Joshua Paine wrote about having a script recursively add and remove to
make the repository reflect what is currently on disk (please correct
me if this isn't a fair summary!).
While this might be sufficient for some usage patterns, I don't think
this is a great solution in general. The major problem I see is that
once a file is renamed externally to fossil, you don't get any
tracking of that in the repository as (I believe) you would get by
issuing a "fossil mv <old-name> <new-name>" command. Instead the
fossil history will contain a "fossil rm <old-name>" and a "fossil add
<new-name>".
Whether this is a huge loss of functionality depends on your usage. An
obvious problem is merging between leaves where in one, a file has
been renamed.
Aside – I think this approach is taken by the "Git" SCM, but it then
has lots of support for reconciliation of where files might have come
from. It's all based on heuristics and reasonable guesses, but I
imagine it must work reasonably well.
Cheers,
Benjohn
_______________________________________________
fossil-users mailing list
[email protected]
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
