Dear Schuler,
>From: schuler@xxxxxxxxxxxxxxxxxxx (Carlos Alejandro Schuler)
>Organization: Aerometrics, Incorporated
>To: netcdfgroup@xxxxxxxxxxxxxxxx
>Subject: netCDF stressed under low memory conditions
>Keywords: 199502081821.AA18595
In the above message you wrote:
>I have found, by experimentation, and by just following the code,
>that netCDF is rather non-robust when run under low memory
>conditions (i.e. with a high probability of malloc failing).
I just checked the code for our netCDF implementation and could not
find (with one caveat) any instance where a failed malloc() or realloc()
was ignored: an error message is always emitted and an error status is
always returned. The one caveat is in the xdrposix_create() function,
which doesn't emit an error message (because XDR functions don't) but
does return an error status to the XDR layer.
Is it possible that you're using a program that doesn't check the return
status of the netCDF functions rigorously enough?
--------
Steve Emmerson <steve@xxxxxxxxxxxxxxxx>
