[bug 0.11] commit succeeds with conflict markers in the file

[Jari Aalto]

John Arbash Meinel <john at arbash-meinel.com> writes:

> My personal feeling is that '--force' is a big hammer, that I want to be
> careful we don't train users to get too used to using. Because it may be
> used in commands that are pretty major things. Like bzr uncommit --force
> used to actually remove the revision data from the store without
> prompting, which was a very destructive operation. Which is a much
> bigger --force than 'commit --force'.
> 
> I would be happier with 'commit --allow-conflicts'. Though I think our
> proposed method of checking at 'bzr resolve' time is a much better fix.

Please do not consider some unknown/(yet another new) option like
--allow-conflicts. The UI should stay clean and use same commands or
options to mean same things. Also the operations used by the other
tools should be consulted  as a reference.

The --force is commonly known and does what the user intends: "Do what
I say/I know what I'm doing". I don't consider it at all "hammer",
but a way to bypass any "smart checks" what a program might be
guessins or preventing to do.

Jari