Would nobody please finish what he's doing.

Would nobody please finish what he's doing.

[Alan Mackenzie]

Hi, Emacs!

I've fixed a bug in cc-mode.el (hi, Martin!), and want to commit it to
the EMACS_22_BASE.  Being the good boy that I am, I've also added an
entry to emacs/lisp/ChangeLog (EMACS_22_BASE branch).

On trying to commit this ChangeLog, I get this error message:

cvs commit: [10:40:08] waiting for nobody's lock in /cvsroot/emacs/emacs/lisp

, which is repeated every thirty seconds to indefinite times.  This part
of Savannah has been in this state for over twelve hours.  I've submitted
a bug report to savannah.

Can anybody shed light on nobody, and on what's going on here?

Thanks!

-- 
Alan Mackenzie (Ittersbach, Germany).

Re: Would nobody please finish what he's doing.

[martin rudalics]

 > I've submitted
> a bug report to savannah.

Thanks to you!  Could someone please post here as soon as
they got rid of nobody?

Re: Would nobody please finish what he's doing.

[Eli Zaretskii]

> Date: Tue, 19 Jun 2007 13:19:08 +0100
> From: Alan Mackenzie <acm@muc.de>
> 
> cvs commit: [10:40:08] waiting for nobody's lock in /cvsroot/emacs/emacs/lisp
> 
> , which is repeated every thirty seconds to indefinite times.  This part
> of Savannah has been in this state for over twelve hours.  I've submitted
> a bug report to savannah.
> 
> Can anybody shed light on nobody, and on what's going on here?

AFAIK, `nobody' is the name of anyone who uses anonymous CVS access.

Re: Would nobody please finish what he's doing.

[Stefan Monnier]

>> cvs commit: [10:40:08] waiting for nobody's lock in /cvsroot/emacs/emacs/lisp
>> 
>> , which is repeated every thirty seconds to indefinite times.  This part
>> of Savannah has been in this state for over twelve hours.  I've submitted
>> a bug report to savannah.
>> 
>> Can anybody shed light on nobody, and on what's going on here?

> AFAIK, `nobody' is the name of anyone who uses anonymous CVS access.

If so, maybe it'd be a good idea to ask the savannah people to change the
anonymous access so that it doesn't lock.  I've seen such "nobody's lock"
several times in the past (usually it's just transient: someone is in the
process of doing an anonymous checkout or update).

The downside of course would be that such an anonymous checkout may result in
an "inconsistent" state, although IIRC the potential inconsistencies aren't
very serious (i.e. it will never see an incomplete RCS file).


        Stefan

Re: Would nobody please finish what he's doing.

[Juanma Barranquero]

On 6/19/07, Stefan Monnier <monnier@iro.umontreal.ca> wrote:

> > AFAIK, `nobody' is the name of anyone who uses anonymous CVS access.
>
> If so, maybe it'd be a good idea to ask the savannah people to change the
> anonymous access so that it doesn't lock.

It's not just nobody. I've been seeing this one since yesterday:

cvs commit: [15:53:33] waiting for sds's lock in /cvsroot/emacs/emacs/src

             Juanma

Re: Would nobody please finish what he's doing.

[Andreas Schwab]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> If so, maybe it'd be a good idea to ask the savannah people to change the
> anonymous access so that it doesn't lock.

Every CVS operation needs a lock, if only to make sure that nobody is
modifying the meta data while you are reading it.  *Note (cvs)Locks::.
*Note (cvs)Concurrency::.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: Would nobody please finish what he's doing.

[Stefan Monnier]

> Every CVS operation needs a lock, if only to make sure that nobody is
> modifying the meta data while you are reading it.  *Note (cvs)Locks::.
> *Note (cvs)Concurrency::.

There's theory and then there's practice.
As mentioned, without a read lock, anonymous checkout risk getting slightly
inconsistent data, but IIRC the possible inconsistency is very minor (and
transient: fixed by the next "cvs update").

The CVS-level (i.e. module-level) metadata is basically inexistent, so the
main issue is the RCS-level (i.e. file-level) metadata: this is updated
pretty much atomically (by writing the new RCS data into a new file and
then moving it into place).

Many CVS repositories have used such unlocked anonymous access with
good results.


        Stefan

Re: Would nobody please finish what he's doing.

[Alan Mackenzie]

Hallo again, Emacs!

On Tue, Jun 19, 2007 at 01:19:08PM +0100, Alan Mackenzie wrote:
> On trying to commit this ChangeLog, I get this error message:

> cvs commit: [10:40:08] waiting for nobody's lock in /cvsroot/emacs/emacs/lisp

> , which is repeated every thirty seconds to indefinite times.  This part
> of Savannah has been in this state for over twelve hours.  I've submitted
> a bug report to savannah.

> Can anybody shed light on nobody, and on what's going on here?

Sylvain Beucler has fixed it, many thanks to him!  They'd moved ViewVC to
a new URL and hadn't updated a config file for it.

-- 
Alan.