On Wed, 19 Feb 2020, Alfred M. Szmidt wrote: > What do people think of these ideas? > > Under `Character Set', I would be against it. Comments, and other > such text, is much better suited if it is ASCII. What would be the > actual usage? The prefered language is English, and that is purley > ASCII, _if_ there is a need to explicitly use some non-ASCII char. we > already recommend UTF-8.
There is one specific case where use of ASCII is desirable even if UTF-8 is otherwise being used: in installed headers that users might include from their own source files. The GCC -finput-charset option applies to all headers read (there's no good way to distinguish, in general, headers that are part of the package being compiled from headers from other packages), and installed headers should be ASCII so they work when included in a source file compiled with -finput-charset= naming any ASCII-compatible character set. -- Joseph S. Myers [email protected]
