On Tue, Dec 20, 2016 at 8:42 AM, Zoltán Herczeg <hzmes...@freemail.hu> wrote: >>>>> Although "compile_bracket_matchingpath" does overwrite "current->top" >>>>> on some paths, it also contains at least one feasible path which does not >>>>> overwrite it. > > I think it expects that current->top must not be NULL. This isn't true, it > can be NULL which represents a "leaf" object in a tree like structure.
So perhaps it can't deduce some implication about current->top being NULL or not NULL in certain paths. Probably a few assertions in the code would shut these warnings. >>>> * it's worth to set up regular scans of it? I can do it weekly. > > I don't mind if you do it. But please check the output manually first and > don't forward a huge report every week for us :) Hum... unfortunately that's outside my control, Coverity sends such reports automatically after a build. But I don't expect them to be huge: the reports are about *new* stuff found in a given scan. For Qt we're talking about ~10 issues found every couple of weeks, so something totally manageable. Thank you, -- Giuseppe D'Angelo -- ## List details at https://lists.exim.org/mailman/listinfo/pcre-dev