Stephan Witt wrote:
> > It would work but it implies external svn call(s) for each opened files no 
> > matter whether the file
> > is under the control :-/ I'm not much happy, but you are the boss now...
> 
> I don't like hard-coding the knowledge how subversion developers organize 
> their meta-data.
> The overhead of forking a process and loading the shared code for the svn 
> client is not too high, IMHO.
> The traversal to the root of the file system is done by the svn client too.
> And if there is some .svn directory upstairs we have to fork anyway.

I understand your reasoning. Now compare with the situation that you have doc 
with 10 childern (I have documents with higher numbers)
and for each you call svn and git binary for nothing... We can try your way and 
check how much people would scream, but I'm afraid
nothing for stable branch :)

Pavel

Reply via email to