bug#11230: 24.1.50; VC doesn't handle internal errors?

[Glenn Morris]

Deniz Dogan wrote:

>> Maybe Emacs should test that the option works, not just that it is
>> present. I don't know; it's a slightly more expensive test. (Make a temp
>> repo, check status, delete it.)
>
> I am not that familiar with VC or Bazaar, but is it really necessary
> to do all of that?  Couldn't we simply check the return code from
> every Bazaar command and act depending on the result of that?

We're talking at cross purposes. I was talking about setting the default
value of vc-bzr-status-switches more rigorously.

The vc-bzr-status command cannot return an error, because it is invoked
all the time, on simply visiting a file in a bzr-controlled directory.
So an error here might make it impossible to edit a file; or even to
build Emacs [1]. The best Emacs can do is convert the error to a message
and report it, which is what it does already. I vaguely recall that it's
a design decision of vc-dir that it puts any error in the actual vc-dir
buffer. A multi-line backtrace from a crashing external program is not a
typical expected case.

[1] http://lists.gnu.org/archive/html/help-gnu-emacs/2012-04/msg00145.html