Detection of reverted commits

[shful]

The changelog may contain commits that have been reverted.
A warning line is added on each of these commits.

Why ? Before we publish, one team member will revise and rework the changelog.
The warning aids him to not accidentially advertise a reverted feature.

[coveralls]

Coverage decreased (-0.5%) to 91.647% when pulling fcb3240 on shful:master into e9759c2 on vaab:master.

[vaab]

Thanks, I've read the code, and understood what this does. A test would have been welcome, as a doc. But before working on that, I suggest you to wait, because I need to think about whether it is a good idea to add this as you did.
What could bother me is that we are adding adhoc code for specifics and this has a limit: we already have too much variables to my taste. Your implementation is fine, don't get me wrong, it follows the current scheme. I'm just wondering if this is the right way to go to keep the maintenance head-ache free.

I was thinking that this type of modification should be possible via the filters (it is not currently feasible because we don't have access to the full commit object in filters, and we probably could). This would allow your whole PR code to be a single filter function, with the addition of being very easily configurable, or removed...

BTW, forcing the unrolling of the log generator is not a good idea. The generator is meant to manage huge list of commits that wouldn't fit in memory. And I don't really understand why you need to make this regex lookup at once before the big loop. Sorry if I missed anything obvious.

[shful]

You are right; I see that GitRepos.log() guarantees reverse chronological order. I wasn't aware of that before, and I will eliminate the log unrolling.
This ordering fits well with any filters that use information from younger commits to manipulate older ones, as I did here.

I also think that a generic solution like filters is preferrable. My gut feeling is that more specific extensions are about to come.
As soon as the concept seems right, I'll add test coverage.

[coveralls]

Coverage decreased (-0.6%) to 91.566% when pulling 2dcfb8f on shful:master into e9759c2 on vaab:master.

[coveralls]

Coverage decreased (-0.6%) to 91.566% when pulling d6da1db on shful:master into e9759c2 on vaab:master.