[RFC] [Bug 363837] proposed patch

[Alexander Belchenko]

Maritza Mendez пишет:
> On Sun, Apr 19, 2009 at 11:13 PM, Alexander Belchenko <bialix at ukr.net
> <mailto:bialix at ukr.net>> wrote:
> 
>     Maritza Mendez пишет:
>     > Although I have expereinced this regression only with
>     BeyondCompare3 as
>     > my external diff tool, this regression is worth addressing for
>     four reasons:
> 
>     I wonder why you see it only with BC. Have you tried other diff tools?
>     E.g. WinMerge? Just to see is this problem specific only to BC.
>     It seems BC does not clean up on exit properly, and thus it's a bug
>     in BC.
> 
> Perhaps.  I would be cautious about rushing to a conclusion.  I have
> done quite a bit of testing of BC and it does clean up.  But there is a
> small delay between the time BC exits and Windows actually releases the
> tempdir.  It's not clear to me to what extent BC can control how fast
> Windows does that.  BC is not exactly lightweight and may leave a lot
> for Windows to clean.  I am communicating with the author of BC about
> that.  But the intrinsic race condition in bzr is catchable and
> fixable.  Whenever possible, bzr should print a friendly erorr rather
> than pass an uncaught exception.  My patch does not even do that, but it
> easily could I think.

Yes, I'm agree: at least bzr should print nicer error message.