Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT sources DOT redhat DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin/>
List-Post: <mailto:cygwin AT sources DOT redhat DOT com>
List-Help: <mailto:cygwin-help AT sources DOT redhat DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-owner AT sources DOT redhat DOT com
Delivered-To: mailing list cygwin AT sources DOT redhat DOT com
From: Chris Faylor <cgf AT cygnus DOT com>
Date: Wed, 27 Sep 2000 00:27:10 -0400
To: cygwin AT sources DOT redhat DOT com
Subject: Re: gmake is slow under cygwin
Message-ID: <20000927002710.A7341@cygnus.com>
Reply-To: cygwin AT sources DOT redhat DOT com
Mail-Followup-To: cygwin AT sources DOT redhat DOT com
References: <20000926194812 DOT D6082 AT cygnus DOT com> <3644885960 DOT 970012527 AT dsl-64-34-95-237 DOT telocity DOT com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.3.6i
In-Reply-To: <3644885960.970012527@dsl-64-34-95-237.telocity.com>; from juliano@cs.unc.edu on Tue, Sep 26, 2000 at 11:55:27PM -0400

On Tue, Sep 26, 2000 at 11:55:27PM -0400, Jeffrey Juliano wrote:
>--On Tuesday, September 26, 2000 Chris Faylor <cgf AT cygnus DOT com> wrote:
>> On Tue, Sep 26, 2000 at 01:57:40PM -0600, Scott Carter wrote:
>>> There is no long delay. When VPATH is set to this (which is incorrect)
>>>
>>> VPATH = //cygdrive/c/temp/test/source
>>
>> Hmm.  I can see why that would cause a huge delay.  This is a very good
>> point.  Thanks for bringing it up.
>>
>> I can easily see someone translating //c to //cygdrive/c and seeing the
>> delay the you describe.
>
>Would it make sense to treat `//cygdrive/[a-zA-Z]' as a special case 
>prefix, and not try to actually find a machine named `cygdrive'?  I suspect 
>this is the direction Chris was thinking.  Of course, the special case 
>logic would have to honor the customizable cygdrive prefix.

No.  I wasn't thinking that way at all.  If someone wants to use //cygdrive
as a machine name I'm not going to stop them.

>Looks like, currently, you're allowed to double-up slashes anywhere except 
>for the first slash.  e.g. `/usr//local' but not `//usr/local'.  That's not 
>commensurate with other unix systems I've tested.

You're not running on UNIX.  You're running on Windows.  That means that
you have a filesystem which is (by default) case preserving but case
insensitive.  You can't create filenames with ':' or '\' in them.  There
are special files called "aux", "com1", "com2", "prn", "lpt1", etc.

We could work around all of those problems with some kind of kludge or
other but I don't think it is a good idea at all.  It would complicate file
parsing for, IMO, little gain.  The net result would be that you'd have
files that were only usable under Cygwin and we'd constantly be explaining
that fact.

So, I'm not going to stop Cygwin from interpreting the leading double
slash or backslash the same way as Windows.  It's part of Windows
proscribed method for accessing remote shares and AFAIK, it's not
improper to treat two leading slashes as "special" anyway.  Windows is
not the only OS which has this distinction.

>My proposed change would make path handling even less symmetric, so perhaps 
>it's not a good idea afterall.  Is this even worth discussing?

No, I don't think so.  I'm not really going to be swayed on this issue.

cgf

--
Want to unsubscribe from this list?
Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com