dolfin team mailing list archive
-
dolfin team
-
Mailing list archive
-
Message #02318
Re: configure/build question
On Mon, Apr 03, 2006 at 11:51:03PM +0200, Mandus wrote:
> On 4/3/06, Anders Logg <logg@xxxxxxxxx> wrote:
> > On Mon, Apr 03, 2006 at 10:59:36PM +0200, Mandus wrote:
> > > Dear dolfin'ers
> > >
> > > After doing a relatively plain configure of dolfin, my 'dolfin-config
> > > --cflags' return the following (among other things):
> > >
> > > -DDEBUG=1 -g -O2 -Wall -Werror -ansi -std=c++98
> > >
> > > The problem is that the -Werror thing cause problem for me, when
> > > linking dolfin and our internal software (pycc for those who knows).
> > >
> > > I'm using dolfin from 'hg', freshly pulled today.
> > >
> > > Is there a way to turn off those flags?
> >
> > The debugging flags are set by default to stop developers (including
> > myself) from checking in code that compiles with warnings.
> >
> > Since it may cause problems, I will disable the debug flags by
> > default. You can either wait for 30 seconds until I have pushed the
> > update to the repository or comment out the following line in
> > configure.ac and then run scripts/preconfigure:
> >
> > enable_debug="yes"
>
> ok - more or less what I already did (actually, I just removed the
> -Werror completely from configure.ac...)
>
> I think this is a wise thing to have in the development tree, so it is
> probably just me using the "wrong" version (hg vs. release)...
Yes, there could be different flags for the development versions and
for the release, but then I would have to remember to turn the flag on
and off every time I make a release... :-)
We'll try without the flag for a while and see how it goes...
> The
> problem I encounter, among other things, is that swig generates
> wrapper code with variables that are never used, which gcc chockes on
> when this flag is on. There are also some issues with python header
> files combined with /usr/include/features.h, which is included
> somehow.
ok, I imagine there may be some issues with PyDOLFIN which is very
fresh.
> Anyway, thanks
>
> btw:
> best regards, _dr_ Åsmund Ø. :-)
Congratulations!
I saw the picture of the cake, very nice... :-)
/Anders
References