Mercurial > hg
annotate .editorconfig @ 33196:439b4d005b4a
tests: demonstrate inconsistencies with dirty state in various commands
Not only is the output of these commands inconsistent with respect to each
other when a file is deleted, they are internally inconsistent depending upon
whether the deleted file is in the top level repo or a subrepo. It seemed
easier to show the problems, rather than describe them. The original goal was
to fix the summary command with respect to deleted files. I haven't fixed any
of the other issues yet, in case anybody believes the current subrepo behavior
is correct.
I think a natural understanding of clean/dirty is that they are two opposite
values of a single binary repo state. If `hg update --clean -r .` changes a
file, then naturally that repo was dirty, and `hg update --check` should have
blocked it. Deleted files are special, in that they don't block a commit. But
they make the filesystem content not the same as a clean checkout.
author | Matt Harbison <matt_harbison@yahoo.com> |
---|---|
date | Wed, 28 Jun 2017 21:30:46 -0400 |
parents | d30fdd6d1bf7 |
children | 1d6066336d7b |