Matthieu, thank you for replying.
> > Look at the initialization of `entry', a local variable.
> >
> > - (entry (if (functionp 'add-log-file-name)
> > - (add-log-file-name buffer-file file-name)
> > - (file-relative-name buffer-file (tla-tree-root))))
> >
> > =>
> >
> > + (entry (add-log-file-name buffer-file file-name))
> >
> > This change is ok?
>
> Actually strange ! Don't know where this comes from.
Ok, I recovered it in my tree(xtla--jet--0.0--patch-172).
> > In *inventory* buffer, I can read my ChangeLog with ?L.
> > However, in the the changelog, only my changelog is appeared.
>
> Yes, "tla changelog" only shows you the logs you have written.
> Alternatively, you can look at your patch-log in your {arch}/
> directory.
>
> Actually, we are abusing tla's "distributed" mechanism: The normal way
> would be to have one "main" archive, with one "main" committer, like
> Tom does for tla itself. Other contributors would only merge from him.
>
> Our way of doing distributed development is a bit more "anarchic":
> Everybody merge from everybody, and we usually do "big merges", i.e
> lots of unrelated patches in the same merge. The result is that our
> ChangeLog is not really clear ...
>
> I think we should continue like we are doing now. That's an
> interesting experiment. Maybe we should reconsider after the release
> of a 1.0, because development should be less active, but breaking a
> feature will become more critical.
I agree anarchic approach is interesting. I feel common ChangeLog is
needed in our approach. Or we have to run dired or C-x C-x in {arch}/
directory to know the intent of Changes.
