This is the mail archive of the
mailing list for the Archer project.
Re: [delayed-symfile] handle incorrect aranges
On Fri, 14 Aug 2009 19:13:02 +0200, Tom Tromey wrote:
> FWIW, I started with the original test case. What happened there was
> that the aranges reader stopped early, then treated some other bytes
> from the middle of an aranges section as a new CU header. This is how
> we ended up with address_size==0. If you 'readelf' the original
> binaries you'll see there is nothing wrong except that bogus (0,0)
Oh, understood now.
> I'll wait until you push before committing, so I can test against your
> test case.
Checked the testcase in as: 77fa7778a37b0d28a7e4e5235f074a10ecf1815d
Fedora 11: https://admin.fedoraproject.org/updates/gdb-126.96.36.19990302-37.fc11