This version is source and binary compatible with the 1.0.1 version of the driver and contains bug fixes and 1 addition to the API.
Changes since the 1.0.1 release include:
- Updated the MongoCollection<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html?com/...>interface to support the save(...)/saveAsync(...) operation as a short cut for either an insert or update. See the JavaDoc<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html>for the full details. - Updated the QueryBuilder.and(...)<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html?com/...>to optimize away the $and operator when the sub-conditions of the query did not collide. In sharded environments, the mongos process does not route $and queries to a single shard and instead always relies on the scatter/gather strategy. Scatter/gather forces client update()s to set the multi-update flag to true. By optimizing the $and operator away, we can avoid having to set the multi-update flag. - Updated parsing of responses to inserts to throw a DuplicateKeyException<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html?com/...>in the case of an "ok" value of 1 but still having an error message. - Improved error handling for the send and receive threads to ensure connections are never into a half open state.
Thanks to all those who provided feedback and helped with testing the driver.
Keep the comments, suggestions, questions, and bug reports coming.
Perhaps you could move the project to GitHub? I have been looking for an alternative MongoDB Java driver that utilizes non-blocking I/O and you seem to have built your driver around that goal. I think community development support would be a good thing to have for this driver. Maybe it could eventually replace the synchronous driver that is officially supported.
> This version is source and binary compatible with the 1.0.1 version of the > driver and contains bug fixes and 1 addition to the API.
> Changes since the 1.0.1 release include:
> - Updated the MongoCollection<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html?com/...>interface to support the save(...)/saveAsync(...) operation as a short cut > for either an insert or update. See the JavaDoc<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html>for the full details. > - Updated the QueryBuilder.and(...)<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html?com/...>to optimize away the > $and operator when the sub-conditions of the query did not collide. In > sharded environments, the mongos process does not route $and queries > to a single shard and instead always relies on the scatter/gather strategy. > Scatter/gather forces client update()s to set the multi-update flag to > true. By optimizing the $and operator away, we can avoid having to set > the multi-update flag. > - Updated parsing of responses to inserts to throw a > DuplicateKeyException<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html?com/...>in the case of an "ok" value of 1 but still having an error message. > - Improved error handling for the send and receive threads to ensure > connections are never into a half open state.
> Thanks to all those who provided feedback and helped with testing the > driver.
> Keep the comments, suggestions, questions, and bug reports coming.
> Perhaps you could move the project to GitHub? I have been looking for an > alternative MongoDB Java driver that utilizes non-blocking I/O and you seem > to have built your driver around that goal. I think community development > support would be a good thing to have for this driver. Maybe it could > eventually replace the synchronous driver that is officially supported.
> Thanks,
> Phil
> On Sunday, October 7, 2012 10:54:40 AM UTC-4, Rob Moore wrote:
>> Version 1.0.1 of the Asynchronous Java Driver is now available:
>> This version is source and binary compatible with the 1.0.1 version of >> the driver and contains bug fixes and 1 addition to the API.
>> Changes since the 1.0.1 release include:
>> - Updated the MongoCollection<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html?com/...>interface to support the save(...)/saveAsync(...) operation as a short cut >> for either an insert or update. See the JavaDoc<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html>for the full details.
>> - Updated the QueryBuilder.and(...)<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html?com/...>to optimize away the >> $and operator when the sub-conditions of the query did not collide. >> In sharded environments, the mongos process does not route $andqueries to a single shard and instead always relies on the scatter/gather >> strategy. Scatter/gather forces client update()s to set the multi-update >> flag to true. By optimizing the $and operator away, we can avoid >> having to set the multi-update flag.
>> - Updated parsing of responses to inserts to throw a >> DuplicateKeyException<http://www.allanbank.com/mongodb-async-driver/apidocs/index.html?com/...>in the case of an "ok" value of 1 but still having an error message.
>> - Improved error handling for the send and receive threads to ensure >> connections are never into a half open state.
>> Thanks to all those who provided feedback and helped with testing the >> driver.
>> Keep the comments, suggestions, questions, and bug reports coming.
>> Rob.
Hi
I am interested in using this driver on android. Does it work?
> I am interested in using this driver on android. Does it work?
To be honest I don't know. It should work as the dependencies within the JVM are from the "core" but I have not tried it.
Have you tried and run into a problem or are you wondering if someone has been successful already?
I have not found an easy method for validating that a third party jar works on the Dalvik runtime. If anyone knows of one I'd be interested in the details.
On Thursday, October 18, 2012 12:20:55 PM UTC-4, Phil Gates-Idem wrote:
> Rob,
> Perhaps you could move the project to GitHub? I have been looking for an > alternative MongoDB Java driver that utilizes non-blocking I/O and you seem > to have built your driver around that goal. I think community development > support would be a good thing to have for this driver. Maybe it could > eventually replace the synchronous driver that is officially supported.
Phil,
Actually, the driver does not use non-blocking I/O. Instead it uses a pair of threads <http://www.allanbank.com/mongodb-async-driver/#Thread_Model>(reader/writer) and fast message passing between the application thread and the read/write threads. A very early version of the driver used NIO and selectors but after testing the performance we found the threaded model was better and the main win for NIO is handling lots of connections which is exactly what the driver is trying to avoid.
On hosting: The driver is still in very young and we are working hard to avoid fragmentation/breaks/cruft in the driver's API. For this and a few other reasons we don't plan to add the code to GitHub.
If there are a features that you would like to see added to the driver you can either post a topic to this group or send me a message directly. I do hope to one day get a public ticket system created. So far the copy/paste cycle for the few issues we have had people report has not caused enough pain to cause me to actually get it setup and other work always jumps in front.
I am looking at using your driver for reading from the oplog collection. I got to the point where i create the db, open the oplog and read the contents. Does the driver support tailable cursor?
On Monday, October 22, 2012 6:35:24 PM UTC-7, Rob Moore wrote:
> On Thursday, October 18, 2012 12:20:55 PM UTC-4, Phil Gates-Idem wrote:
>> Rob,
>> Perhaps you could move the project to GitHub? I have been looking for an >> alternative MongoDB Java driver that utilizes non-blocking I/O and you seem >> to have built your driver around that goal. I think community development >> support would be a good thing to have for this driver. Maybe it could >> eventually replace the synchronous driver that is officially supported.
> Phil,
> Actually, the driver does not use non-blocking I/O. Instead it uses a pair > of threads <http://www.allanbank.com/mongodb-async-driver/#Thread_Model>(reader/writer) and fast message passing between the application thread and > the read/write threads. A very early version of the driver used NIO and > selectors but after testing the performance we found the threaded model was > better and the main win for NIO is handling lots of connections which is > exactly what the driver is trying to avoid.
> On hosting: The driver is still in very young and we are working hard to > avoid fragmentation/breaks/cruft in the driver's API. For this and a few > other reasons we don't plan to add the code to GitHub.
> If there are a features that you would like to see added to the driver you > can either post a topic to this group or send me a message directly. I do > hope to one day get a public ticket system created. So far the copy/paste > cycle for the few issues we have had people report has not caused enough > pain to cause me to actually get it setup and other work always jumps in > front.
> I am looking at using your driver for reading from the oplog > collection. I got to the point where i create the db, open the oplog and
> read the contents. Does the driver support tailable cursor?
In theory all of the pieces exist. I never exposed the capability through the interfaces since I did not feel it had seen enough testing and was worried about the impact it would have on other requests via the same Mongo instance.
If you are willing to run a snapshot version and give me some feedback on how it is interacting with other requests/usage of the client/connection I'll send you a link to a version with the capability exposed.
On Friday, December 14, 2012 4:52:20 PM UTC-8, Rob Moore wrote:
> On 12/14/2012 04:22 PM, anucekay wrote: > > Hai Rob,
> > I am looking at using your driver for reading from the oplog > > collection. I got to the point where i create the db, open the oplog and > > read the contents. Does the driver support tailable cursor?
> In theory all of the pieces exist. I never exposed the capability > through the interfaces since I did not feel it had seen enough testing > and was worried about the impact it would have on other requests via the > same Mongo instance.
> If you are willing to run a snapshot version and give me some feedback > on how it is interacting with other requests/usage of the > client/connection I'll send you a link to a version with the capability > exposed.
