Re: Should Inline be private in the private part of a package spec?

["Randy Brukardt" <randy@rrsoftware.com>]

"Georg Bauhaus" <rm.dash-bauhaus@futureapps.de> wrote in message 
news:5020e0e1$0$6570$9b4e6d93@newsspool3.arcor-online.net...
...
> Advances in compiler technology, or even varying the
> compilation processes, should not find a way into
> public parts of public packages made for re-use by others.
>
> package Reusable is
>
>   procedure Op (X : T)
>      with Inline => True,
>           Optimize => Time,
>        GNAT_Eliminate => Possibly,
>       Rational_Debug_Hook => Yes,
>        SofCheck_Inspectable => X;
>
> That's all relevant information in some hypothetical
> specific setup, but it is terribly at odds with re-usable Ada.
> These aspect are pragmatic, they are not relevant for re-use,
> they are extrinsic to Op's meaning.

I agree that most of this is junk, but not for the same reason. First, I 
don't think compiler-specific directives should even exist -- if they do, 
that shows a significant lack in the portable part of the language. (And 
yes, Ada has this sort of lack.) And I don't think they ought to be given 
anywhere, in any way (compiler "options" are just as harmful -- they're a 
perfect way to get vendor-lock-in.) Pragmatic considerations sometimes 
override good sense, but in that case, where or what they look like is 
already irrelevant -- you're already on the evil side of programming, 
nothing you can do will ever make these worthwhile. (Visible part, private 
part, project file, all are evil.)

The same goes for things like Inline and Pack (the latter as it is intended 
as "best efforts", and not as it is defined in Ada, which is seriously 
conflicted about what it means).

The *only* thing here should be an implementation-independent indicator of 
how this should be used (Optimization would work, but the existing choices 
don't quite give enough information. Probably a decent first cut, though.) 
And I'd argue that it has to be in the visible specification, as such a 
directive has to allow the creation of a compilation dependence (else most 
compilers could not inline).

                                          Randy.