On Mon, Oct 2, 2023 at 2:53 PM Tom de Vries via Gdb-patches <
gdb-patches@sourceware.org> wrote:

> Factor out usage of cooked_indexer::m_die_range_map into new class
> parent_map
> with member functions find_parent and set_parent.
>
> Tested on x86_64-linux.
> ---
>  gdb/dwarf2/cooked-index.h | 22 ++++++++++++++++++++++
>  gdb/dwarf2/read.c         | 23 +++++++++++++++++------
>  2 files changed, 39 insertions(+), 6 deletions(-)
>
>
> It seems breakpoint-locs.exp test does not fail anymore with this patch
applied on ppc64le Fedora Rawhide. But I wouldn't expect that
from just a factoring out. Is that a coincidence?
< Breakpoint 1 at 0x10010208: file
/root/build/gdb/testsuite/../../../binutils-g
db/gdb/testsuite/gdb.cp/breakpoint-locs.h, line 23.
< (gdb) FAIL: gdb.cp/breakpoint-locs.exp: break N1::C1::baz
< testcase
/root/build/gdb/testsuite/../../../binutils-gdb/gdb/testsuite/gdb.cp/breakpoint-locs.exp
completed in 0 seconds
---
> Breakpoint 1 at 0x10010208: N1::C1::baz. (2 locations)
> (gdb) PASS: gdb.cp/breakpoint-locs.exp: break N1::C1::baz
> testcase
/root/build/gdb/testsuite/../../../binutils-gdb/gdb/testsuite/gdb.cp/breakpoint-locs.exp
completed in 1 seconds