Greg Hudson <ghudson@MIT.EDU> writes:
> How can anyone feel this way?
>
> equal-rites% svn log -r1 http://svn.collab.net/repos/svn/branches/1.0-stabilization
> svn: REPORT request failed on '/repos/svn/!svn/bc/1/branches/1.0-stabilization'
> svn: '/repos/svn/!svn/bc/1/branches/1.0-stabilization' path not found
Our real problem with error messages isn't with the fact that they
don't always follow proper conventions. It's with the fact that
they're often completely disconnected from their ultimate source :-).
Issue #1652 doesn't address that.
> Granted, #1652 doesn't fix the really egregious problems like this.
Right :-).
> But
> our error messages are not "fine"; they are really bad. Given that 1652
> has been verified by two people to only change the text of error
> messages and only in helpful ways, I don't see why anyone would want to
> object to it.
I think I understand what's going on here...
You saw this in the STATUS file
"(seem to be doing fine with our current msgs so far)"
and, not unreasonably, interpreted it to mean that gstein and I think
our error messages are fine. Don't worry, we don't think that! What
we meant was: the things that the #1652 patch improves are already in
an acceptable state, so #1652 is not compelling for 1.0 IOHO.
Also, the patch is huge and widespread, so it may take a significant
amount of time to apply (it will probably conflict syntactically with
other changes).
It is true that many of our error message *are* really bad, but we
have other issues filed about that kind of badness (#1254, for
example). The patch in #1654 doesn't fix this. In most places, the
patch is about conforming to conventions, though in some places it
also adds a little more information.
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Jan 6 22:23:33 2004