On Thu, Nov 01, 2018 at 06:15:28PM +0000, Ben Dooks wrote: > On 01/11/18 18:11, Ben Dooks wrote: > > I'm not reposting the parsing bit, unless people would like it > > again, it seems to have already been well reviewed. > > > > This is a new evaluate patch and the test suite to add some > > validation to go with it. It requires the ptr_list_nth_entry() > > from Luc which i've fudged into the series. I've put a patch in > > to help clean the validation output as it generates a lot of > > files. > > > > I've improved the printf handling since the previous version, > > it now runs a table of formats and has handling for positional > > arguments, widths, precisions and other fun bits of the printf > > formatting that I am becoming to think is a bit crazy. > > > > The current issues are: > > > > I've still not got the pointer evaluationm right, but it sort > > of works enough to get it tested. This will need a bit of work > > but does that need to be done for the initial merge? > > > > Do we want to have -Wformat as default for the initial merge? > > If not i'll add a patch to default it off, so we can revert it > > once this is sorted > > I forgot to mention, do you think some of the small fixes and the > parsing bit can go on to a branch for merging? It would be nice to > have these out of the way. I can already take the small fixes, everything that is stand-alone but Wformat, parsing of attribute and check/evaluation are interdependent and I'll thus not apply them separately. -- Luc
