| commit | 59f7b5a4086f6c3f3268055688c2b405a4ea2170 | |
| author | Dan Carpenter <dan.carpenter@oracle.com> | |
| Thu, 30 Jul 2015 10:37:22 +0000 (30 13:37 +0300) | ||
| committer | Dan Carpenter <dan.carpenter@oracle.com> | |
| Thu, 30 Jul 2015 10:37:22 +0000 (30 13:37 +0300) | ||
| tree | 6791e7dded203c929c3df1e0c7805f8ab8d488af | treesnapshot (tar.gz zip) |
| parent | 0ee2670447927b7498fd7b1756d4b35b322c8b98 | commitdiff |
ranges: don't allow so many inverted ranges
It turns out that we were getting a lot of ranges like 0-(-1). Part of
the problem was the nested struct issue fixed in 0ee267044792 ('type:
improve get_member_type_from_key()'). But also when we add a range we
should do it with add_range_type(). I moved that function around to
avoid the forward declaration.
If that fails, and we still get and inverted type then just use the whole
range at that point.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
It turns out that we were getting a lot of ranges like 0-(-1). Part of
the problem was the nested struct issue fixed in 0ee267044792 ('type:
improve get_member_type_from_key()'). But also when we add a range we
should do it with add_range_type(). I moved that function around to
avoid the forward declaration.
If that fails, and we still get and inverted type then just use the whole
range at that point.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
| smatch_ranges.c | diffblobblamehistory |