This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [patch] gdb_assert -> complaint for weird DWARF
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: Doug Evans <dje at google dot com>
- Cc: gdb-patches <gdb-patches at sourceware dot org>
- Date: Tue, 25 Feb 2014 09:13:48 +0100
- Subject: Re: [patch] gdb_assert -> complaint for weird DWARF
- Authentication-results: sourceware.org; auth=none
- References: <20140224214314 dot GA5700 at host2 dot jankratochvil dot net> <CADPb22RpU2pMtWnPN-Rmc8J1sQDS6-dsUQ4e39WsJi+QKK6K7A at mail dot gmail dot com> <20140224220439 dot GA7121 at host2 dot jankratochvil dot net> <CADPb22SWQMhT0VQ=ryAQd_a4LNMTeGfke+qcgwewxaS3_qQKAg at mail dot gmail dot com>
On Mon, 24 Feb 2014 23:59:50 +0100, Doug Evans wrote:
> Can you send me the binaries for repro?
It was sent in the previous mail:
http://people.redhat.com/jkratoch/gcc-debuginfo-4.8.2-7.fc20.x86_64-gnatbind.debug
> We could probably generate a good testcase for gdb from that.
I was thinking about it but:
* There is not much to test gdb_assert vs. complaint.
* Currently I believe the generated DWARF is incorrect. Detecting that such
incorrect DWARF is identified as an incorrect one has limited sense IMO.
So far I find it just a GCC bug.
> Another worry I have is that if my expectation that we shouldn't be
> recursively calling process_die (even for bad debug info) is wrong,
> then is there some obscure case where possible accidental re-reading
> of a DIE is actually needed by the current code to get the right
> answer (IOW is making this a complaint and returning also introducing
> a bug? Less of a bug than crashing or infinite recursion of course,
> but IWBN to invest some time to dig deeper given that we have a repro
> at hand).
As I said so far I do not find the DWARF (in the gnatbind case) to be
meaningful.
Jan