[gdal-dev] rfc 24 - jpip

Frank Warmerdam warmerdam at pobox.com
Tue Jul 7 13:49:46 EDT 2009


Norman Barker wrote:
> Hi
> 
>  
> 
> http://trac.osgeo.org/gdal/wiki/rfc24_progressive_data_support
> 
>  
> 
> and the code at
> 
>  
> 
> http://svn.osgeo.org/gdal/sandbox/normanb/

Norman,

I pretty much agree with the things Even mentioned a few weeks ago.  A few
notes:

1) I would like you to provide the developer docs for GDALAsyncRasterIO and
the new methods on GDALDataset.  Well, you can skip the GDALAsyncRasterIO
accessor methods, but it is imperative (IMHO) that details docs be provided
for methods like BeginAsyncRasterIO(), GetNextUpdatedRegion(), LockBuffer(),
and UnlockBuffer().

2) Currently the JPIPKAK dataset does not appear to support normal data
access.  I do not believe this is acceptable.  If nothing else, it will
be educational for the rest of us if you implement the "normal" file
raster access mechanisms by invocation of appropriate async services.

3) I think we should provide a default GDALAsyncRasterIO() implementation
based on calls to the regular raster io functionality that works properly
except that it does not necessarily honour the timeout functionality
properly (it is effectively synchronous).

4) The GDALAsyncRasterIO object should be renamed GDALAsyncRasterReader
or something that does not imply that writing is supported unless that
is intended for the future.

5) The GDALAsyncRasterIO implementation should not be gdaldataset.cpp
(except for the begin call).

I am supportive of the general approach.

PS. I've successfully built the support on linux with only a few minor
changes.  I have yet to test it though.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent



More information about the gdal-dev mailing list