On Sat, Sep 17, 2016 at 06:52:44AM +0000, Daniel Shahaf wrote:
> I think end-users do care about the error code. We print it out in the
> command-line tools too and one of the reasons for that is so scripts can
> parse it.
>
> Moreover, as per the TODO's in the patch, this is groundwork for passing
> the error (code|chain) over the wire, which would allow libsvn_ra/libsvn_client
> consumers to distinguish "post-commit hook error" from "post-commit FS
> processing error" through means more robust than strstr(). (We already
> have code to marshal svn_error_t over ra_svn.)
OK, so you've intentionally expanded your original problem definition.
Given that, I believe I understand where you're going with this and
I agree it would be nice to have.
Received on 2016-09-17 09:24:06 CEST