From mboxrd@z Thu Jan 1 00:00:00 1970 From: julia.lawall@lip6.fr (Julia Lawall) Date: Mon, 7 Apr 2014 17:34:01 +0200 (CEST) Subject: [Cocci] How to exclude volatile data accesses in expressions with SmPL? In-Reply-To: <5342C434.6020805@users.sourceforge.net> References: <5307CAA2.8060406@users.sourceforge.net> <53342A9C.9070406@users.sourceforge.net> <5334705A.9050303@users.sourceforge.net> <5342BF46.8030405@users.sourceforge.net> <5342C434.6020805@users.sourceforge.net> Message-ID: To: cocci@systeme.lip6.fr List-Id: cocci@systeme.lip6.fr On Mon, 7 Apr 2014, SF Markus Elfring wrote: > >> How do you think about to make it possible to check attributes for expressions > >> in SmPL constraints? > > > > I tried to do this at one point, but it introduced a lot of parsing > > problems, because attributes can appear in a variety of places. > > Which kind of syntax had you got in mind for such a functionality? > > Is there a need to distinguish property checks from support for extensions from > a popular compiler implementation? > http://gcc.gnu.org/onlinedocs/gcc-4.8.2/gcc/Type-Attributes.html > > > > The benefit didn't seem to be worth the risk. > > Can we clarify your concerns a bit more here? To my recollection, there were problems of ambiguity in the parser. Also, I think it made the abstract syntax tree unwieldy, because attributes had to be taken into account in many positions. julia