This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [gimple] assignments to volatile


This discussion appears to have gone quiet. From the table in http://gcc.gnu.org/wiki/VolatileAccessComparison I think it clear that:

* GCC's current behaviour is inconsistent

* Historically GCC's behaviour has changed. GCC 4.3's behaviour is particularly unfortunate WRT C and C++ consistency.

* The 2 other compilers tested (I'm lumping all the EDG-based ones together) implement different access patterns. One of which appears to be the same as the current GCC development semantics.

* I missed another disambiguating testcase, namely '0, vobj = x;' in C the result of the comma operator is not an lvalue -- its result has the same properties as the result of the conditional operator. That argues for them behaving the same -- which currently does not happen in GCC.

I think there is also consensus that what the standard literally says are not useful semantics.

Is there consensus on what the semantics should be?

Any suggestions as to how to move forward?

nathan

--
Nathan Sidwell    ::   http://www.codesourcery.com   ::         CodeSourcery


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]