bzr resolve

[Bosco Rama]

On 07/26/12 08:59, Aaron Bentley wrote:
> On 12-07-26 10:43 AM, Stephen J. Turnbull wrote:
>> Nevertheless, I think the rationale for "bzr resolve FILE" not
>> checking by default (at least FILE itself) is pretty weak.
> 
> What is the rationale for specifying FILE when you don't want to
> override the automatic behaviour?

It's more for adherence to the "first, do no harm" precept.  If, for
example, I fumble-finger a glob:
   bzr resovle m *.c

instead of:
   bzr resolve m*.c

However, if the resolve command was dangerous in all its forms
and came with a big warning:
    "WARNING: kills angels and makes your toes fall off!!"

(or words to that effect :-p )

then, I may be able to see the case for not detecting that the user
had undone the changes that the merge system did.

Maybe it's just me but it still feels strange for a system to alter
a file, flag it as conflicted and then make no attempt to at least
check that the file's conflicts had been addressed.

Bosco.