The latest version of stash.test exposes a quirk in the stash diff
command, that occurs after stashing a fossil rm.
Boiled down, I did the following in stash.test. First get a new
repository with files f1, f2, and f3, and commit them:
repo_init
write_file f1 "f1"
write_file f2 "f2"
write_file f3 "f3"
fossil add f1 f2 f3
fossil commit -m "c1" --tag c1
Add a new file f0, and edit f2:
write_file f0 "f0"
fossil add f0
write_file f2 "f2.1"
Now snapshot the changes and try stash diff to see that the folder and
the stash agree:
fossil stash snapshot -m "snap 1"
fossil stash diff
fossil stash show
Remove f1, and snapshot:
file delete f1
fossil rm f1
fossil stash snapshot -m "snap 2"
fossil stash diff
Note that diff fails, and says:
DELETE f1
no such file: c:/Users/Ross/Documents/tmp/fbuild/stash/f1
fossil stash show shows a sensible collection of changes in the snaphot.
I also renamed f3 and took another snapshot. Since f1 is still missing,
I still see the same error from diff, but the final fossil stash show is
also sensible.
Is it sensible to have stash diff fail because a file that the stash was
going to delete is already deleted? I was expecting that stash diff
would show that the working directory already has all the stashed
changes applied, not an error.
The version of stash.test as of [94b95307] demonstrates what I tried.
--
Ross Berteig r...@cheshireeng.com
Cheshire Engineering Corp. http://www.CheshireEng.com/
+1 626 303 1602
_______________________________________________
fossil-dev mailing list
fossil-dev@mailinglists.sqlite.org
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/fossil-dev
