This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [patch 2/2] Accelerate lookup_symbol_aux_objfile 8x
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: Doug Evans <xdje42 at gmail dot com>
- Cc: "gdb-patches at sourceware dot org" <gdb-patches at sourceware dot org>
- Date: Thu, 23 Oct 2014 20:29:23 +0200
- Subject: Re: [patch 2/2] Accelerate lookup_symbol_aux_objfile 8x
- Authentication-results: sourceware.org; auth=none
- References: <20141020214500 dot GC22011 at host2 dot jankratochvil dot net> <CAP9bCMTUYw0zLjpxNZ=Ya17BOOmhaivkbkbMd++CZi3WXjO=fA at mail dot gmail dot com>
On Wed, 22 Oct 2014 11:40:18 +0200, Doug Evans wrote:
> When I think of primary and non-primary symtabs I think of .c/.cc
> (primary) and .h (non-primary).
> Now we've got "secondary" dwz-created symtabs.
dwz-created symtabs are primary, they can contain symbols (although that
happens rarely).
> Therefore while this patch avoids searching dwz secondary symtabs
> multiple times, we will now search symtabs for headers whereas before
> we wouldn't (and thus search symtabs for .c/.cc files an extra time
> for each included header).
Addressed in the new patch sent in the other mail, thanks.
> [Our symbol table data structures and code sucks in so many major ways.
> What'd I'd like to see happen is some deeper changes, and have some
> planned when I get the time, but I'm not going to impose that on this
> patch set. :-)]
There are many possible improvements of GDB as usual. Just I find this
patchset already done and providing in fact good-enough-for-me acceleration.
Jan