On Mon, 25 Apr 2005, [UTF-8] Branko Ä^Libej wrote:
> Peter N. Lundblad wrote:
>
> >On Fri, 22 Apr 2005, [UTF-8] Branko �^Libej wrote:
> >
> >
> >
> >>I vote for fixing this bug, not preserving it indefinitely. Our
> >>compatibility rules don't say that we have to maintain bugs that people
> >>happen to be relying on, because they didn't read the documentation.
> >>
> >>
> >>
> >What exactly do you want to be fixed? We currently support
> >file://machine/share syntax, and happen to support file:///\machine/share.
> >file:////machine/share isn't supported since one slash is dropped by our
> >canonicalization routine.
> >
> >The problem with removing support for file:///\achine/share is that it was
> >the only way to use ra_local on a network share in 1.1. So we will break
> >peoples workng copies. Even if they can just switch --relocate, they will
> >be annoyed by this.
> >
> >
> We did not document file:///\... as being supported in 1.1, and we did
> say not to use ra_local on network mounted filesystems. I see no
Where did we say not to use FSFS repositories over ra_local on a network
share? Note that according to Simon Large, the TSVN people were
"recommended" to use this workaround in 1.1. I don't know where that mail
is though.
> particular reasons to keep supporting a bug that happened to be "used"
> by people who don't believe our documentation.
But you seem to see a reason not to answer my first question. Exactly
what's the bug you want fixed?
We could introduce code to detect this hack on Windows, but I don't see a
reason to annoy people just for the sake of it.
Regards,
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Apr 25 14:14:39 2005