Frank Warmerdam <warmerdam <at> pobox.com> writes: > On 11-09-08 07:48 AM, Jim Pendleton wrote: > > Is there any model for redirecting CPLDebug output from stdout to an > > alternative path? > > There is no need to overwrite the high level CPLError and CPLDebug functions. > Instead install a handler for the messages with CPLSetErrorHandler().
That was the information I needed. For the sake of expediency, I've ifdef'ed __android_log_print() calls adjacent to the printf() statements in the existing CPLDefaultErrorHandler rather than duplicating the code and adding another source file to the build. If I eventually proceed the approved GDAL development way is there a standard place to call CPLSetErrorHandler() when loading GDAL as a library? Since I'm running in Java, I wouldn't necessarily have a pointer to my C++ function at runtime without some extra hackery (and there's no SWIG binding presently for that routine, regardless.) I would like to perturb as little of the original GDAL source as possible. If only there existed a gdal.SetConfigOption interface for CPLSetErrorHandler() that would allow me to pass an entry point name instead of a function pointer... Thanks for your help, Jim P. _______________________________________________ gdal-dev mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/gdal-dev
