Re: shells/oh-my-posh

[Jonathan Perkin <jperkin%mnx.io@localhost>]

* On 2023-03-21 at 11:30 GMT, pin wrote:

> > No.  That's only for if a package will only ever during its entire lifetime
> > ...
> > It also means that the failures never show up in bulk builds so people are
> > not made aware of them and be able to do something about it.
>
> I understand this reasoning, in that case something like BROKEN_ON_PLATFORM ?

No, just leave it.

> It's simply weird to have a package in pkgsrc proper without a hint that it
> doesn't build on NetBSD. At least a commented note on the package Makefile.
>
> It really doesn't matter what/how but, something. Right now, it's deceiving.

Go and look at the pkgsrc-bulk mailing list.  There are thousands of 
broken packages across our supported operating systems.  Maintaining 
accurate ONLY_*/BROKEN for every package would literally be a full time 
job for a large team of people, and if you aren't going to do it for all 
packages then you've introduced inconsistency and confusion into pkgsrc.

Build failures are a _good_ thing.  It shows exactly what the state of a 
platform is.  It encourages users and/or maintainers to dig in and fix 
things.  It helps those who might be interested in getting involved to 
see in advance what the problem is.  It provides a way for those who are 
seeing failures to double check against another build to see if they are 
seeing the same failures.  It gives upstream maintainers visibility into 
how portable their software is, and what the issues are.

Masking failures is always a terrible idea.

--
Jonathan Perkin   -   mnx.io   -   pkgsrc.smartos.org
Open Source Complete Cloud   www.tritondatacenter.com