sshd on win2k [revisit]

sshd on win2k [revisit]

[Harry Putnam]

Its been a while since I checked on the availability of packages that
installs sshd on a win2k.

Is it still amatter of tinkering or is there a package out that just
installs it, and allows win2k to run a sshd daemon?

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

more info on cvs

[Heitzso]

After reading FAQ/Doc I'm noting that goal of Cygwin is
for all programs to run under cmd.exe with backslashed file names
just fine.

I assume that cvs code needs a patch to convert
backslashed names to forward slashes or to
change the test in last_component() to find
backward as well as forward slashes.

Note I'm not the cvs cygwin maintainer.
I emailed the maintainer awhile ago
and haven't heard back.

Should I go into cvs tree and dig out
cvs patch code and get setup to alter
the cvs cygwin patch?

Thanks,
Heitzso


--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: more info on cvs

[Charles Wilson]

Technically, I think the cvs protocol itself expects that paths are
specified using backslash.  However, certain implementations/ports of
cvs to native windows have relaxed that, so that Windows users can use
"normal" backslashes when specifying pathnames.  Then, netbeans/forte
took that laxness as license (or perhaps the developers at Sun just
said, "Hey, windows -- paths have backslashes -- okay, we'll do that
too")

What we're left with is this: we can either (a) limit pathnames to using
only '/' as specified by the spec (which also, BTW, explicitly *allows*
'\' as a normal char in a filename !!!), or (b) violate the spec and
adapt cvs on cygwin to explicitly convert '\' to '/' every time.

The problem is, some remote system to which your local cygwin-cvs client
talks, might have a file with a *legal* name containing '\'.  What do we
do then?

I'm in favor of (b), and treat the "problem" above as another
windows-ism (like the fact that "aux" is an illegal name on windows but
is FINE everywhere else on the planet...)

--Chuck

P.S. Sorry for being slow to respond, Heitzso, but I've been very busy
with (a) my job (b) cygwin release 1.3.0^H1, (c) ncurses release 5.2-5
lately. :-)  I "maintain" about 20 packages and it's getting to be
pretty overwhelming...

Heitzso wrote:
> 
> After reading FAQ/Doc I'm noting that goal of Cygwin is
> for all programs to run under cmd.exe with backslashed file names
> just fine.
> 
> I assume that cvs code needs a patch to convert
> backslashed names to forward slashes or to
> change the test in last_component() to find
> backward as well as forward slashes.
> 
> Note I'm not the cvs cygwin maintainer.
> I emailed the maintainer awhile ago
> and haven't heard back.
> 
> Should I go into cvs tree and dig out
> cvs patch code and get setup to alter
> the cvs cygwin patch?
> 
> Thanks,
> Heitzso
> 
> --
> Want to unsubscribe from this list?
> Check out: http://cygwin.com/ml/#unsubscribe-simple

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: more info on cvs

[Christopher Faylor]

On Mon, Apr 30, 2001 at 03:48:32PM -0400, Charles Wilson wrote:
>Technically, I think the cvs protocol itself expects that paths are
>specified using backslash.  However, certain implementations/ports of
>cvs to native windows have relaxed that, so that Windows users can use
>"normal" backslashes when specifying pathnames.  Then, netbeans/forte
>took that laxness as license (or perhaps the developers at Sun just
>said, "Hey, windows -- paths have backslashes -- okay, we'll do that
>too")
>
>What we're left with is this: we can either (a) limit pathnames to using
>only '/' as specified by the spec (which also, BTW, explicitly *allows*
>'\' as a normal char in a filename !!!), or (b) violate the spec and
>adapt cvs on cygwin to explicitly convert '\' to '/' every time.
>
>The problem is, some remote system to which your local cygwin-cvs client
>talks, might have a file with a *legal* name containing '\'.  What do we
>do then?
>
>I'm in favor of (b), and treat the "problem" above as another
>windows-ism (like the fact that "aux" is an illegal name on windows but
>is FINE everywhere else on the planet...)
>
>--Chuck
>
>P.S. Sorry for being slow to respond, Heitzso, but I've been very busy
>with (a) my job (b) cygwin release 1.3.0^H1, (c) ncurses release 5.2-5
>lately. :-)  I "maintain" about 20 packages and it's getting to be
>pretty overwhelming...

I'd actually be in favor of the "do nothing since it isn't broken"
scenario.

While the Cygwin DLL does try to do the right thing with MS-DOS paths I
don't see any urgent need to adapt every tool that comes with Cygwin to
do the same.  The point of Cygwin is that it is supposed to provide a
UNIX environment.  I certainly don't want to see package maintainers
spending their precious time fiddling with sources to accomodate MS-DOS
(IMO) brain-damage.

Whether you do take time to do this is, of course, up to you Chuck.
I appreciate your maintainership very much but I don't want to add
to your burden unduly.

cgf

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: more info on cvs

[Charles Wilson]

Charles Wilson wrote:
> 
> Technically, I think the cvs protocol itself expects that paths are
> specified using backslash.  However, certain implementations/ports of
                  ^^^^^^^^^
        I meant, forward slash. '/'

--Chuck

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: more info on cvs

[Christopher Faylor]

On Mon, Apr 30, 2001 at 02:29:35PM -0400, Heitzso wrote:
>After reading FAQ/Doc I'm noting that goal of Cygwin is for all
>programs to run under cmd.exe with backslashed file names just fine.

Cygwin does work fine with backslashed file names.  This does not mean
that individual utilities understand backslashed file names.  If the
utility comes from UNIX it is unlikely that it will handle backslashed
filenames correctly.

We don't guarantee that every package maintainer will take the
not-insubstantial amount of time required to adapt a package to MS-DOS
pathnames.

cgf

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: sshd on win2k [revisit]

[Dmitrii Pasechnik]

On 30 Apr 2001, Harry Putnam wrote:
>
>Its been a while since I checked on the availability of packages that
>installs sshd on a win2k.
>
>Is it still amatter of tinkering or is there a package out that just
>installs it, and allows win2k to run a sshd daemon?
>
it is still quite far from automatic.
As you know, there are two separate problems there:
1) giving  the right permissions to the users;
2) installing sshd as a service.

Regarding 2), I saw at least one GPL-ed program that happily
installs and manages itself as a service, etc, namely WinVNC
(see http://www.uk.research.att.com/vnc/ )
It most probably could be used as a code sample to build
such a thing for sshd.
 
Could be a nice project for a student, or for someone with a knowledge of
Win32 registry programming...

-- 
Dmitrii Pasechnik
e-mail: d.pasechnik@twi.tudelft.nl 
http://ssor.twi.tudelft.nl/~dima/


--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: sshd on win2k [revisit]

[Charles Wilson]

Dmitrii Pasechnik wrote:
> 
> On 30 Apr 2001, Harry Putnam wrote:
> >
> >Its been a while since I checked on the availability of packages that
> >installs sshd on a win2k.
> >
> >Is it still amatter of tinkering or is there a package out that just
> >installs it, and allows win2k to run a sshd daemon?
> >
> it is still quite far from automatic.
> As you know, there are two separate problems there:
> 1) giving  the right permissions to the users;
> 2) installing sshd as a service.
> 
> Regarding 2), I saw at least one GPL-ed program that happily
> installs and manages itself as a service, etc, namely WinVNC
> (see http://www.uk.research.att.com/vnc/ )
> It most probably could be used as a code sample to build
> such a thing for sshd.

Corinna's version of inetd also does this.  I (actually not me, but I
was an accessory to the crime) stole that code for ipc-daemon and now IT
does this too.  The problem, IIRC, is that Corinna is trying to keep the
OpenSSH code as close to the official version as possible -- thus, most
changes are submitted right back to the openSSH people for official
inclusion.

But the strange hacks necessary for windows service installation are
meeting with, err, resistance.  So, either we maintain an official but
external patch, or we just continue status quo.

Is that about right, Corinna?

> Could be a nice project for a student, or for someone with a knowledge of
> Win32 registry programming...

--Chuck

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: sshd on win2k [revisit]

[Corinna Vinschen]

On Mon, Apr 30, 2001 at 04:10:40PM -0400, Charles Wilson wrote:
> Dmitrii Pasechnik wrote:
> > 
> > On 30 Apr 2001, Harry Putnam wrote:
> > >
> > >Its been a while since I checked on the availability of packages that
> > >installs sshd on a win2k.
> > >
> > >Is it still amatter of tinkering or is there a package out that just
> > >installs it, and allows win2k to run a sshd daemon?
> > >
> > it is still quite far from automatic.
> > As you know, there are two separate problems there:
> > 1) giving  the right permissions to the users;
> > 2) installing sshd as a service.
> > 
> > Regarding 2), I saw at least one GPL-ed program that happily
> > installs and manages itself as a service, etc, namely WinVNC
> > (see http://www.uk.research.att.com/vnc/ )
> > It most probably could be used as a code sample to build
> > such a thing for sshd.
> 
> Corinna's version of inetd also does this.  I (actually not me, but I
> was an accessory to the crime) stole that code for ipc-daemon and now IT
> does this too.  The problem, IIRC, is that Corinna is trying to keep the
> OpenSSH code as close to the official version as possible -- thus, most
> changes are submitted right back to the openSSH people for official
> inclusion.
> 
> But the strange hacks necessary for windows service installation are
> meeting with, err, resistance.  So, either we maintain an official but
> external patch, or we just continue status quo.
> 
> Is that about right, Corinna?

Yep. And since I have no problems to install sshd as a service in about
5 minutes I'm suffering an extrem letargy to change that. Personally
I hate the special service code. I'm very happy with the SRVANY solution.

Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Developer                                mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

RE: sshd on win2k [revisit]

[Michael Erdely]

Especially since Karl found the "-D" switch and you can stop the service
now too!

-ME

-----Original Message-----
From: cygwin-owner@sources.redhat.com
[ mailto:cygwin-owner@sources.redhat.com ] On Behalf Of Corinna Vinschen
Sent: Wednesday, May 02, 2001 9:23 AM
To: cygwin@cygwin.com
Subject: Re: sshd on win2k [revisit]

<snip>

Yep. And since I have no problems to install sshd as a service in about
5 minutes I'm suffering an extrem letargy to change that. Personally I
hate the special service code. I'm very happy with the SRVANY solution.

Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Developer                                mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple



--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple