Grzegorz Jaskiewicz <gj@pointblue.com.pl> writes:
> usual round of updates to the scan report.
> Today's report available at:
> http://zlew.org/postgresql_static_check/scan-build-2009-09-12-1/
Looks like the clang guys still have some work to do. The
null-dereference reports, in particular, seem to be willing to make
self-contradictory assumptions in order to claim there is a possibility
of a null dereference. The clearest example I found was this one:
http://zlew.org/postgresql_static_check/scan-build-2009-09-12-1/report-Ybdv3J.html#EndPath
where to conclude that lp might be null, clang first assumes
PageGetMaxOffsetNumber(page) < offnum (at line 4251); but it then
must assume that that is *false* in order to suppose that control
can arrive at the dereference inside ItemIdIsNormal at line 4254.
regards, tom lane