This is the mail archive of the
mailing list for the GCC project.
Re: IA32 bootstrap error
Richard Henderson wrote:
> On Sun, Dec 30, 2001 at 05:07:37PM -0500, Richard Kenner wrote:
> > So there is *no* version of GDB that will work?
> The gdb from cvs head mostly works.
My results weren't nearly as positive. Should I start chasing GDB problems
or GCC problems?
Starting program: /play/negcs-trunk/gcc/cc1 libgcc2.i
Breakpoint 1, fancy_abort (file=warning: Cannot convert floating-point register
value to non-floating-point type.
0x0, line=warning: Cannot convert floating-point register value to non-floating-
0, function=warning: Cannot convert floating-point register value to non-floatin
1450 internal_error ("Internal compiler error in %s, at %s:%d",
Initial frame selected; you cannot go up.
> Use cvs gdb. And cvs binutils.
> It's all one product, you see. If the parts of the toolchain
> are not tested together then it's virtually certain that they
> won't work together.
I can see that they're somewhat co-dependent as they're each the primary
consumer/producer of this sort of stuff, but that's not the promise that
ELF and Dwarf were trying to make. The goal was to be interoperable
across the object file level, wasn't it?
Additionally, the "it's all one product" approach leads to creep
between the modules. Today, I had to find and fix two GCC-isms and one
GLIBC-ism that have crept into GDB. So testing everything as one big
unit, in addition to raising the price tag of intellecual manageability
for working on these things, has problems of its own.
So while it's understandable that for specific cases they have to be
lockstepped, let's please be careful to not go too far down this path.
"Oh, and you need cvs GLIBC. And cvs linux kernel..."