[Svnmerge] Bugs when merging revisions that modify and deletefiles

[Giovanni Bajo]

On 1/10/2007 8:15 PM, Jack Repenning wrote:

>> Is there something *really* risky about using "svn merge --force"? I 
>> can't think of any, especially for clean working copies (like those 
>> required by svnmerge.py to operate).
> 
> Well, "especially for clean wc's" is of course an added assumption, 
> something additional that the users would have to learn, and hence a 
> source of surprises for the uninitiated.

They don't have to learn this: svnmerge.py refuses to operate by default 
if it detects local changes in the working copy. So, in a way, it's 
something users learn very very early.

> The risk is that the behavior with "--force" is silent, while the 
> behavior without it does announce itself.  Some people really *want* to 
> review the output line-by-line, and they want that easy-to-miss 
> "skipping" message because the no-message-at-all case is even easier to 
> "miss."

Uhm. I think I shall rephrase my question. Which scenarios do you 
devise, given a clean working copy, in which the "skipping" message 
might appear during a svnmerge merge operation?
-- 
Giovanni Bajo