[GRASS5] Re: r.poly update

GRASS GRASS developer
1

Hi Andrea,
hi all,

(cc to grass5)

On Mon, Mar 26, 2001 at 08:16:31AM +0200, Andrea Aime wrote:

Markus Neteler wrote:
> On Fri, Mar 23, 2001 at 04:40:43PM +0100, Andrea Aime wrote:
> > Hi Markus,
> > here you will find my modified version of r.poly. I'm not sure about
> > it, but I think that only bound.c changed.
>
> Thanks, the file is in CVS now (was the island/label fix, is that right?).

Yes, but another fix is required as soon as something is decided about
the use of zero values as vector categories

I need a recommendation of the others here...
(see:
http://www.geog.uni-hannover.de/cgi-bin/minorweb.pl?A=READ&L=grass5/2001/3/12/11
)

[...]

PS: as I'm playing with GRASS sources I've noticed in head.in that you
specify -g as a compiler flag (that produces debug symbols) and then -s in
the linker flags, that strip away any debug symbol. Would it be possibile
to specify some flag to gmake5 in order to obtain a debug build (maybe
with -ggdb flag that produces best debug information for the GNU debugger)
and then maybe no flags for a production build (what about optimizations?
computation intensive software can gain up to 30% speed-up with the proper
compiler optimizations).

Currently I don't know how to tell configure to use
-O2
instead of
-g

I have checked configure.in, the related line seems to be line 26
AC_PROG_CC
Obviously something needs to be predefined for selecting -O2 on default?

Markus

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

2

On Mon, Mar 26, 2001 at 09:23:18AM +0100, Markus Neteler wrote:

Hi Andrea,
hi all,

(cc to grass5)

On Mon, Mar 26, 2001 at 08:16:31AM +0200, Andrea Aime wrote:
> Markus Neteler wrote:
> > On Fri, Mar 23, 2001 at 04:40:43PM +0100, Andrea Aime wrote:
> > > Hi Markus,
> > > here you will find my modified version of r.poly. I'm not sure about
> > > it, but I think that only bound.c changed.
> >
> > Thanks, the file is in CVS now (was the island/label fix, is that right?).
>
> Yes, but another fix is required as soon as something is decided about
> the use of zero values as vector categories

I need a recommendation of the others here...
(see:
http://www.geog.uni-hannover.de/cgi-bin/minorweb.pl?A=READ&L=grass5/2001/3/12/11
)

Zero is not a valid category for vectors as far as I understand. I
suspect there will be other problems if we try to generate vectors that
have a zero category number. It would be okay for a category value
though. Category numbers should be positive integers. I don't know
what a good solution would be...

[...]
> PS: as I'm playing with GRASS sources I've noticed in head.in that you
> specify -g as a compiler flag (that produces debug symbols) and then -s in
> the linker flags, that strip away any debug symbol. Would it be possibile
> to specify some flag to gmake5 in order to obtain a debug build (maybe
> with -ggdb flag that produces best debug information for the GNU debugger)
> and then maybe no flags for a production build (what about optimizations?
> computation intensive software can gain up to 30% speed-up with the proper
> compiler optimizations).

Currently I don't know how to tell configure to use
-O2
instead of
-g

I have checked configure.in, the related line seems to be line 26
AC_PROG_CC
Obviously something needs to be predefined for selecting -O2 on default?

You don't wont -O2 if you want debugging. I just modify the flags
in head.$(ARCH), removing any -O flags, and removing the -s flag from
the linker flags. Then recompile whatever you're interested in.

--
Eric G. Miller <egm2@jps.net>

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'