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

Re: Proposed resolution: Standardizing on UTF-8 isn't enough

From: B. Smith-Mannschott <benpsm_at_gmail.com>
Date: 2007-07-19 13:42:04 CEST

On 7/19/07, Erik Huelsmann <ehuels@gmail.com> wrote:
> On 7/19/07, Justin Erenkrantz <justin@erenkrantz.com> wrote:
> > On 7/18/07, Erik Huelsmann <ehuels@gmail.com> wrote:
> > > (1) Recode all paths on Mac to NFC and assume all other systems submit
> > > NFC without checking.
> >
> > +1.
> >
> > "The perfect is the enemy of the good."
> >
> > #1 solves the current problem.
> >
> > -1 (vote, not veto) on #2 and #3.
> >
> > > Option (1) doesn't have this effect: MacOS X has functions built in to
> > > normalize to NFC. No additional dependencies would be required
> > > anywhere.
> >
> > Which is why I favor #1 by a wide margin.
> >
> > > At least, that's what I was going to propose until I realized there
> > > were mixed client version concerns. Now, I think the only option is to
> > > go with (2).
> >
> > I tend to disregard this aspect - it would just be broken right now,
> > so I'm not concerned about how we deal with older versions.
>
> We would be breaking a mixed-version setup in an all-Mac project,
> since when all users commit NFD paths, there's no problem. After this
> change, some clients would recode to NFC and others would submit
> NFD...

I see it that way too. It would just be replacing a {Win,Linux}/Mac
incompatibility with a Mac[1.4]/Mac[1.4+n] incompatibility. That's a
little better, but not much.

The only way to handle that would be to provide tools to normalize the
paths in existing repositories. And those would again require a
library (e.g. ICU) to do proper NFC normalization, unless we're
willing to say the tool for *server-side* normalization will only work
on a mac.

So, would we be really saving ourselves from a dependency by choosing
the "easy" solution?

// bsmith@occs

p.s. IMHO, since Subversion has made the choice to use unicode (UTF-8)
internally, it should do so *properly*. Perhaps it's overkill to do so
in connection with this particular defect, but it should be done.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Jul 19 13:41:20 2007

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

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.