Align version with Engine about oq-hazardlib HOT 5 OPEN

LGTM

from oq-hazardlib.

I do agree that the hazardlib is stable enough to have a version >= 1.0.

I am not sure the version of the hazardlib should be equal to that of the engine. If we make a change to the engine without changing the hazard lib do we intend to change the version number of the hazardlib without any other changes?

from oq-hazardlib.

That's a good point @pslh but it never happened in 4 years. We always had versions aligned. We should consider pros and cons of both the options.

We need to take into account that "hazardlib" includes baselib, which is a fundamental part of the engine and changes to the engine often require changes to baselib. Am I wrong @micheles ?

from oq-hazardlib.

The dependencies of hazardlib all apply to the engine but not vice-versa. If we add support for different parallel processing frameworks, I could imagine that we might well change the engine without changing the underlying libraries.

If it is really true that a new oq-engine version implies a new hazardlib version and vice-versa, perhaps we should review our decision to have them in separate repositories.

It is possible to use the hazardlib without the engine, so it should be possible to download/install the hazardlib without having to download/install the engine.

We expect code changes in hazardlib not to break the engine, so hazardlib developers already need both repositories.

from oq-hazardlib.

it is really true that a new oq-engine version implies a new hazardlib version and vice-versa, perhaps we should review our decision to have them in separate repositories.*

I have been pointing out this for years. I do not remember any major version release having hazardlib changes independent from engine changes.

from oq-hazardlib.