Johan Corveleyn wrote on Wed, Jun 01, 2011 at 16:14:32 +0200:
> On Wed, Jun 1, 2011 at 4:10 PM, Julian Foad <julian.foad_at_wandisco.com> wrote:
> > On Wed, 2011-06-01 at 01:39 +0200, Johan Corveleyn wrote:
> >> On Wed, Jun 1, 2011 at 1:22 AM, Stefan Sperling <stsp_at_elego.de> wrote:
> >> > On Wed, Jun 01, 2011 at 01:15:56AM +0200, Johan Corveleyn wrote:
> >> >> [[[
> >> >> The usual namespace rules apply: only names that begin with "SVN_"
> >> >> and don't contain double underscores are considered part of the public
> >> >> API. Everything else is not officially supported.
> >> >> ]]]
> >> >>
> >> >> Thinking: "I don't think this should be part of the public API", so I
> >> >> can just dispense with the "SVN_" prefix.
> >> >
> >> > In this case I would advocate for SVN__DISABLE_PREFIX_SUFFIX_SCANNING :)
> >> > You're right that it doesn't need to be part of the public API.
> >>
> >> I went for SVN_DISABLE_PREFIX_SUFFIX_SCANNING in the end (r1129965).
> >> Who knows, maybe it can prove useful someday, if some special build
> >> needs to disable this f.i. because it has problems with the seeking
> >> around of the prefix/suffix scanning (jumping to the end of the file,
> >> then jumping to the start again) or something ...
> >
> > I think the implication that these #defines are considered part of the
> > public API is not true.
> >
> > This one's purpose is definitely private (debugging/testing/optimising);
> > there's no point in speculating that it might turn out to be useful.
>
> Ok, I guess you're right. But then I'll still leave it with "SVN_"
> prefix, because that doesn't imply anything (as you say), and it's
> consistent with most other #defines in knobs.
If it's not public, then IMO it should get the double-underscore treatment.
The goal is the same as for API functions: to make it easier to see from
the name of a symbol whether it's a public API or not.
Received on 2011-06-01 16:32:26 CEST