This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH] Fix up msymbol type of dll trampoline to mst_solib_trampoline
- From: Tom Tromey <tromey at redhat dot com>
- To: Yao Qi <yao at codesourcery dot com>
- Cc: <gdb-patches at sourceware dot org>
- Date: Wed, 10 Jul 2013 10:56:33 -0600
- Subject: Re: [PATCH] Fix up msymbol type of dll trampoline to mst_solib_trampoline
- References: <1372043502-4618-1-git-send-email-yao at codesourcery dot com> <874ncjmgkl dot fsf at fleche dot redhat dot com> <51CD0054 dot 9040401 at codesourcery dot com> <87ehbmkzqr dot fsf at fleche dot redhat dot com> <51D36FB3 dot 4070006 at codesourcery dot com>
>>>>> "Yao" == Yao Qi <yao@codesourcery.com> writes:
Yao> That is, after the hash table of minsyms of OBJFILE is built up, we
Yao> can iterate all minsyms, if symbol is _imp_x, look up x in the hash
Yao> table. If found, modify the found's type.
Yao> + for (i = objfile->minimal_symbol_count; i > 0; i--)
This should use ALL_OBJFILE_MSYMBOLS.
I'm mildly concerned that this exposes an implementation detail of the
minsym storage -- namely, it assumes that it is ok to modify a minsym
after the minsym is installed.
This approach would also block constification of the minsym API.
Neither of these seem like blocking considerations though. Presumably
bugs arising from the first would be caught in testing; and for the
second, casting away const would be obviously ok.
So, ok with the ALL_OBJFILE_MSYMBOLS change.
Tom