Re: Broken operating systems....

Friday, 16 June 2000

        On 13 Jun 2000, Jan Vroonhof <vroonhof(a)math.ethz.ch&gt; wrote:

Sorry for delayed reply; this got missed in a busy inbox.

...
 Daniel Pittman <daniel(a)danann.net&gt; writes:

 Does FSF Emacs have the same problem. I don't think our saving
 routines are very much different? 
I didn't check. I am fairly sure that they would, however, be equally
broken by this change.

...
> In the short term[1], is it worth my while putting together a
patch
> to fix this in XEmacs by resolving the output filename and only ever
> passing the real name to the system?

 It seems sensible to access a file using one consistent name. However
 if it is short term only, does playing with back-up-by-copying and
 friends help? 
Setting 'backup-by-copying-when-linked' should fix the issue: the
problem only occurs when using O_CREAT through a symlink to a
non-existent file.

That would serve as a work around for the issue. I am still, personally,
inclined to count it as a nasty bug in the kernel though.

...
> I had a glance at the code and this is, um, non-trivial to do, I
> think. Especially as it would mean potentially dealing with GC. I
> will put in the effort if needed, though.

 If you want you can try writing the code first without the GC-stuff. I
 don't think the file handling code does any GC-ing as long as
 file-name-handler-alist 
Hmmmm. Well, with that assurance in mind, I feel less worried about it.
I might have a shot at changing the code, assuming that I can find any
free time over the next few days...

Thanks,
        Daniel

-- 
It is well to remember that the entire universe, with one trifling exception,
is composed of others.
        -- John Andrew Holmes

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998