"James Gregory" <james.jrg@gmail.com> wrote in message news:beb51c260505051117392703d0@mail.gmail.com...
You are right that there is a clash. But it seems to me std::runtime_error should be the base class, so it is the implementation that is wrong.
Comments?
--Beman
1. The boost guidelines at http://www.boost.org/more/error_handling.html say "Derive your exception class from std::exception. Except in *very* rare circumstances where you can't afford the cost of a virtual table, std::exception makes a reasonable exception base class"
2. But runtime_error does derive from std::exception, and the error is a run time error
Yes, so in future versions I'll change the derivation to runtime_error. Thanks for reporting this.
3. Whilst I'm here, I would also like to point out the following silly code I need to use boost filesystem for a program designed to run in Linux and Windows:
#ifdef __linux__ fs::path::default_name_check(fs::windows_name); #else //ie we are compiling for Windows fs::path::default_name_check(fs::native); #endif
If I use fs::native at all times then I cannot use paths with spaces in under Linux, but if I use fs::windows_name at all times then I cannot use a colon to specify drive names in Windows (e.g. C:\).
That whole problem will disappear with the coming major revision, planned for 1.34. The distinction between a native and portable path will not have to be explicitly identified. Cheers, --Beman