On Mon, Aug 24, 2009 at 17:55, Hyrum K.
Wright<hyrum_wright_at_mail.utexas.edu> wrote:
> On Aug 23, 2009, at 7:32 PM, Greg Stein wrote:
>> On Thu, Aug 20, 2009 at 15:30, Hyrum K. Wright<hyrum_at_hyrumwright.org>
>> wrote:
>>>
>>> ...
>>> +++ trunk/subversion/libsvn_wc/entries.c Thu Aug 20 12:30:08 2009
>>> (r38890)
>>> ...
>>> @@ -3264,7 +3263,7 @@ svn_wc__walker_default_error_handler(con
>>> {
>>> /* Note: don't trace this. We don't want to insert a false "stack frame"
>>> onto an error generated elsewhere. */
>>> - return err;
>>> + return svn_error_return(err);
>>> }
>>
>> Really?! Come on, man. Read the comment **right above that return**
>
> Maybe the comment is bogus. I actually found the included stack frame to be
> valuable in attempting to debug a problem. But in the end, I suppose I just
> like my bike sheds chartreuse.
The comment isn't bogus.
The error is generated by A, wrapped by B, wrapped by C. And then D
takes it and asks somebody "wanna do anything with this?", then
returns a wrapped error. So you get D(C(B(A))). The "somebody" is not
in the call chain. They're "over there".
Cheers,
-g
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2386816
Received on 2009-08-27 00:05:50 CEST