|
History
| Log In
|
|
|
|
[
Permlink
| « Hide
]
Sarah Allen - 11/Sep/07 11:49 AM
oddly, when Maynard removed the Debug.write the replicator ERROR went away. I wouldn't have expected that.
Is there a test case for this bug? I don't see how this could happen since the error is being signaled two lines after a call to setData, which should ensure there is a datapath?
Would a datapath still be ensured if setData is called with null?
setData creates a datapath if there is not one.
As I said to Maynard in email: if you could just turn on backtraces and get me a copy of the backtrace from the ERROR message that would be a start. In the developer console, check the backtrace button. Run the app. Click on the error message to inspect it, then click on the backtrace to inspect it, and copy/paste the output to me (or into the Jira bug). Can you inspect Frame #17 and inspect the `this` value (and any other args) and send me that too? Thanks!
I understand what's happening. The pagingdatasetreplicator overrides setData() but doesn't create a datapath until data is received from the server. I think the right thing to do is to create the datapath immediately to avoid this error. If this works, we may be able to resolve this bug won't fix.
It appears that it isn't the datapath but the datanode in line 203 that is null.
So, can you inspect the this value in frame 17? I still don't see how datanode gets to be null. Are you setting the nodes array somehow so that the datanode is null? This is in the code that Max added to support dataselectionmanager, so I'm trying to figure out what should be done to the selection when there's no datanode.
I sent Pablo a change to review that should make having a null datanode innocuous.
The datanode can be null because, in the case of paging data, the data for a node may not be there until it's received from the server.
r6440 | ptw | 2007-09-12 15:57:55 -0400 (Wed, 12 Sep 2007) | 19 lines
Changed paths: M /openlaszlo/branches/wafflecone/lps/components/utils/replicator/replicator.lzx Change 20070912-ptw-m by ptw@dueling-banjos.local on 2007-09-12 13:21:29 EDT in /Users/ptw/OpenLaszlo/wafflecone for http://svn.openlaszlo.org/openlaszlo/branches/wafflecone Summary: Take care updating selection when node is null Bugs Fixed: Technical Reviewer: pkang (Message-ID: <Pine.OSX.4.64.0709121247430.22280@pkang-mac.local>) QA Reviewer: max (Message-ID: <46E83BBD.3090603@openlaszlo.org>) Details: If there is no datanode, you can't be selected. Tests: Pablo to try with this test case Pablo has verified that the code now checks to be sure that the node is not null before evaluating. No test case exists. Verified as per Pablo Kang.
From Tucker: Pablo had a test case, I think it was huge (maybe all of webtop), which is why I had to rely on him confirming it was fixed. You could ask him if he has any sort of a simple test. On 2007-10-24, at 12:15 EDT, Steve O'Sullivan wrote: > Hello Tucker! > > I am looking at this bug with an eye toward verifying it as fixed. > However, there is no test case included in the bug notes that I can > see, and I do not know how best to approach it. > > Could you please forward a test case that I can use to verify this > bug, or explain which of our pre-existing test cases I should run? > > I believe that I will see the error "ERROR @utils/replicator/ > replicator.lzx#203: undefined object does not have a property '[]' > " if this test fails and the bug still is open, correct? > > Thanks for any and all assistance. > > Regards, > > > SteveO |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
Task
Verified
P1
Explicit replicator displays warning when a clone's datapath is null
