[mapserver-dev] Implement msPostgisLayerGetExtent

[Tamas Szekeres]

2013/12/31 thomas bonfort <thomas.bonfort at gmail.com>

> > As far as I can see in the discussion, a global setting to disable the
> > driver extent calculation for the GetCapabilities request would be a good
> > consensus, isn't it?
>
> My opinion is that this would be a bare minimum. The difficult
> question is whether the expensive extent calculation should set to on
> or off by default:
> - ON by default: people who have not updated their mapfiles, or people
> who do not explicitly disable, are faced with potential denial of
> service if they have layers with many features
> - OFF by default: the wrong extent calculation continues to cause
> unexpected behaviour that isn't trivial to debug
>
>
To be specific currently msLayerGetExtent is called from the following
locations:

1. msOWSGetLayerExtent
2. msRASTERLayerGetExtent (for tile indexes)
3. getExtent (mapscript)

I don't see any other use cases.

For #2 and #3 it is expected to return a valid extent instead of a failure.
The only issue we should talk about is  msOWSGetLayerExtent whether or not
to use the driver specific extent calculation. Seems like we intend to
continue supporting an undocumented "side effect" which I'm quite unsure
that is the correct route to follow.

Tamas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/mapserver-dev/attachments/20131231/01b5530e/attachment.html>