This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: GDB internal error in pc_in_thread_step_range
> Cc: gdb-patches@sourceware.org
> From: Simon Marchi <simon.marchi@polymtl.ca>
> Date: Sun, 23 Dec 2018 18:31:45 -0500
>
> On 2018-12-23 11:10 a.m., Eli Zaretskii wrote:
> > Like this?
> >
> > --- gdb/coffread.c~1 2018-07-04 18:41:59.000000000 +0300
> > +++ gdb/coffread.c 2018-12-23 10:24:15.758116900 +0200
> > @@ -874,8 +874,10 @@ coff_symtab_read (minimal_symbol_reader
> > int section = cs_to_section (cs, objfile);
> >
> > tmpaddr = cs->c_value;
> > - record_minimal_symbol (reader, cs, tmpaddr, mst_text,
> > - section, objfile);
> > + /* Don't record unresolved symbols. */
> > + if (!(cs->c_secnum <= 0 && cs->c_value == 0))
> > + record_minimal_symbol (reader, cs, tmpaddr, mst_text,
> > + section, objfile);
> >
> > fcn_line_ptr = main_aux.x_sym.x_fcnary.x_fcn.x_lnnoptr;
> > fcn_start_addr = tmpaddr;
>
> That looks good to me.
Should I push it to the master branch?
> > I'm sorry, I don't have such a setup, either. Can anyone else run the
> > regression tests? If not, how are MinGW related changes tested when
> > submitted here?
>
> I thought you were the MinGW reference around here :).
I'm happy to help with this, but setting up the GDB test suite needs a
Cygwin installation as well, and I don't have that.
> > The change above causes the following behavior, when stepping out of
> > 'main' with "next":
> >
> > 6 return 0;
> > (gdb) n
> > 7 }
> > (gdb)
> > 0x00401288 in ?? ()
> > (gdb) n
> > Cannot find bounds of current function
> > (gdb) c
> > Continuing.
> > [Inferior 1 (process 9228) exited normally]
> > (gdb) q
> >
> > If this is fine with you, it's fine with me.
>
> That seems good to me, given the information GDB works with.
I agree, so I'd like to install this patch.