Building gdb from cvs

Building gdb from cvs

[Frank Middleton]

Using the supplied SPARC version of gdb, printing of strings works

/usr/bin/gdb.6.3.50 --version
GNU gdb 6.3.50_2004-11-23-cvs
...
3	char *c = "ABCD";
(gdb) n
4	return 0;
(gdb) p c
$1 = 0x105a0 "ABCD"

But, in order to debug on SPARC, I am using a locally patched version
and printing of strings doesn't seem to work any more

/usr/local/bin/gdb --version
GNU gdb (GDB) 6.8.50.20090427-cvs
...
3	char *c = "ABCD";
(gdb) n
4	return 0;
(gdb) p c
$1 = 0x105a0 "

This must be an artifact of the way this was built (it uses gnu libiconv,
for example), or some kind of regression (unlikely).

Since it was built with a motley collection of Solaris and gnu libraries,
this is probably unreproducible. Can you guess as to which libraries you
think might be causing this so I can relink with the gnu ones instead?

Thanks

Re: Building gdb from cvs

[Joel Brobecker]

> (gdb) p c
> $1 = 0x105a0 "

I've had something similar on AIX:
http://www.sourceware.org/ml/gdb-patches/2009-04/msg00704.html

I've just posted a new fix for the issue on AIX. Perhaps it's
the same for you as well.
http://www.sourceware.org/ml/gdb-patches/2009-05/msg00159.html

-- 
Joel

Re: Building gdb from cvs

[Frank Middleton]

On 05/07/09 20:01, Joel Brobecker wrote:
>> (gdb) p c
>> $1 = 0x105a0 "
>
> I've had something similar on AIX:
> http://www.sourceware.org/ml/gdb-patches/2009-04/msg00704.html
>
> I've just posted a new fix for the issue on AIX. Perhaps it's
> the same for you as well.
> http://www.sourceware.org/ml/gdb-patches/2009-05/msg00159.html

Sure was, although the patch seems to have already been applied
to the current weekly snapshot.FWIW this has now been checked
out on SPARC/SunOS (snv103)...

Thanks!