On Tue, Jun 17, 2014 at 3:18 PM, Paul Franklin <pf.98052@gmail.com> wrote:
> I thought it breaks the logic of plugins check!

That's the impression I have too.

But I wasn't around when the plugin system was started
and so I don't claim to be able to accurately reproduce the
rationale behind it.   Or debate it from all points of view.

However I do have the distinct impression that the concept
of having to manually force such a version-number increase
was integral to the whole philosophy, to magically somehow
force developers into making sure that older things were
upgraded, when necessary.

But as I pointed out over a year ago, IMHO that is not done
at all and instead a brute-force upgrade of the number is done
instead, always -- just like we are talking about doing now.

Including on the gramps-addons repo, where it is even more
likely that older things will not have been updated.

So as everybody knows I'd like to see the whole plugin system
drastically changed, but as everybody also knows as a junior
developer I have no chance at all to cause that to happen.

The addon numbering was designed to only allow the right versions of plugins to work. For example:

* if you just take some random plugin/addon and drop it in, it probably won't work. This was happening a lot in the old days as people were sharing their homemade plugins, and they didn't work. We can now tell why it doesn't work... oh, it was designed for 2.4.
* addons can be made to only work to a point, or only after a point (for example, after 5.2.3)

* more specific versions will override less specific versions

Yes, the numbering must be bumped up on each major.minor version release. That can be as easy as as "sed -i ..." or might involve each addon being examined, and updated. 

I would say that we try to pick the best system, regardless if it comes from a "junior developer" (there is no such thing).

If you have a better idea, please make a GEP, write some code, and put it out there as a pull request :)



HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
Find What Matters Most in Your Big Data with HPCC Systems
Open Source. Fast. Scalable. Simple. Ideal for Dirty Data.
Leverages Graph Analysis for Fast Processing & Easy Data Exploration
Gramps-devel mailing list