Re: [Gnu-arch-users] tla update

[Michael Teichgräber]

Denys Duchier <address@hidden> writes:

> Did it actually revert anything, or this just a heading where
> reverted changes would be listed if there were any? If the latter,
> then either omitting it or printing something like "(none)" after it
> would be less confusing.

One could test in arch_undo() whether all rel_table members of struct
arch_make_changeset_report have zero elements, and in this case print
a different message, e.g. "* no changes to revert\n", and perhaps not
revert and reapply the empty changeset at all.

In such a case, and if only messages that are written by the update
command would start in column 0, it could look like this:

* setting aside local changes temporarily
  + linting the source tree
  + computing changeset   *)
  + no changes to revert

* updating for new patches in archive
  + patching for revision address@hidden/mozart--main--0--
    A   ./{arch}/mozart/mozart--...-ebuilds/patch-log/patch-5
    M   ./mozart-1.2.5-r4.ebuild

-- 
Michael

[*] "computing changeset" won't be printed in the most current tla version