one note:
now when I run the tests, i get in IntegrationXmlRpcTests, 3 errors, and 9
failures.
Errors:
testCreateIndexer
testDropIndexerTwice
testListIndexers
Failures:
testListInitialCollections
testGetParentCollection
testGetMultipleTextNodes
testFetchMultipleAttributeNode
testFetchAttributeNode
testFetchBoolean
testFetchString
testFetchNumber
testConflictingPrefix
do others get the same errors and failures?
dave
-----Original Message-----
From: Viner, David [mailto:[EMAIL PROTECTED]
Sent: Friday, August 08, 2003 4:46 PM
To: '[EMAIL PROTECTED]'
Subject: RE: MetaService and Metadata
Hi Vadim,
You are correct. I was only thinking of the rpc aspect. But, I'm glad
you checked in your changes and a test case to ensure proper function for
the future.
I'll take a look at the test cases, and see how/if a user can change
modified and created time.
dave
-----Original Message-----
From: Vadim Gritsenko [mailto:[EMAIL PROTECTED]
Sent: Friday, August 08, 2003 3:58 PM
To: [EMAIL PROTECTED]
Subject: Re: MetaService and Metadata
Viner, David wrote:
>First, let me say... you rock. Thanks for getting this done!
>
>Second, the streamToXML(Document doc) should invoke streamToXML(doc, true).
>That was simply an oversight on my part.
>The streamFromXML(Element source) I am not so sure about. If we invoke
>streamFromXML(source, true), then any "created" or "modified" times listed
>in the xml will be used by the MetaData object;
>
I think you are thinking only about rpc part. Note that the exact same
method is used to read meta data from the storage. Meaning, if there is
"false", then timestamps (if any) stored in meta.tbl will be ignored.
So, in order to obtain those values from the storage, you have to set
this value to true.
>that is the MetaData will be
>updated to contain the created and modified times listed in the incoming
>xml. I'm not sure that this is the desired or intended behavior. It seems
>to me that the "modified" time (and "create" time) should be controlled by
>the internals of xindice by default, and by default the user should not be
>able to alter them by submitting new xml. If, however, the user truly
wants
>to change the modified time externally, he or she is empowered by calling
>the streamFromXML with true as the second argument.
>
This will have to be investigated later... Will need to extend test case
for this.
Vadim
>Does that make sense? I'm not sure that I'm correct; this is just my gut
>instinct.
>
>dave
>
>
...
