Naming conventions for defining macros

Naming conventions for defining macros

[Philipp Stephani]

Hi,

the manual page about coding conventions says: "Constructs that define a
function or variable should be macros, not functions, and their names
should start with `define-'." This recommendation is at odds with the usual
recommendation to prefix all non-local names with the package name and
doesn't seem to be widely followed by either Emacs (e.g. ert-deftest,
cl-defstruct) or popular third-party packages (e.g. magit-define-command).
Should we get rid of this recommendation? It amounts to special-casing a
certain class of names without much benefit.

Thanks,
Philipp

Re: Naming conventions for defining macros

[Barry Margolin]

In article <mailman.727.1424760332.31049.help-gnu-emacs@gnu.org>,
 Philipp Stephani <p.stephani2@gmail.com> wrote:

> Hi,
> 
> the manual page about coding conventions says: "Constructs that define a
> function or variable should be macros, not functions, and their names
> should start with `define-'." This recommendation is at odds with the usual
> recommendation to prefix all non-local names with the package name and
> doesn't seem to be widely followed by either Emacs (e.g. ert-deftest,
> cl-defstruct) or popular third-party packages (e.g. magit-define-command).
> Should we get rid of this recommendation? It amounts to special-casing a
> certain class of names without much benefit.

I think you're expected to combine the conventions, hence 
magit-define-command. It obeys the recommendation to use the package 
name prefix, then the define- prefix.

-- 
Barry Margolin, barmar@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***

Re: Naming conventions for defining macros

[Philipp Stephani]

Barry Margolin <barmar@alum.mit.edu> schrieb am Di., 24. Feb. 2015 um
16:45 Uhr:

> In article <mailman.727.1424760332.31049.help-gnu-emacs@gnu.org>,
>  Philipp Stephani <p.stephani2@gmail.com> wrote:
>
> > Hi,
> >
> > the manual page about coding conventions says: "Constructs that define a
> > function or variable should be macros, not functions, and their names
> > should start with `define-'." This recommendation is at odds with the
> usual
> > recommendation to prefix all non-local names with the package name and
> > doesn't seem to be widely followed by either Emacs (e.g. ert-deftest,
> > cl-defstruct) or popular third-party packages (e.g.
> magit-define-command).
> > Should we get rid of this recommendation? It amounts to special-casing a
> > certain class of names without much benefit.
>
> I think you're expected to combine the conventions, hence
> magit-define-command. It obeys the recommendation to use the package
> name prefix, then the define- prefix.
>
>
Instead of trying to formalize this convention, would it do much harm to
get rid of the 'define-' convention altogether? AFAIK there's no other kind
of function with a conventional midfix; the other conventions are either
suffixes (e.g. '-hook') or prefixes (the package name).