Re: [PATCH] a little step in issue 897: non-recursivesvn_handle_error()
From: Files <files_at_poetryunlimited.com>
Date: 2003-10-26 22:13:33 CET
I just noticed that while I was playing around with a few other things, I
When I retried compiling the patch my proposed patch, I noticed the problem.
Am now rerrunning compile/regression tests on my Mandrake RPMs all the way
There should not have been a declaration in the for statement - not ANSI C.
I apologize for the oversight.
Correction attached. :)
Any other comments, feel free to let me know. :) :) :) :)
-- Shamim Islam BA BS Files said: > I noticed a bug in the last posting. (there was an == instead of an !=). > > What's w/ this "print_message" parameter though. That's really not it's role. > It's role is "new_message", no? :) :) :) :) > > The outside world should know nothing of the choices made inside the black > box. It should only supply the information. > > The black box should decide what to do when it sees a new message. :) :) :) > > Also, I disagree w/ the fatal trigger to abort being in the secondary > function. The secondary function really isn't *handling* anything. It's focus > should be printing. > > Why the static function that "prints" error messages is called "handle_error" > is beyond me. I've called it "print_error". It really doesn't > open/close/finalize anything. Or did I miss that part? > > And lastly, I keep thinking that the debug pretty printing should happen in > the main handler, since it really doesn't make use of the buffers and what > not, but I think that's 6 of one and half dozen of the other. > > I do agree that the functions need to be small. I also agree w/ you Greg that > the changes need to be idiomatic. Esp. in reference to the use of 'for'. > > That being said, here is my proposal. I've attached a bz2 file to preserve > formatting. If you like, great. If not, it was fun to do. :) :) :) :) >
This is an archived mail posted to the Subversion Dev mailing list.