Hi Brad,
On Sat, Jun 11, 2005 at 02:15:30PM +0200, grass@intevation.de wrote:
Author: brad
Update of /grassrepository/grass6/lib/image
In directory doto:/tmp/cvs-serv30756Modified Files:
close.c filbuf.c flsbuf.c name.c open.c pix.c rdwr.c rle.c
row.c
Log Message:
Added return types and proper returns to functions.
Added proper header files.
thank you a lot for fixing this library!
Markus
On Sat, Jun 11, 2005 at 06:05:22AM -0700, Brad Douglas wrote:
> > Update of /grassrepository/grass6/lib/image
...
Can I replace i_seterror() and associated functions with
G_set_error_routine()?
Probably (hopefully) yes: AFAIK the image is very old and was
not updated when the G_*() functions were implemented.
But I am not very familiar with it. The lib/ogsf/ is making use
of lib/image, so subsequent (few) changes there may be required.
Markus
On Sat, 2005-06-11 at 16:25 +0200, Markus Neteler wrote:
On Sat, Jun 11, 2005 at 06:05:22AM -0700, Brad Douglas wrote:
> > > Update of /grassrepository/grass6/lib/image
...
>
> Can I replace i_seterror() and associated functions with
> G_set_error_routine()?
>Probably (hopefully) yes: AFAIK the image is very old and was
not updated when the G_*() functions were implemented.
I suppose the real question is: Do we want to make the various libraries
in lib dependent upon lib/gis or should they be as independent as
possible?
Rewriting the lib/image error routines are simple enough, so either way
works for me.
But I am not very familiar with it. The lib/ogsf/ is making use
of lib/image, so subsequent (few) changes there may be required.
AFAIK, lib/ogsf is the only usage of lib/image. The biggest reason I
started working on lib/image was to allow lib/ogsf to compile without
explicitly adding -fPIC to compiler flags.
PS -
Oddly, lib/image has been kinda fun to work with. Reminiscent of
writing video card kernel drivers and I learned the header format of a
SGI IRIS images.
--
Brad Douglas <rez@touchofmadness.com>