Hello All, Have anyone experienced any similar behavior? Is there any robot black-listing mechanism active in Wave? Do any of our thoughts make any sense? We are not sure what's going on.
Maybe it has anything to do with this bug report: http://code.google.com/p/google-wave-resources/issues/detail?id=494 Thanks. -- http://blog.thewe.net http://twitter.com/theWE_ On Wed, Nov 18, 2009 at 12:23 PM, Avital Oliver <[email protected]> wrote: > Hello friends, > > We are working on a mechanism for replication of portions of gadget states > in different waves using robots. Basically, our robot responds to a > BLIP_SUBMITTED event on one wave and sends one or more > DOCUMENT_ELEMENT_MODIFY_ATTRS operations to one or more gadgets. We are > writing our robot in Clojure without any of the APIs by manually generating > the operation JSONs. Our robot is mostly working but we seem to be affected > by several possible bugs and since we don't have logs from within the Wave > server or the Robot Gateway it's making our debugging work somewhat > difficult. All of our work has been done by verifying through the App Engine > logs that the operations are being sent. > > Questions: > > 1. We seem to have learned that if an operation in sent with illegal JSON > structure (for example sending null for operations.list instead of []) then > that robot stops working, or perhaps stops working in the context of the > wave it was trying to act on. Are we correct in this observation? Is there > some defined state of a certain robot not being allowed to operate on a > specific wave? When does this happen? > > 2. We think that if we send two distinct operations modifying the same > gadget the operation doesn't happen and perhaps also then puts the robot in > the "not working" mode (as described in issue 1). When we united the two > operations into one operation it seems to be working. Is it true that this > is the case or are we confusing several different issues into one? > > 3. As the complement of issue 2, we also have started to believe that you > can't send any two DOCUMENT_ELEMENT_MODIFY_ATTRS operations in one bundle, > even if it is to two different gadgets in different waves. When we sent only > one such operation and maintained the second one to be sent as a response to > the next DOCUMENT_CHANGED event it seemed like we got things to be working > (but other things kept not working). Are we correct in our guessy > observation? > > 3.1. This also may have to do with the fact that the two gadgets have the > same URL (in two different waves). Could this be the cause of some of our > problems? > > 4. We tried sending a DOCUMENT_ELEMENT_MODIFY_ATTRS operation with a null > value and it did not delete the key (we assumed it might work this way > because of the way submitDelta works in the gadget API). We also looked for > a different operation type that might remove a key but couldn't find one. > How can we delete a gadget key using a robot? > > Thanks, > Avital and Ayal. > > -- > http://thewe.net > > -- > You received this message because you are subscribed to the Google Groups > "Google Wave API" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]<google-wave-api%[email protected]> > . > For more options, visit this group at > http://groups.google.com/group/google-wave-api?hl=. > -- You received this message because you are subscribed to the Google Groups "Google Wave API" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/google-wave-api?hl=.
