This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug c/18624] GCC does not detect local variable set but never used
- From: "dcb314 at hotmail dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 11 Jun 2009 12:57:08 -0000
- Subject: [Bug c/18624] GCC does not detect local variable set but never used
- References: <bug-18624-9596@http.gcc.gnu.org/bugzilla/>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Comment #14 from dcb314 at hotmail dot com 2009-06-11 12:57 -------
(In reply to comment #13)
> We need someone to write the patch. Just that.
I've got some spare time now, so I'd like to have a go.
I can see that used_flag bitfield can be expanded from 1 bit to 2 bits.
I can see that TREE_USED macro can be reworked to match.
What concerns me is all the different places (currently 170) that the
text pattern "TREE_USED.*=" occurs.
I suspect a manual review of each of those 170 places would be
prone to failure. Even without considering the various languages
in sub directories, there are still 84 uses of "TREE_USED.*=" left.
Considering solely the test program I originally posted, how
would I go about finding which subset of the 84 places get
visited by the test program ?
gdb with 84 breakpoints doesn't sound feasible to me.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18624