[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: svn commit: r11509 - branches/1.1.x

From: Ben Reser <ben_at_reser.org>
Date: 2004-10-20 19:49:30 CEST

On Wed, Oct 20, 2004 at 07:23:44PM +0200, Branko ??ibej wrote:
> No it't not. The RFC is explicit in stating that URLs using the file:
> schema can have two forms:
>
> file:///path
> file://localhost/path
>
> (that is, the host name part must be either empty or "localhost"). We've
> supported both in the past, therefore we should support both now.

We still support both. However, there was a small error where we
weren't removing a duplicate path separator properly in some URLs.

My patch should have disallowed accepting a leading // if the path was a
uri. I fixed this in r11514.

> The form file://machine/path violates the RFC if machine isn't "localhost".

No it doesn't. Go read the RFC. A blank name or localhost are a
special case. We've only ever supported the special case. Now that
it's okay to put fsfs repos on a share we should probably start
supporting the non-special case.

> But they're closer to being valid URLS according to the RFC.

Well I guess the RFC doesn't really speak to duplicate path separators.
But I think it's bogus to shove a UNC path into a file URL when the file
URLs already have a valid method of representing the data and shoving
them in the way the URLs I said I'd veto implies a local path not a path
to a remote machine.

-- 
Ben Reser <ben@reser.org>
http://ben.reser.org
"Conscience is the inner voice which warns us somebody may be looking."
- H.L. Mencken
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Oct 20 19:50:00 2004

This is an archived mail posted to the Subversion Dev mailing list.