[OpenWrt-Devel] [PATCH] package: use PKG_VERSION only for upstream packages

[Adrian Schmutzler]

> -----Original Message-----
> From: Paul Oranje [mailto:por at oranjevos.nl]
> Sent: Sonntag, 9. August 2020 11:02
> To: Adrian Schmutzler <freifunk at adrianschmutzler.de>
> Cc: OpenWrt Development List <openwrt-devel at lists.openwrt.org>
> Subject: Re: [OpenWrt-Devel] [PATCH] package: use PKG_VERSION only for
> upstream packages
> 
> Op 24 feb. 2020, om 17:08 heeft Adrian Schmutzler
> <freifunk at adrianschmutzler.de> het volgende geschreven:
> >
> > In the package guidelines, PKG_VERSION is supposed to be used as "The
> > upstream version number that we're downloading", while PKG_RELEASE is
> > referred to as "The version of this package Makefile".
> > Thus, the variables in a strict interpretation provide a clear
> > distinction between "their" (upstream) version in PKG_VERSION and
> > "our" (local OpenWrt trunk) version in PKG_RELEASE.
> 
> As still some discussion exists about how to use PKG_VERSION and
> PKG_RELEASE, currently around local projects, a thought of some time ago
> might be worth consideration.
> 
> The wiki states that PKG_RELEASE must be reset to 1 whenever
> PKG_VERSION is updated [1], which makes it subordinate to the upstream
> version of the package. Would just keep incrementing PKG_RELEASE not be
> less entangled (of orthogonal purposes) ?

I personally think the current solution of resetting PKG_RELEASE is more helpful.
Having PKG_RELEASE orthogonal would establish two parallel versions of the same package, an internal and an external one. IMO this would just create more confusion.

> 
> Also, might it be an idea to rename PKG_VERSION to
> PKG_UPSTREAM_VERSION ? (or something alike but more concise) That
> would make its meaning much more obvious.

We already have PKG_SOURCE_VERSION, which however describes the commit hash.
Therefore, I think going anywhere close to that with the current PKG_VERSION variable name again would just increase confusion, and not reduce it.

Apart from that, changing PKG_VERSION variable name would be a major PITA considering how abundant and widely distributed it is. I'd only do that for a very, very good reason.

Essentially, I personally think the two variables are actually fair to handle and understand if you care and read the guidelines. And with the multitude of variables for BuildPackage to cover all the cases in place, I don't think we will be able to get a solution which can be understood without looking at the docs at all anyway.

So, we should just keep it as is IMO.

Best

Adrian

> 
> [1] https://openwrt.org/docs/guide-developer/package-policies
> 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: openpgp-digital-signature.asc
Type: application/pgp-signature
Size: 834 bytes
Desc: not available
URL: <http://lists.openwrt.org/pipermail/openwrt-devel/attachments/20200811/6f26a316/attachment.sig>