John Evans <jevans@xxxxxxxxxxxxxxxxxx> writes:
> On Friday 15 July 2005 09:46 am, Ed Hartnett wrote:
>> This is in the C interface, I'm not familiar with the MatLab
>> interface, perhaps it exposes nc_create but not nc__create.
>
> The current release of the matlab interface (mexcdf.sourceforge.net)
> does not expose nc__create, but it took only about 10 minutes to do it.
> I can have a new release out shortly, once I update the docs and do a
> bit more testing on it.
>
>
Whoops! As my boss Russ just pointed out to me, what you want is in
the nc__enddef function, not nc__create. Here's what Russ sez:
"What you want is to provide a non-zero argument for the "h_minfree"
parameter of the nc__enddef() call, as documented here, for example:
http://www.unidata.ucar.edu/packages/netcdf/docs/netcdf-c/nc_005f_005fenddef.html
which specifies the amount of extra space to reserve in the header for
future additions. There is an equivalent function in the Fortran and
C++ interfaces, but I'm also not sure whether the Matlab interface
exposes the "underbar underbar" functions such as nc__create(),
nc__open(), and nc__enddef()."
Sorry about the confusion!
Thanks,
Ed
>
> --
> John Evans: IMCS Rutgers University
> 71 Dudley Rd, New Brunswick, NJ 08901-8521
> ph: 732-932-3435
>
>
--
Ed Hartnett -- ed@xxxxxxxxxxxxxxxx
