Hi there,
there is another thread happening which is suggesting or proposing a
future version of "fossil rm" which is not identical to the current one.
What is the specific desired future behaviour?
I think there are three possible future commands to be considered --
whether implemented as commands or as arguments is not of concern here.
repo-only-rm -- mark as "not part of the project"
repo-and-file-rm -- repo-only-rm, plus remove from the filesystem if safe
repo-and-file-really-rm -- repo-only-rm, plus remove from the filesystem
if possible
Have I missed, or mischaracterised, any? (file-only-rm presumably remains
outside of the fossil tool.)
Currently, "repo-only-rm" is spelled "rm"; and the other two do not exist
within fossil, so to achieve them one must use external scripting which
reimplements at least part of the fossil logic to determine which files
is a subdirectory should be considered. Avoiding the need to reimplement
that logic is probably a good thing.
Consider a straightforward case; files only (no directories) and nothing
in ignore-glob:
echo X >a; echo X >b; echo X >c;
fossil add a b
fossil commit -m "a and b = X"
echo Y >b; echo Y >c;
fossil new-rm a
fossil new-rm b
fossil new-rm c
fossil commit -m "all gone"
What is the intended behaviour of each of the new-rm commands here?
Which, if any, of the new-rm commands fail?
If commands fail, what advice is given on how to make them succeed?
What files exist in the filesystem after them?
If files are gone, is there some way I can recover the "Y" for b and c?
Thanks,
f
--
Francis Daly [email protected]
_______________________________________________
fossil-users mailing list
[email protected]
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
