[gdal-dev] Memory leaks?

[Paul Selormey]

Hello Ed,
Thanks for the clarifications.

Best regards,
Paul.

----- Original Message -----
From: "Ed McNierney" <ed at t...>
To: <gdal-dev at yahoogroups.com>
Sent: Tuesday, August 27, 2002 12:06 PM
Subject: RE: [gdal-dev] Memory leaks?


>
> Paul -
>
> No, that's not the same thing at all. The call to GDALAllRegister ()
> *loads* all the configured drivers, but it does not *open* them. For
> example, in the GeoTIFF driver, the call to GDALAllRegister () calls (in
> the driver) GDALRegister_GTiff (), which registers the driver by
> allocating a data structure for it and filling it in. This is pretty
> much what all drivers do at registration time. There is no equivalent
> "GDALUnregister_Gtiff" function to undo this operation, and so there is
> no equivalent "GDALUnregistter_Gtiff" function in the driver.
>
> The constructor and destructor for the class GTiffDataset are not called
> at registration time. They're called much later if and when there's an
> actual dataset for that driver to handle, and the destructor is indeed
> properly called when the dataset is no longer needed. If it were not
> being called, your memory leak tester would be reporting allocations
> from other than GDALAllRegister that were left over.
>
> At the GDALAllRegister level, there's no "destructor" to be called, so
> the failure to call it doesn't create a problem for any driver.
>
> - Ed
>
> -----Original Message-----
> From: Paul Selormey [mailto:paul at t...]
> Sent: Monday, August 26, 2002 10:41 PM
> To: gdal-dev at yahoogroups.com
> Subject: Re: [gdal-dev] Memory leaks?
>
>
>
> Hello Ed,
>
> > (at the risk of pushing this thread over the edge into a religious
> > debate...)
>
> I am not interested in debating an issue for the sake of debate, there
> is not time for this.
>
> > While it would be nice to offer a complementary "close and clean up"
> > function for the library, I can't agree that it's wrong to not have
> > one.
>
> I think you have already made the point. If you agree, however, that the
> destructors are not called, then have a look at the destructors for each
> driver to see what they really do, and decide whether it is a choice to
> clean up or not - this concludes my contribution to this issue.
>
> Best regards,
> Paul.
>
>
>
>
>
> Your use of Yahoo! Groups is subject to
> http://docs.yahoo.com/info/terms/
>
>
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>