) Date: Fri Nov 11 11:17:22 2005 Subject: [sync-services-developers] Sync Services not taking "no" for an answer? In-Reply-To: Message-ID: On 11/11/05 10:59:48 AM, "Nick Zitzmann" wrote: > The sync client description uses many entities with cascading one-to- > many relationships, and when the engine pushes a delete change for a > parent instance of an entity, previously it was refusing the delete > changes on the child entities that were now irrelevant. But since - > clientRefusedChangesForRecordWithIdentifier: apparently does nothing, > it was getting bombarded with delete changes that were previously > refused, which made sync sessions slower (and call histories longer) > as more objects were deleted. That's why I originally brought this up... OK, but why were you refusing the delete changes if they were irrelevant? Just accept them and the problem goes away. I was under the impression that there was no good reason to refuse a delete. It's not going to cascade further as a result. From mildm8nnered at gmail.com Wed Nov 16 21:01:58 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Wed Nov 16 21:02:01 2005 Subject: [sync-services-developers] syncing only some Address Book groups Message-ID: <9f38cd3c0511162101y29a939d2g80eaba57067a83b7@mail.gmail.com> This is probably the most frequent feature request from my users. I hadn't really looked into filters until now, but my first naive idea, having looked at the docs, is to grab a truth snapshot of the relevant data, and to have my filter inspect the pulled records, and walk up to the parent, to determine whether they belong to a group that should be synced or not. My understanding is that filter only work when pulling, so I would add similar logic to the pushing side of my app (although actually I might not need to, as I will already have filtered them out when pulling [from a different truth store]). Does this sound viable, or am I on the wrong track? Has anyone else tried to do this, or anything like it? What happens if someone enables a group that has previously been unsynced. Will they get all the updates that they're missing (i.e. that they filtered on previous syncs)? cheers, Martin From btruex at markspace.com Thu Nov 17 07:25:48 2005 From: btruex at markspace.com (Brent Truex) Date: Thu Nov 17 07:26:02 2005 Subject: [sync-services-developers] syncing only some Address Book groups In-Reply-To: <9f38cd3c0511162101y29a939d2g80eaba57067a83b7@mail.gmail.com> References: <9f38cd3c0511162101y29a939d2g80eaba57067a83b7@mail.gmail.com> Message-ID: Martin, Have you looked at the ISyncFilter class? It saves you the trouble of managing the record comparison, and is quick to implement if you can live with the way it works. > What happens if someone enables a group that has previously been > unsynced. Will they get all the updates that they're missing (i.e. > that they filtered on previous syncs)? > With ISyncFilters, yes. Conversely, removing a group from the sync list will cause all the entries for that group to be deleted from the client. - Brent --------------------------------------------------------------- Brent Truex btruex@markspace.com From mildm8nnered at gmail.com Thu Nov 17 12:09:39 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Thu Nov 17 12:37:49 2005 Subject: [sync-services-developers] syncing only some Address Book groups In-Reply-To: References: <9f38cd3c0511162101y29a939d2g80eaba57067a83b7@mail.gmail.com> Message-ID: <9f38cd3c0511171209v1ec3d37cpa0a541c9bcf50514@mail.gmail.com> Thanks Brent, sorry, I should have been clearer - I meant in the context of using ISyncFilter. The reason I thought I might need to grab a snapshot was that, if, for example, a phone number is updated, won't I get passed the phone number record in the ISyncFilter filter method, whereas to determine whether to pass this on or not, I need to look at the parent contact record. I'll have a play around anyway ... cheers, m. On 11/17/05, Brent Truex wrote: > Martin, > > Have you looked at the ISyncFilter class? It saves you the trouble of > managing the record comparison, and is quick to implement if you can > live with the way it works. > > > What happens if someone enables a group that has previously been > > unsynced. Will they get all the updates that they're missing (i.e. > > that they filtered on previous syncs)? > > > > With ISyncFilters, yes. Conversely, removing a group from the sync > list will cause all the entries for that group to be deleted from the > client. > > - Brent > > > > > --------------------------------------------------------------- > > Brent Truex > btruex@markspace.com > > > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers > From btruex at markspace.com Thu Nov 17 14:51:46 2005 From: btruex at markspace.com (Brent Truex) Date: Thu Nov 17 14:51:51 2005 Subject: [sync-services-developers] syncing only some Address Book groups In-Reply-To: <9f38cd3c0511171209v1ec3d37cpa0a541c9bcf50514@mail.gmail.com> References: <9f38cd3c0511162101y29a939d2g80eaba57067a83b7@mail.gmail.com> <9f38cd3c0511171209v1ec3d37cpa0a541c9bcf50514@mail.gmail.com> Message-ID: On Nov 17, 2005, at 3:09 PM, Martin Redington wrote: > The reason I thought I might need to grab a snapshot was that, if, for > example, a phone number is updated, won't I get passed the phone > number record in the ISyncFilter filter method, whereas to determine > whether to pass this on or not, I need to look at the parent contact > record. My session code is set up to pull from the parent entity first. Then if the client gets child records it doesn't have a parent for, it refuses the change. I had problems using a session snapshot inside a filter, but that was before Tiger was released, so it may work correctly now.. - Brent --------------------------------- Brent Truex btruex@markspace.com -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.markspace.com/pipermail/sync-services-developers/attachments/20051117/d2964d3a/attachment.html From mildm8nnered at gmail.com Fri Nov 18 02:59:12 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Fri Nov 18 02:59:15 2005 Subject: [sync-services-developers] New Syncrospector? In-Reply-To: References: <22901BDD-27FE-49C9-A947-F9D09D3E3132@markspace.com> Message-ID: <9f38cd3c0511180259r6b239dd7t55979261dc84fecc@mail.gmail.com> Does anyone know where the new one is on ADC. I can't find it on the site. cheers, m. On 11/3/05, Eric Seidel wrote: > I believe there is a newer one available from ADC. > > -eric > > On 02.11.2005, at 11:52, Scott Gruby wrote: > > > Does anyone know if there is a Syncrospector later than 0.5 that > > works with 10.4.3? It launches, but doesn't work and dumps the > > following in Console: > > > > 2005-11-02 11:50:36.877 Syncrospector[23219] *** -[NSCFArray > > isd_enumerateObjectsAsArgumentToFunction:]: selector not recognized > > [self = 0x6079d30] > > 2005-11-02 11:50:37.125 Syncrospector[23219] [404340] ##### RAISE > > NSInvalidArgumentException: *** -[NSCFArray > > isd_enumerateObjectsAsArgumentToFunction:]: selector not recognized > > [self = 0x6079d30] > > > > > > -- > > Scott Gruby > > Lead Engineer > > Mark/Space, Inc. > > > > > > Please visit for assistance > > with Mark/Space products. > > > > _______________________________________________ > > sync-services-developers mailing list > > sync-services-developers@lists.markspace.com > > http://lists.markspace.com/mailman/listinfo/sync-services-developers > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers > From lwdupont at gmail.com Fri Nov 18 08:20:10 2005 From: lwdupont at gmail.com (Lucien W. Dupont) Date: Fri Nov 18 08:20:18 2005 Subject: [sync-services-developers] New Syncrospector? In-Reply-To: <9f38cd3c0511180259r6b239dd7t55979261dc84fecc@mail.gmail.com> References: <22901BDD-27FE-49C9-A947-F9D09D3E3132@markspace.com> <9f38cd3c0511180259r6b239dd7t55979261dc84fecc@mail.gmail.com> Message-ID: <84BAB5BB-8A8C-40C2-A137-4C1A77CF3C97@gmail.com> There is a new one that was just posted Wednesday, click the Download link after you log in, and scroll down a bit, you should see it. Syncrospector 2.1 (v186) On Nov 18, 2005, at 2:59 AM, Martin Redington wrote: > Does anyone know where the new one is on ADC. > > I can't find it on the site. > > cheers, > m. > > On 11/3/05, Eric Seidel wrote: >> I believe there is a newer one available from ADC. >> >> -eric >> From mildm8nnered at gmail.com Fri Nov 18 11:44:56 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Fri Nov 18 11:45:01 2005 Subject: [sync-services-developers] New Syncrospector? In-Reply-To: <84BAB5BB-8A8C-40C2-A137-4C1A77CF3C97@gmail.com> References: <22901BDD-27FE-49C9-A947-F9D09D3E3132@markspace.com> <9f38cd3c0511180259r6b239dd7t55979261dc84fecc@mail.gmail.com> <84BAB5BB-8A8C-40C2-A137-4C1A77CF3C97@gmail.com> Message-ID: <9f38cd3c0511181144g54432bb3qd4fa09e56180824b@mail.gmail.com> I looked there, and searched all of the download pages - no trrace of it :-( I'm not a paying ADC member though - I'vejust got a freebie login. I mailed ADC, but no reply so far. On 11/18/05, Lucien W. Dupont wrote: > There is a new one that was just posted Wednesday, click the Download > link after you log in, and scroll down a bit, you should see it. > > Syncrospector 2.1 (v186) > > > On Nov 18, 2005, at 2:59 AM, Martin Redington wrote: > > > Does anyone know where the new one is on ADC. > > > > I can't find it on the site. > > > > cheers, > > m. > > > > On 11/3/05, Eric Seidel wrote: > >> I believe there is a newer one available from ADC. > >> > >> -eric > >> > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers > From bcriscuolo at markspace.com Fri Nov 18 11:50:42 2005 From: bcriscuolo at markspace.com (Brian M. Criscuolo) Date: Fri Nov 18 11:50:43 2005 Subject: [sync-services-developers] New Syncrospector? In-Reply-To: <9f38cd3c0511181144g54432bb3qd4fa09e56180824b@mail.gmail.com> References: <22901BDD-27FE-49C9-A947-F9D09D3E3132@markspace.com> <9f38cd3c0511180259r6b239dd7t55979261dc84fecc@mail.gmail.com> <84BAB5BB-8A8C-40C2-A137-4C1A77CF3C97@gmail.com> <9f38cd3c0511181144g54432bb3qd4fa09e56180824b@mail.gmail.com> Message-ID: <810D9967-55C4-40F2-BF54-9E82F46D66A8@markspace.com> Very strange - I'm a freebie member as well! And I see it on the main downloads link. Where are you located? Maybe its a country by country thing? b On Nov 18, 2005, at 2:44 PM, Martin Redington wrote: > I looked there, and searched all of the download pages - no trrace of > it :-( I'm not a paying ADC member though - I'vejust got a freebie > login. > > I mailed ADC, but no reply so far. > > On 11/18/05, Lucien W. Dupont wrote: >> There is a new one that was just posted Wednesday, click the Download >> link after you log in, and scroll down a bit, you should see it. >> >> Syncrospector 2.1 (v186) >> >> >> On Nov 18, 2005, at 2:59 AM, Martin Redington wrote: >> >>> Does anyone know where the new one is on ADC. >>> >>> I can't find it on the site. >>> >>> cheers, >>> m. >>> >>> On 11/3/05, Eric Seidel wrote: >>>> I believe there is a newer one available from ADC. >>>> >>>> -eric >>>> >> _______________________________________________ >> sync-services-developers mailing list >> sync-services-developers@lists.markspace.com >> http://lists.markspace.com/mailman/listinfo/sync-services-developers >> > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers ******************** Brian M. Criscuolo Senior Software Engineer Mark/Space, Inc. bcriscuolo@markspace.com From mildm8nnered at gmail.com Fri Nov 18 17:03:50 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Fri Nov 18 17:03:55 2005 Subject: [sync-services-developers] New Syncrospector? In-Reply-To: <810D9967-55C4-40F2-BF54-9E82F46D66A8@markspace.com> References: <22901BDD-27FE-49C9-A947-F9D09D3E3132@markspace.com> <9f38cd3c0511180259r6b239dd7t55979261dc84fecc@mail.gmail.com> <84BAB5BB-8A8C-40C2-A137-4C1A77CF3C97@gmail.com> <9f38cd3c0511181144g54432bb3qd4fa09e56180824b@mail.gmail.com> <810D9967-55C4-40F2-BF54-9E82F46D66A8@markspace.com> Message-ID: <9f38cd3c0511181703j71f675cv8a1306cb02c9a750@mail.gmail.com> I'm in the UK (well, Paris right now, but I was in the UK yesterday). Apple wrote back saying Syncrospector is in the StickiesExample, but that's just the old one (or was yesterday). >From what you guys are saying there is now a separate Syncrospector link. I'll get back to Apple... On 11/18/05, Brian M. Criscuolo wrote: > Very strange - I'm a freebie member as well! And I see it on the main > downloads link. > > Where are you located? Maybe its a country by country thing? > b > > On Nov 18, 2005, at 2:44 PM, Martin Redington wrote: > > > I looked there, and searched all of the download pages - no trrace of > > it :-( I'm not a paying ADC member though - I'vejust got a freebie > > login. > > > > I mailed ADC, but no reply so far. > > > > On 11/18/05, Lucien W. Dupont wrote: > >> There is a new one that was just posted Wednesday, click the Download > >> link after you log in, and scroll down a bit, you should see it. > >> > >> Syncrospector 2.1 (v186) > >> > >> > >> On Nov 18, 2005, at 2:59 AM, Martin Redington wrote: > >> > >>> Does anyone know where the new one is on ADC. > >>> > >>> I can't find it on the site. > >>> > >>> cheers, > >>> m. > >>> > >>> On 11/3/05, Eric Seidel wrote: > >>>> I believe there is a newer one available from ADC. > >>>> > >>>> -eric > >>>> > >> _______________________________________________ > >> sync-services-developers mailing list > >> sync-services-developers@lists.markspace.com > >> http://lists.markspace.com/mailman/listinfo/sync-services-developers > >> > > > > _______________________________________________ > > sync-services-developers mailing list > > sync-services-developers@lists.markspace.com > > http://lists.markspace.com/mailman/listinfo/sync-services-developers > > > ******************** > Brian M. Criscuolo > Senior Software Engineer > Mark/Space, Inc. > bcriscuolo@markspace.com > > > > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers > From mildm8nnered at gmail.com Sun Nov 20 03:04:00 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Sun Nov 20 03:04:04 2005 Subject: [sync-services-developers] Sync Services deleting all data unexpectedly. Message-ID: <9f38cd3c0511200304i1f74c807m9aab84be000d23e6@mail.gmail.com> Has anyone else seen Sync Services delete all their data unexpectedly? I've had three user reports of this now - the symptoms seem to be that one machine suddenly decides that it doesn't need its data any more, and then propagates the deletes via sync services to other sync clients, but I'm still collecting more detailed diagnostics to send to Apple. I've heard some anecdotal reports of data loss when people moved to Tiger, but these were all end-user reports IIRC. I always sync in default (fast sync) mode, so its hard to see how all of the data could get trashed ... cheers, Martin From bcriscuolo at markspace.com Mon Nov 21 05:38:54 2005 From: bcriscuolo at markspace.com (Brian M. Criscuolo) Date: Mon Nov 21 05:38:59 2005 Subject: [sync-services-developers] Sync Services deleting all data unexpectedly. In-Reply-To: <9f38cd3c0511200304i1f74c807m9aab84be000d23e6@mail.gmail.com> References: <9f38cd3c0511200304i1f74c807m9aab84be000d23e6@mail.gmail.com> Message-ID: <729683CC-AEE5-4E29-930D-201C7D1BD2EB@markspace.com> Martin, I've seen it in slow sync mode a few times. If the entire data set isn't pushed into the truth whatever is missing is flagged as a delete. I've quickly deleted entire groups of calendars in iCal that way - along with all the events! I've heard the anecdotal reports as well. Here's one from a customer of ours. I'm sure that we're involved in some way, but this isn't something that I can reproduce: "I decided to try a little more testing (nothing better to do on a sunny Saturday afternoon apparently). On the handheld I created a fake contact and did a (contacts only) sync. I was initially happy because the fake contact was in my Address Book on the Mac! Yikes, I then noticed that most (but not all) of the groups that I had created (weeks ago on the Mac) were GONE from Address Book. So I reverted to the Address Book database that I had saved this morning before using 5.0.2. Sure enough, they were still there. So I synced again, assuming that the contact that had been added to the Palm would once again be added to the Address Book and I'd watch and see what happened to my groups. The groups were fine this time (what the heck is going on?) but the contact on the Palm was not added to Addy Book. Hmmm, so I checked the Palm, it wasn't there anymore either. OK...So I created the Fake Contact on the Palm again and synced again. Now the groups are still there in the Addy Book and so is the Fake Contact. So I deleted the Fake Contact on the Mac and synced again. Now it's gone on the Mac and on the Palm and my groups seem OK. So everything is now OK....right? I wish I knew." I'll try to pay more attention to things like this and maybe we can work out a pattern. b On Nov 20, 2005, at 6:04 AM, Martin Redington wrote: > Has anyone else seen Sync Services delete all their data unexpectedly? > > I've had three user reports of this now - the symptoms seem to be that > one machine suddenly decides that it doesn't need its data any more, > and then propagates the deletes via sync services to other sync > clients, but I'm still collecting more detailed diagnostics to send to > Apple. > > I've heard some anecdotal reports of data loss when people moved to > Tiger, but these were all end-user reports IIRC. > > I always sync in default (fast sync) mode, so its hard to see how all > of the data could get trashed ... > > cheers, > Martin > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers ******************** Brian M. Criscuolo Senior Software Engineer Mark/Space, Inc. bcriscuolo@markspace.com From mildm8nnered at gmail.com Mon Nov 21 09:03:45 2005 From: mildm8nnered at gmail.com (Martin Redington) Date: Mon Nov 21 09:03:49 2005 Subject: [sync-services-developers] Sync Services deleting all data unexpectedly. In-Reply-To: <729683CC-AEE5-4E29-930D-201C7D1BD2EB@markspace.com> References: <9f38cd3c0511200304i1f74c807m9aab84be000d23e6@mail.gmail.com> <729683CC-AEE5-4E29-930D-201C7D1BD2EB@markspace.com> Message-ID: <9f38cd3c0511210903p63041e39sa83f7a66e8421615@mail.gmail.com> I think I've got some good data on this, and, if it keep happening (grrr) I'll be collecting even better data with my next release. When I get a chance, I'll sit down and analyse what I've got so far. cheers, m. On 11/21/05, Brian M. Criscuolo wrote: > Martin, > > I've seen it in slow sync mode a few times. If the entire data set > isn't pushed into the truth whatever is missing is flagged as a > delete. I've quickly deleted entire groups of calendars in iCal that > way - along with all the events! > > I've heard the anecdotal reports as well. Here's one from a customer > of ours. I'm sure that we're involved in some way, but this isn't > something that I can reproduce: > > "I decided to try a little more testing (nothing better to do on a > sunny Saturday afternoon apparently). On the handheld I created a > fake contact and did a (contacts only) sync. I was initially happy > because the fake contact was in my Address Book on the Mac! Yikes, I > then noticed that most (but not all) of the groups that I had created > (weeks ago on the Mac) were GONE from Address Book. So I reverted to > the Address Book database that I had saved this morning before using > 5.0.2. Sure enough, they were still there. So I synced again, > assuming that the contact that had been added to the Palm would once > again be added to the Address Book and I'd watch and see what > happened to my groups. The groups were fine this time (what the heck > is going on?) but the contact on the Palm was not added to Addy Book. > Hmmm, so I checked the Palm, it wasn't there anymore either. OK...So > I created the Fake Contact on the Palm again and synced again. Now > the groups are still there in the Addy Book and so is the Fake > Contact. So I deleted the Fake Contact on the Mac and synced again. > Now it's gone on the Mac and on the Palm and my groups seem OK. So > everything is now OK....right? I wish I knew." > > I'll try to pay more attention to things like this and maybe we can > work out a pattern. > > b > > On Nov 20, 2005, at 6:04 AM, Martin Redington wrote: > > > Has anyone else seen Sync Services delete all their data unexpectedly? > > > > I've had three user reports of this now - the symptoms seem to be that > > one machine suddenly decides that it doesn't need its data any more, > > and then propagates the deletes via sync services to other sync > > clients, but I'm still collecting more detailed diagnostics to send to > > Apple. > > > > I've heard some anecdotal reports of data loss when people moved to > > Tiger, but these were all end-user reports IIRC. > > > > I always sync in default (fast sync) mode, so its hard to see how all > > of the data could get trashed ... > > > > cheers, > > Martin > > > > _______________________________________________ > > sync-services-developers mailing list > > sync-services-developers@lists.markspace.com > > http://lists.markspace.com/mailman/listinfo/sync-services-developers > > > ******************** > Brian M. Criscuolo > Senior Software Engineer > Mark/Space, Inc. > bcriscuolo@markspace.com > > > > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers > From lwdupont at gmail.com Mon Nov 21 13:39:58 2005 From: lwdupont at gmail.com (Lucien Dupont) Date: Mon Nov 21 13:40:06 2005 Subject: [sync-services-developers] New Syncrospector? In-Reply-To: <9f38cd3c0511181703j71f675cv8a1306cb02c9a750@mail.gmail.com> References: <22901BDD-27FE-49C9-A947-F9D09D3E3132@markspace.com> <9f38cd3c0511180259r6b239dd7t55979261dc84fecc@mail.gmail.com> <84BAB5BB-8A8C-40C2-A137-4C1A77CF3C97@gmail.com> <9f38cd3c0511181144g54432bb3qd4fa09e56180824b@mail.gmail.com> <810D9967-55C4-40F2-BF54-9E82F46D66A8@markspace.com> <9f38cd3c0511181703j71f675cv8a1306cb02c9a750@mail.gmail.com> Message-ID: <55c896120511211339x4084fae6xaa5e9dae419669e3@mail.gmail.com> It might be because I or others have a Software Seed Key (which gives people access to not yet released software).. I currently see it by logging in, then clicking on Downloads, then Developer Tools. Lucien On 11/18/05, Martin Redington wrote: > I'm in the UK (well, Paris right now, but I was in the UK yesterday). > > Apple wrote back saying Syncrospector is in the StickiesExample, but > that's just the old one (or was yesterday). > > >From what you guys are saying there is now a separate Syncrospector link. > > I'll get back to Apple... > > On 11/18/05, Brian M. Criscuolo wrote: > > Very strange - I'm a freebie member as well! And I see it on the main > > downloads link. > > > > Where are you located? Maybe its a country by country thing? > > b > > > > On Nov 18, 2005, at 2:44 PM, Martin Redington wrote: > > > > > I looked there, and searched all of the download pages - no trrace of > > > it :-( I'm not a paying ADC member though - I'vejust got a freebie > > > login. > > > > > > I mailed ADC, but no reply so far. > > > > > > On 11/18/05, Lucien W. Dupont wrote: > > >> There is a new one that was just posted Wednesday, click the Download > > >> link after you log in, and scroll down a bit, you should see it. > > >> > > >> Syncrospector 2.1 (v186) > > >> > > >> > > >> On Nov 18, 2005, at 2:59 AM, Martin Redington wrote: > > >> > > >>> Does anyone know where the new one is on ADC. > > >>> > > >>> I can't find it on the site. > > >>> > > >>> cheers, > > >>> m. > > >>> > > >>> On 11/3/05, Eric Seidel wrote: > > >>>> I believe there is a newer one available from ADC. > > >>>> > > >>>> -eric > > >>>> > > >> _______________________________________________ > > >> sync-services-developers mailing list > > >> sync-services-developers@lists.markspace.com > > >> http://lists.markspace.com/mailman/listinfo/sync-services-developers > > >> > > > > > > _______________________________________________ > > > sync-services-developers mailing list > > > sync-services-developers@lists.markspace.com > > > http://lists.markspace.com/mailman/listinfo/sync-services-developers > > > > > > ******************** > > Brian M. Criscuolo > > Senior Software Engineer > > Mark/Space, Inc. > > bcriscuolo@markspace.com > > > > > > > > > > _______________________________________________ > > sync-services-developers mailing list > > sync-services-developers@lists.markspace.com > > http://lists.markspace.com/mailman/listinfo/sync-services-developers > > > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers > From bcriscuolo at markspace.com Tue Nov 22 08:07:02 2005 From: bcriscuolo at markspace.com (Brian M. Criscuolo) Date: Tue Nov 22 08:07:11 2005 Subject: [sync-services-developers] Record Pushing Issue Message-ID: <3537700B-6F00-4AC4-B467-7AB4DFEA7B96@markspace.com> Hey everyone, I'm running across a really strange issue that did not happen on 10.4.2, but now does on 10.4.3. When I push records from a device (Palm) into the truth, I don't necessarily have *all* the data that makes up an entity. For instance, the Palm address book only supports three street address fields. That is all I can push. If the truth contains 5 street addresses, three of them are reflected on the device. Previously, I would push the street address record containing the changes into the truth and whatever change would be merged in (updating one of the 5 records, for example). After the sync, all five records existed in the truth with the changes pushed. Now, on 10.4.3, if I only push three records, the other two records are DELETED! This is a change from the previously existing behavior, and I'm not sure why. Does this statement have anything to do with it? "ISyncSession can push "clear" deltas for properties that were added when a client accepted a record but did not subsequently provide. For example, if a client accepts record A and adds a property Q in the formatted record, then pushes a change to X that does not include Q, ISyncSession should not issue a "clear" for Q." Is it because I'm accepting all addresses when pulling the original record but only providing a subset of them when pushing a change back, as outlined above? Did Apple change the behavior of: "SyncSession should not issue a "clear" for Q." but not document it? There were numerous Sync Services changes in 10.4.3 but none were described in any detail. We need some detail about them, as shipping code behaves differently. Does anyone have any thoughts or suggestions to this perplexing issue? Thanks, b ******************** Brian M. Criscuolo Senior Software Engineer Mark/Space, Inc. bcriscuolo@markspace.com From juds at letterrip.com Tue Nov 22 09:00:11 2005 From: juds at letterrip.com (Jud Spencer) Date: Tue Nov 22 09:00:28 2005 Subject: [sync-services-developers] Record Pushing Issue In-Reply-To: <3537700B-6F00-4AC4-B467-7AB4DFEA7B96@markspace.com> Message-ID: On 11/22/05 8:07:02 AM, "Brian M. Criscuolo" wrote: > Is it because I'm accepting all addresses when pulling the original > record but only providing a subset of them when pushing a change > back, as outlined above? Did Apple change the behavior of: > "SyncSession should not issue a "clear" for Q." but not document it? > > There were numerous Sync Services changes in 10.4.3 but none were > described in any detail. We need some detail about them, as shipping > code behaves differently. > > Does anyone have any thoughts or suggestions to this perplexing issue? I'm new to the sync services game. I essentially started with 10.4.3. I have found that in a case where you are pushing a subset of entities related to a parent, that you need to provide the ID's of all the sub-entities in the parent, not just the ones that are being pushed. What we are doing in our app is keeping track of the entities when we have a schema incompatibility, and then pushing them on any update. For example, our app only supports two addresses. When we pull a record with more than two, we keep track of the ids of the ones that we accepted but didn't write to our database. When we later push the record back to the truth, we write the id's of the addresses that we write as well as the ids of the ones that we don't write into the street addresses property in .Contact. This works for us. Jud From bcriscuolo at markspace.com Tue Nov 22 12:32:51 2005 From: bcriscuolo at markspace.com (Brian M. Criscuolo) Date: Tue Nov 22 12:33:47 2005 Subject: [sync-services-developers] Record Pushing Issue In-Reply-To: References: Message-ID: That's fine if you're in sync. How about if a slow sync occurs? The case I have is this: - Contacts in sync. - Add phone number to record on the desktop - Force a slow sync by calling clientWantsToPushAllRecordsForEntityNames - Push record to desktop After the mingle, the desktop record does not have the phone number we just added. In what you're doing (I do the same thing for some cases) as long as you have already processed that address/phone/etc, you're good. But if it is new, it won't be staying around after your sync. Of course, I could be doing something wrong and this is a non-issue. b On Nov 22, 2005, at 12:00 PM, Jud Spencer wrote: > On 11/22/05 8:07:02 AM, "Brian M. Criscuolo" > > wrote: > >> Is it because I'm accepting all addresses when pulling the original >> record but only providing a subset of them when pushing a change >> back, as outlined above? Did Apple change the behavior of: >> "SyncSession should not issue a "clear" for Q." but not document it? >> >> There were numerous Sync Services changes in 10.4.3 but none were >> described in any detail. We need some detail about them, as shipping >> code behaves differently. >> >> Does anyone have any thoughts or suggestions to this perplexing >> issue? > > I'm new to the sync services game. I essentially started with > 10.4.3. I have > found that in a case where you are pushing a subset of entities > related to a > parent, that you need to provide the ID's of all the sub-entities > in the > parent, not just the ones that are being pushed. What we are doing > in our > app is keeping track of the entities when we have a schema > incompatibility, > and then pushing them on any update. For example, our app only > supports two > addresses. When we pull a record with more than two, we keep track > of the > ids of the ones that we accepted but didn't write to our database. > When we > later push the record back to the truth, we write the id's of the > addresses > that we write as well as the ids of the ones that we don't write > into the > street addresses property in .Contact. > > This works for us. > > Jud > > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers ******************** Brian M. Criscuolo Senior Software Engineer Mark/Space, Inc. bcriscuolo@markspace.com From btruex at markspace.com Tue Nov 22 13:26:46 2005 From: btruex at markspace.com (Brent Truex) Date: Tue Nov 22 13:26:53 2005 Subject: [sync-services-developers] Record Pushing Issue In-Reply-To: References: Message-ID: It looks like the behavior of clientWantsToPushAllRecordsForEntityNames changed at some point. Once this is called in 10.4.3, the SyncEngine takes what the client pushes as the total collection of records for an entity, even if the client explicitly refused some records from SyncServices in a previous session, or hasn't seen the records at all yet. Jud, what happens in your code in the case where a phone number (or other related record) is added to the Mac just before a slow sync? - Brent On Nov 22, 2005, at 3:32 PM, Brian M. Criscuolo wrote: > That's fine if you're in sync. How about if a slow sync occurs? > > The case I have is this: > > - Contacts in sync. > - Add phone number to record on the desktop > - Force a slow sync by calling > clientWantsToPushAllRecordsForEntityNames > - Push record to desktop > > After the mingle, the desktop record does not have the phone number > we just added. > > In what you're doing (I do the same thing for some cases) as long > as you have already processed that address/phone/etc, you're good. > But if it is new, it won't be staying around after your sync. > > Of course, I could be doing something wrong and this is a non-issue. > > b > > On Nov 22, 2005, at 12:00 PM, Jud Spencer wrote: > >> On 11/22/05 8:07:02 AM, "Brian M. Criscuolo" >> >> wrote: >> >>> Is it because I'm accepting all addresses when pulling the original >>> record but only providing a subset of them when pushing a change >>> back, as outlined above? Did Apple change the behavior of: >>> "SyncSession should not issue a "clear" for Q." but not document it? >>> >>> There were numerous Sync Services changes in 10.4.3 but none were >>> described in any detail. We need some detail about them, as shipping >>> code behaves differently. >>> >>> Does anyone have any thoughts or suggestions to this perplexing >>> issue? >> >> I'm new to the sync services game. I essentially started with >> 10.4.3. I have >> found that in a case where you are pushing a subset of entities >> related to a >> parent, that you need to provide the ID's of all the sub-entities >> in the >> parent, not just the ones that are being pushed. What we are doing >> in our >> app is keeping track of the entities when we have a schema >> incompatibility, >> and then pushing them on any update. For example, our app only >> supports two >> addresses. When we pull a record with more than two, we keep track >> of the >> ids of the ones that we accepted but didn't write to our database. >> When we >> later push the record back to the truth, we write the id's of the >> addresses >> that we write as well as the ids of the ones that we don't write >> into the >> street addresses property in .Contact. >> >> This works for us. >> >> Jud >> >> >> _______________________________________________ >> sync-services-developers mailing list >> sync-services-developers@lists.markspace.com >> http://lists.markspace.com/mailman/listinfo/sync-services-developers > > > ******************** > Brian M. Criscuolo > Senior Software Engineer > Mark/Space, Inc. > bcriscuolo@markspace.com > > > > > _______________________________________________ > sync-services-developers mailing list > sync-services-developers@lists.markspace.com > http://lists.markspace.com/mailman/listinfo/sync-services-developers --------------------------------- Brent Truex btruex@markspace.com