[fdo-internals] Defects in PostGIS provider

[Mateusz Loskot]

Chuck Wilbur wrote:
> Hi Everybody,

Hi Chuck,

Thank you for writing to the fdo list.
It should overcome my inability to work on these problems in short time.

> How do I go about submitting defect reports against the PostGIS provider in
> trunk\Providers\PostGIS? I've downloaded and built the code to track down
> some problems I was having, and I've uncovered what look like three issues
> with the provider itself that are blocking me from proceeding with
> development:
> 
> 1) InsertCommand::Execute can only insert NULL geometry values successfully.
> This is the worst of the three defects, because there's no way to work
> around it. Internally the code is building a non-parameterized statement
> string and the WKB format fields are being interpreted by PostGIS as long
> integers, which causes insertion to fail.

The provider generates WKB stream from FDO geometry and inserts it
directly into INSERT INTO statement.

Does it crash for every geometry or may be if you use XYZ or XYZM
dimensions?

> 2) ApplySchemaCommand::Execute always issues CREATE TABLE statement(s) - it
> cannot be used to delete a feature class.

Hmm, I'm not sure I understand.
ApplySchema command is not used for removing feature classes (tables).

> 3) Some internal checks are case-sensitive after Postgres has changed the
> case to all lower, so (for example) a newly-created schema with a multi-case
> name cannot be accessed or used with the name it was created under.

I try to follow strategy with quoting all schema identifiers.
Perhaps, there are some places where I forgot to add quotes.

I'll scan the code.

> P.S. I would be interested in helping to fix these if I knew my efforts
> could be integrated into the code stream. What process is in place for
> getting write access to the Subversion vault?

I think the easiest and suggested way of sending codes to FDO is to open
new Ticket and add patch.
You can generate patch very easily by calling "svn diff > my-fix.patch"
file.


Thank you very much for your analysis.
I also apologize for not providing you with reasonable support earlier.

Cheers
-- 
Mateusz Loskot
http://mateusz.loskot.net