[Haskell-cafe] Is bumping the version number evil, if it's not mandated by the PVP?

[Ross Paterson]

On Sat, Aug 14, 2010 at 11:13:19AM +0200, Sebastian Fischer wrote:
> I wonder whether (and how) I should increase the version number of a
> library when the API does not change but the implementation gets
> more efficient.
> 
> Should I bump a.b.C or even a.B to signal that it's worth using the
> new version or should I bump only a.b.c.D such that packages that
> depend on a.b get installed with the new version automatically?
> 
> When bumping only a.b.c.D, the new version is not installed as a
> dependency if the old version already is installed (unless the new
> version is explicitly demanded.) It seems bumping a.b.c.D has
> advantages for some users and disadvantages for others.

How would bumping the major version change that?  Let's suppose a
package depending on your declares a version constraint a.b.* (which
would be good practice).  If you bump the patch level, anyone installing
the other package with get the latest version of your package; if you
bump the major version, they won't, until the maintainer of the other
package releases a new version to update their dependencies.  (Some
maintainers might be annoyed by that; if you annoy maintainers too
much, they might look for less annoying dependencies.)

Whenever you bump a version number, you are signalling that the new
version is worth using.  If you bump the patchlevel, you're also
signalling that the new version is compatible with the old, so using
it is a no-brainer.