[postgis-devel] [geos-devel] GEOS C++ API deprecated?

Greg Troxel gdt at lexort.com
Sun Oct 8 17:04:15 PDT 2017


Mateusz Loskot <mateusz at loskot.net> writes:

>> It is a problem to have APIs that are not stable, regardless of
>> language.  Others, once they realize this, will try not to use APIs at
>> all :-)
>
> Greg, I do understand it. I still keep my position (no longer relevant
> for the GEOS decision outcome)
> Let me summarise my point,
>
> I develop a software. I use lib X. I use whichever lib X's API I prefer.
> Someone spots my software and packages.
> After a few released versions, lib X breaks API.
> The old version works for me, so I stick to it.
> Now, I'm the troublemaker.
> Maintainer of the package is trying to force my project to switch.
> She/He fails to convince me and turns the spears towards the lib X.
> Now, the lib X is the bad guy.

> Now, that is a very unacceptable practice for me, no matter how high and
> good are intentions of the maintainer.

I see your point.

>From the packaging point of view, one wants to update X, and keep your
software.  And not to keep multiple versions of libraries.

The combination of API changes, packages that don't have releases with
the new API, packaging systems not wanting to have multiple versions of
many packages, and wanting to keep as much software packaged as
possible, leads to not having a solution in some cases.  I think your
preference is just to drop the package that doesn't build anymore,
without malice, and without complaining.  That seems totally reasonable
and I think I'm coming around to that, since it's essentially what would
have happened in the end, without as much angst.

> That, combined with some smaller incidents or commentary
> directed to such disobedient open source project(s)
> may raise the blood pressure quite a lot.

Understood.  I'm sorry if I contributed to that; I was really just trying
to figure out how to decide how much pain to impose on users of which
package, given that it wasn't really possible to make everything happy
at once.

> However, in GEOS, I acknowledge there is significant difference of
> opinions here and, without any window dressing, I've now respected
> that.
>
> Greg and Bas, any disprespect that I might have generated against you
> was unintentional, please accept my apologies.

No problem - while it's clear there was disagreement, I always believed
everyone was working towards the overall interests of the Free Software
world.  Thank you for taking the time to explain your thoughts.

Greg



More information about the postgis-devel mailing list