On Tue, 2004-02-24 at 13:16, kfogel@collab.net wrote:
> Greg Hudson <ghudson@MIT.EDU> writes:
> > I'd like to commit my transactionless reporter code, which I described
> > in
> > http://subversion.tigris.org/servlets/BrowseList?list=dev&by=thread&from=143289
> > and put in the issue tracker in
> > http://subversion.tigris.org/issues/show_bug.cgi?id=1679
> > a couple of months ago.
> >
> > There are no API changes involved; it's just a performance
> > improvement, and there should be no problem to have it in svn 1.1.
> >
> > Object now or don't.
>
> You're talking about replacing our current reporter.c with the new
> code, so that the transactionless reporter becomes the new default?
>
> I don't object to this; I already reviewed the patch back when you
> first posted it. Spooling the file also feels funny to me, but if
> pressed I have to admit that it's not worse than spooling into the
> database.
>
> However, it might be wise to add this as a conditional part of the
> build first, e.g.,
>
> $ ./configure --enable-txnless-reporter
>
> ...or something, so we can get some more testing first. After a few
> weeks, we could switch over completely.
I'm -0 on this. Similar configure time switches have been proposed in
the past, and they've been shot down. We want everyone working on the
bleeding edge (i.e. trunk) to be testing the fixes and features that go
into Subversion--adding a switch to enable something means that it's not
going to be tested as widely.
-Fitz
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Feb 25 01:46:33 2004