wrong default for backup-by-copying-when-linked?

wrong default for backup-by-copying-when-linked?

[Dave Love]

This can't be a new issue, but I don't recall discussion of it.  It
seems to me that backup-by-copying-when-linked should default to t, as
I actually thought it did.  I've just been bitten by breaking links to
some system configuration files.

Is there a good reason for the current default of nil?  If so, I think
it should at least be commented.

Re: wrong default for backup-by-copying-when-linked?

[Andreas Schwab]

Dave Love <d.love@dl.ac.uk> writes:

|> This can't be a new issue, but I don't recall discussion of it.  It
|> seems to me that backup-by-copying-when-linked should default to t, as
|> I actually thought it did.  I've just been bitten by breaking links to
|> some system configuration files.
|> 
|> Is there a good reason for the current default of nil?  If so, I think
|> it should at least be commented.

The current default is convenient if you work on a "cp -al" copy of a
source tree and want to do a fast diff between the original tree and the
modified tree.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux AG, Deutschherrnstr. 15-19, D-90429 Nürnberg
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: wrong default for backup-by-copying-when-linked?

[Dave Love]

Andreas Schwab <schwab@suse.de> writes:

> The current default is convenient if you work on a "cp -al" copy of a
> source tree and want to do a fast diff between the original tree and the
> modified tree.

I'm sure it's more convenient in some cases, but I don't think that
outweighs the trouble it can cause.  When it's clobbered links you
normally have no indication, and it's likely to lead to obscure
problems.  It took quite a while to discover what had happened in the
case that prompted this.

Re: wrong default for backup-by-copying-when-linked?

[Kevin Rodgers]

Dave Love wrote:

> Andreas Schwab <schwab@suse.de> writes:
> 
> 
>>The current default is convenient if you work on a "cp -al" copy of a
>>source tree and want to do a fast diff between the original tree and the
>>modified tree.
>>
> 
> I'm sure it's more convenient in some cases, but I don't think that
> outweighs the trouble it can cause.  When it's clobbered links you
> normally have no indication, and it's likely to lead to obscure
> problems.  It took quite a while to discover what had happened in the
> case that prompted this.

Indeed.  The very first thing I put in every Emacs installation's default.el
file is

(setq backup-by-copying-when-linked t)	; preserve hard links to original file

-- 
<a href="mailto:&lt;kevin.rodgers&#64;ihs.com&gt;">Kevin Rodgers</a>