Re: [Mingw-users] Ridiculous debug executable sizes?

[Michael G. <mg...@te...>]

> >>So now, all is looking much better, and maybe it's time to modify that
> >>x-compiler build script, so it doesn't build the libraries with debuggi=
ng
> >>symbols in the first place.  What do you think, Michael?
> >=20
> >   I'm not Michael, but I think that the best solution would be to adapt=
 the
> > x-compiler build script so you can download the actual win32 native min=
gw
> > headers and libs from the distribution and use them in a --with-sysroot=
 (or
> > perhaps just --with-headers/--with-libs) configure time option.  That w=
ay the
> > cross-toolchain is *guaranteed* to produce consistent interoperable cod=
e to
> > the native one.
>=20
> I'm not Michael either, but I disagree.  If you can't build the compiler =
from=20
> scratch, then you have no chance of doing anything funky like applying pa=
tches=20
> to the compiler, or using a different compiler than the 'standard' distri=
buted=20
> version.
>=20
> This *should* work, and it needs to be corrected.  If it doesn't work, th=
en we=20
> don't understand whats going on.  A gap in knowledge is a doorway for bug=
s to=20
> mistakes to enter, and it needs to be plugged.

Well I *am* Michael ;-) and I've been trying to understand what actually
happens. Due to tight deadlines at work I'm only able to devote limited
time to this but so far I have not yet verified that the xcompiled libs
are indeed build with debugging info enabled (note I'm not saying they
aren't; while I write this I have another idea to give a try :)

I also tried to create a xcompiler with the new binutils Danny just
published but so far no luck. Again I had no time to look into the logfile
to check what went wrong -- all I know at this point in time is it did
not finish correctly :-(

Anyway, I'll give my other idea a try in a minute and will report the
results here (today I'd expect).

Best,
Michael
=2D-=20
 Vote against SPAM - see http://www.politik-digital.de/spam/
 Michael Gerdau       email: mg...@te...
 GPG-keys available on request or at public keyserver