Message-ID: <46940E85.FF4A114D@dessent.net>
Date: Tue, 10 Jul 2007 15:56:05 -0700
From: Brian Dessent <brian AT dessent DOT net>
MIME-Version: 1.0
To: cygwin AT cygwin DOT com
Subject: Re: Starting some notes on large paths... what's the right place?(Re:    large paths in cygwin.)
References: <f65u0f$k99$1 AT sea DOT gmane DOT org> <4686950F DOT E3D9ED56 AT dessent DOT net> <f710v5$n33$1 AT sea DOT gmane DOT org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: cygwin AT cygwin DOT com
Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com

Robert North wrote:

> If I have time I may investigate what locations in code need to be
> changed, and technical details of possible internal UCS-2 (Or is that
> UTF16 ) implementations.

http://blogs.msdn.com/michkap/archive/2005/05/11/416552.aspx

From the standpoint of the low-level filesystem, it doesn't matter
whether you call it UCS-2 or UTF-16 -- it just accepts a series of wide
characters without knowledge of their semantics or validity.  Prior to
Windows XP, the entire system was surrogate-neutral, whereas starting
with XP at the shell level there was added support added for
surrogates.  So this is why some tell you that prior to XP, Windows used
UCS-2 and XP and later use UTF-16.  But this is really making a
distinction where there doesn't need to be one, since all valid UCS-2 is
also valid UTF-16, and because the filesystem is completely agnostic to
the presence or absence of surrogates, as they just appear as additional
wide characters in the filename.  And thus I suspect that this
distinction will also be irrelevant for the implementation of path
handling code in Cygwin as well.

> Ideally I'd like to place this kind of thing in a cygwin wiki,
> but as Cygwin doesn't have one,  my preference is to place the notes in 
> this mailing list, where they will be searchable.
> Is this a good idea?
> Is there a better place to put them?
> --such as the developer's list?

The project has traditionally relied on the mailing list as the sole
place for information for the simple reason that having something like a
wiki would just mean that users would start using it to ask questions
and they would go unanswered because monitoring two places is twice the
work of monitoring one place.  Or they would ask something and a random
uninformed stranger would reply with something totally misleading or
wrong and it would go unchecked because we're only reading the mailing
list.  I can totally see the following edits happening on a wiki page:

"Does anyone know how to get ssh daemon working on windows server 2003? 
---foo"

"Yes, give up, it doesn't work because Cygwin sucks and is broken on
2k3?  ---bar"

If this nonsense had been uttered on the mailing list the user would
have been told the correct steps (run ssh-host-config) and that 2k3 is
fully supported and works fine.

Anyway, for collecting information I suppose a thread on this list is
the best place.  The developers list isn't the right place for random
musings and such, but if you have a specific implementation plan (i.e.
you're referring directly to parts of the code) then it would probably
be appropriate there.

Brian

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/