Re: [patch] Speed up find_pc_section

[Paul Pluzhnikov <ppluzhnikov at google dot com>]

On Fri, Sep 11, 2009 at 1:36 AM, Pierre Muller <muller@ics.u-strasbg.fr> wrote:

> Maybe you could alter the complaint so that it becomes
> clear which of the two sections is discarded from the mapping?

Patch with all the suggestions so far incorporated.
Tested on Linux/x86_64 with no regressions.

Here is what the complaint looks like for gdb.base/solib-overlap.exp
(if I turn complaints on):

  During symbol reading, unexpected overlap between (A) section
`.hash' from `/usr/local/google/gdb/testsuite/gdb.base/solib-overlap-lib1-0x40000000.so'
[0x158, 0x19c) and (B) section `.hash' from
`/usr/local/google/gdb/testsuite/gdb.base/solib-overlap-lib2-0x40000000.so'
[0x158, 0x19c). Will ignore section B.

Thanks,
-- 
Paul Pluzhnikov

2009-09-11  Paul Pluzhnikov  <ppluzhnikov@google.com>

     * objfiles.c (qsort_cmp): Remove asserts.
     (insert_section_p, filter_debuginfo_sections): New function.
     (filter_overlapping_sections): Likewise.
     (update_section_map): Adjust.

Attachment: gdb-find_pc_section-20090911.txt
Description: Text document