[mapguide-internals] Re: Motion: vote RFC 77 - Create Feature Source

[Leaf Li]

Hi Trevor,

Thanks for your clarification.

MgFileFeatureSourceParams should be a concrete implementation if there is enough similarity in the Fdo provider implementations for SDF, SHP, and Sqlite.  This seems to be indicated by the email discussion and class hierarchy but it is not specifically called out as such in the RFC.

Based on Jason's comment regarding class naming, I agree that MgCreateSdfParams should not be used to implement creators for SHP and SQLite.  If MgFileFeatureSourceParams can implement all the functionality of MgCreateSdfParams then I think we should mark MgCreateSdfParams as deprecated.  This would leave us with a single MgFileFeatureSourceParams class with support for three providers.  This could be done under a separate RFC related to RFC 77.
[Leaf] Yes. MgCreateSdfParams isn't used to implement creators for SHP and SQLite. Actually MgFileFeatureSourceParams is enough for SDF, SHP and SQLite. We keep MgCreateSdfParams in case some existing application still uses it and this change doesn't break their application. I will mark mark MgCreateSdfParams as deprecated in comments for this class. In the future, we can remove it completely.

What is MgFeatureService::CreateFeatureSourceSQLite?  This is not documented anywhere in the RFC.  Hopefully this is just a typo.  I think it would be good to avoid adding provider specific APIs to MgFeatureService public APIs.
[Leaf] This is a typo. It should be MgFeatureService::CreateFeatureSource(...).

Thanks,
Leaf Li