Re: [Fresco-devel] huge so's: are they normal?

Hi, Andy:

El jue, 06 de 03 de 2003 a las 05:21, Andy Sy escribió:
> > Nick Lewycky wrote:
> > > Andy Sy wrote:
> > 
> > > > In Fresco M1, checking out the sizes of some of the so's in the
> > > > Berlin/modules/ directory, a lot of the generated so's are huge.
> > > > PSDrawingKit.so for example is 14MB in size even though the sources
> > > > are quite small (around 20k).
> > > 
> > > That's abnormal. It's only 950k here. The largest is ToolKit.so, at 5.3Mb.
> > > 
> > > > Anyone know what's going on?
> > > 
> > > Not a clue. Is something being statically linked by accident? Something big?
> > 
> > You didn't configure with debugging enabled, did you? That might explain
> > it, since the build bloats incredibly, or it did the last time I tried
> > (and ran out of disk space on the system I was building on!)
> 
> Hmmm... this is exactly what happens although I never turned on
> debug.  All I did was ./configure --prefix=/usr/local.
> Could it have something to do with omniORB-4.0.0?  My omniORB
> .so's and .a's are 2-3MB in size - is this typical?
> 

Yes, I think it's typical.  What compiler are you using?  It turns out
that in some versions (I don't know if this is the normal behavior)
AC_PROG_CC adds "-g" automatically if it finds the gcc compiler.  Can
you check your Fresco's config.status file to see if it contains the
"-g" flag?:

grep -e "-g" config.status

If so, the big size can be caused by the debug flag. Also, you can check
the size without debugging information by running split on those huge
libraries.  If the size decreases, you're sure it is due to the debug
overhead.

	Best regards.
	diego

-- 
Diego Sevilla Ruiz -- http://ditec.um.es/~dsevilla/ -- [email protected]
Dep. Ingeniería y Tecnología de Computadores, Facultad de Informática
Univ.de Murcia,Campus Espinardo,30080 Murcia (SPAIN),Tel.+34968367571
lynx -dump http://ditec.um.es/~dsevilla/face |uncompface -X | display
PGP:  http://pgp.rediris.es:11371/pks/lookup?op=get&search=0xC9B964B7
(lambda x,f=lambda x,f:x and x[-1]+f(x[:-1],f):f(x,f))('se.mu@allivesd')