Ping: https://gcc.gnu.org/ml/gcc-patches/2018-10/msg00473.html
On 10/08/2018 05:22 PM, Martin Sebor wrote:
Attached is an updated patch with the INTEGRAL_TYPE_P test added
to detect constant non-integer arguments like (void*)0, and with
quoting made unconditional. I also removed the pretty printer
business to avoid the "value%s" format, and modified the manual
to clarify when each of the attributes are applicable and what
their meaningful argument values are.
On 10/07/2018 04:38 PM, Martin Sebor wrote:
While still testing an enhancement in the area of attributes
I ran across bugs and inconsistencies in how different handlers
deal with positional arguments. The bugs are either an ICE due
to the handlers not consistently converting positional arguments
to constants (i.e., CONST_DEL to INTEGER_CST in C++) for which
downstream code is unprepared (PR 87541), or errors for valid
code (PR 87542), or failing to diagnose invalid arguments.
The other inconsistencies are simply in responding to the same
invalid condition differently for different attributes .
The attached patch introduces a new function to do validate
positional arguments in a uniform way and replaces the existing
handling with it.
As a consequence of the handling being made consistent a number
of tests needed adjusting. In addition, some invalid arguments
that were previously rejected with a hard error are diagnosed
with just a warning (invalid argument values in attribute format),
and in one other instance what previously triggered a warning is
now accepted without one (attribute alloc_size on a function
without a prototype). I'd be up tightening things up if that's
preferable as long it's done consistently.
Tested on x86_64-linux.
Martin
PS It would be nice to have a general policy for how to respond
to invalid attribute arguments (warning or error). Even better,
it would be ideal to move the validation from the front-ends and
back-ends into the middle-end. I.e., describe attribute arguments
in more detail in struct attribute_spec and have decl_attributes
validate nut just the number of arguments but also their types
and, when appropriate, expected values.
