Thorsten Behrens wrote:
Stephan Bergmann <[EMAIL PROTECTED]> writes:
The most pragmatic approach (and the one at least I have taken during
the last couple weeks when working on warnings01) is probably to not
touch the headers delivered from external projects, but instead use
pragmas at the places where those headers are included to disable
warnings from within the headers. In case a delivered header is
included in multiple places, this can be made DRY by creating a new
header that does the pragma stuff and then includes the original
header.
Hm - are you suggesting to switch off the warning temporarily (while
the problematic header gets parsed), or completely for the compilation
unit including this header?
The former. You are right, there are cases where this does not work,
but in those cases where it does work, I think it is a reasonable approach.
-Stephan
I think both ways have problems - the first one, because it doesn't
work for some of gcc's warnings in template code (the warning fires
only at the time of template instantiation), the second one because it
potentially makes the whole effort moot - just imagine including on of
the infringing boost headers in a central cppuhelper include.
Puzzled,
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
