This is the mail archive of the mailing list for the glibc project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFC] Canonical Standards Names

On Wed, 21 Jun 2017, Joseph Myers wrote:

> What I'm getting at is: since SVID and BSD are not standards supported by 
> glibc (do not have any feature test macros distinct from _DEFAULT_SOURCE), 
> should they be documented at all in @standards rather than MISC or 
> DEFAULT?  Likewise for XPG[123], since _XOPEN_SOURCE was used without 
> versioning for all of XPG[1234] so it's not possible to support the older 
> versions in the headers.  Likewise anything else that doesn't map to 
> feature test macros.

To put this another way: there are three things we need to decide: (a) 
what standards need names for @standards at all (e.g., do BSD and SVID 
need separate names); (b) given which standards need names, what those 
names should be; (c) how @standards should describe the standards with an 
interface when it's not simply one standard and, by implication, all those 
that are supersets (or nearly supersets) of it.  (a) is important before 
we can know what an answer for (b) should look like.  For (c) I think 
C-style && || ! expressions is the obvious answer, with details such as 
how everything is presumed to be visible in GNU unless stated explicitly 

(Note for (c) the manual would always describe what standards define the 
interface to be in particular headers; it's possible the headers also make 
the interface visible in cases where it's permitted but not required to be 
in those headers.)

Joseph S. Myers

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]