On 15 Jan 2015 at 8:11, Beman Dawes wrote:
Should Boost have policy to clear these warnings?
Thoughts?
Personally I'd be opposed to setting such a policy for all libraries. Speaking for myself, I intentionally shadow variable names as I find it makes the code clearer e.g. when catching exceptions, I catch to a variable named e. I also, in implementation code, sometimes use macros which hard code the use of certain variable names always being a certain meaning e.g. in AFIO, the size_t variable named id always refers to the "current" op id code, where "current op" can be a bit fuzzy when you're actually dealing with several ops at the time. I have no opposition to each library setting its own policy here. Niall -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/