Re: Treat "Circular dependency...dropped" as a hard error instead?

[JZB]

On 5/11/24 06:11, Dmitry Goncharov wrote:
> On Wed, May 8, 2024 at 2:04 PM JZB <jzb@z2zcorp.com> wrote:
> > ... perhaps a command line option or a ".CIRCULARERROR:" rule
> > or some way to allow this to be turned on by those who need it?
>
> There is currently no mechanism (option or otherwise) to fail on a
> circular dependency.
> On top of that, there are situations when make drops a circular
> dependency silently.
> See https://savannah.gnu.org/bugs/?60736.
>
>
> regards, Dmitry

Thanks Dmitry, that provides some background.

To be honest, I don't quite understand Paul's logic, in that his 
approach to a global .EXTRA_PREREQS silence should be the exception, not 
the default.  I can screw-up a global .EXTRA_PREREQS just as readily as 
I can a target-specific one, maybe even more so.

However, if there existed a rule, such as what I proposed, a 
".CIRCULARERROR:", and it should be global, never target specific 
(should be easier to implement), because I either care about circular 
dependencies being errors or I don't (and that is project-wide, not 
per-target), and have that take precedence over the 
silent/not-silent-warning issue, it would actually be a compromise 
between the two sides on

> https://savannah.gnu.org/bugs/?60736.

In other words, current behaviour stays as Paul has it, but if the user 
wants circular dependencies to be errors, then Dmitry's algorithm has 
prioriry, and yes, that should include .EXTRA_PREREQS as well.


--jzb