dired-do-hardlink should update link count we see

dired-do-hardlink should update link count we see

[Dan Jacobson]

I can't believe the dired-do-hardlink command still doesn't update the
link count for at least the file we are on... like the M command
updates the modes.  Well, at least it could say in the minibuffer
"press g to refresh..."
-- 
http://jidanni.org/ Taiwan(04)25854780

Re: dired-do-hardlink should update link count we see

[Eli Zaretskii]

On 8 Jul 2002, Dan Jacobson wrote:

> I can't believe the dired-do-hardlink command still doesn't update the
> link count for at least the file we are on...

I think Dired doesn't automatically refresh the directory view as a 
matter of principle.  It's not something specific to dired-do-hardlink.

Re: dired-do-hardlink should update link count we see

[David Kastrup]

eliz@is.elta.co.il (Eli Zaretskii) writes:

> On 8 Jul 2002, Dan Jacobson wrote:
> 
> > I can't believe the dired-do-hardlink command still doesn't update the
> > link count for at least the file we are on...
> 
> I think Dired doesn't automatically refresh the directory view as a
> matter of principle.  It's not something specific to
> dired-do-hardlink.

So if you rename a file from within dired, dired displays the same
file name?  If you change its permissions from within dired, it
displays the same permissions?  If you move it or delete it from
within dired, it stays there as if it existed before?  If you copy to
some other directory under dired, the files don't arrive in the
display there?

Could you substantiate your claim?  I can't seem to find any basic
dired operation that does _not_ update the view (except that when you
are renaming files in a sorted buffer, the buffer does not get
resorted.  But the name is still changed in the display).

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum
Email: David.Kastrup@t-online.de

Re: dired-do-hardlink should update link count we see

[Eli Zaretskii]

On 9 Jul 2002, David Kastrup wrote:

> > I think Dired doesn't automatically refresh the directory view as a
> > matter of principle.  It's not something specific to
> > dired-do-hardlink.
> 
> So if you rename a file from within dired, dired displays the same
> file name?  If you change its permissions from within dired, it
> displays the same permissions?  If you move it or delete it from
> within dired, it stays there as if it existed before?  If you copy to
> some other directory under dired, the files don't arrive in the
> display there?
> 
> Could you substantiate your claim?

The claim is that Dired never queries the filesystem after any operation, 
it just manipulates the buffer contents.

Re: dired-do-hardlink should update link count we see

[David Kastrup]

eliz@is.elta.co.il (Eli Zaretskii) writes:

> On 9 Jul 2002, David Kastrup wrote:
> 
> > > I think Dired doesn't automatically refresh the directory view as a
> > > matter of principle.  It's not something specific to
> > > dired-do-hardlink.
> > 
> > So if you rename a file from within dired, dired displays the same
> > file name?  If you change its permissions from within dired, it
> > displays the same permissions?  If you move it or delete it from
> > within dired, it stays there as if it existed before?  If you copy to
> > some other directory under dired, the files don't arrive in the
> > display there?
> > 
> > Could you substantiate your claim?
> 
> The claim is that Dired never queries the filesystem after any operation, 
> it just manipulates the buffer contents.

Well, that's what Dan complained about: that it did not manipulate the
buffer contents after doing a hard link.  So if your claim was to be
what you claim, it was utterly irrelevant to the report.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum
Email: David.Kastrup@t-online.de

Re: dired-do-hardlink should update link count we see

[Eli Zaretskii]

On 9 Jul 2002, David Kastrup wrote:

> So if your claim was to be
> what you claim, it was utterly irrelevant to the report.

I don't think it was irrelevant.

Re: dired-do-hardlink should update link count we see

[David Kastrup]

eliz@is.elta.co.il (Eli Zaretskii) writes:

> On 9 Jul 2002, David Kastrup wrote:
> 
> > So if your claim was to be
> > what you claim, it was utterly irrelevant to the report.
> 
> I don't think it was irrelevant.

It is perfectly understandable if one thinks some things not worth
fixing or doing correctly.  There is something like priorities.
Playing semantic games and going into denial is not necessary if one
does not want to indulge personally into fixing a problem, and it may
keep others from doing so.

Short summary:
Almost all dired operations update the buffer (without actually
rereading the directory) according to their knowledge about the
affected change.  dired-do-hardlink doesn't adjust the link count
which changes predictably.  That's what Dan reported.

You first claimed that dired operations don't update the buffer, then
you claimed that this was supposed to mean dired does not reread the
directory in order to update the buffer.  Since rereading the
directory is not necessary for predicting the changed link count,
this interpretation of your original claim would make it irrelevant.

You beg to differ.  Since you are in effect asking that the current
behavior not be changed for the sake of some perceived consistency,
it might be good to back up your claim before some misguided soul
might be tempted into changing Emacs' behavior in that regard.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum
Email: David.Kastrup@t-online.de

Re: dired-do-hardlink should update link count we see

[Eli Zaretskii]

On 9 Jul 2002, David Kastrup wrote:

> It is perfectly understandable if one thinks some things not worth
> fixing or doing correctly.

I didn't say this isn't worth fixing, I just pointed out that it's a part 
of a general pattern.

Re: dired-do-hardlink should update link count we see

[David Kastrup]

eliz@is.elta.co.il (Eli Zaretskii) writes:

> On 9 Jul 2002, David Kastrup wrote:
> 
> > It is perfectly understandable if one thinks some things not worth
> > fixing or doing correctly.
> 
> I didn't say this isn't worth fixing, I just pointed out that it's a
> part of a general pattern.

Oh, surely, surely.  Other operations (like changing permissions, the
file name, moving files, copying them) update the info in the dired
buffer without actually rereading the directory info, so the general
pattern would certainly imply that updating the info in the dired
buffer without actually rereading the directory info would be
something entirely different.

Again, if you are planning to make sense, you better try harder.  And
just removing all relevant context from posts of mine (which you have
now done two times already) in order to repeat your unsubstantiated
claim is starting to look silly.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum
Email: David.Kastrup@t-online.de

Re: dired-do-hardlink should update link count we see

[Dan Jacobson]

D> Again, if you are planning to make sense, you better try harder.  And
D> just removing all relevant context from posts of mine (which you have
D> now done two times already) in order to repeat your unsubstantiated
D> claim is starting to look silly.

Yes, I have been noticing loyalty problems with Eli for quite some
time now.

It seems he may not grasp fully the goals of the movement.

He might need reeducation.

We will have an assistant talk with him. 
-- 
http://jidanni.org/ Taiwan(04)25854780

Re: dired-do-hardlink should update link count we see

[David Kastrup]

jidanni@dman.ddts.net (Dan Jacobson) writes:

> D> Again, if you are planning to make sense, you better try harder.  And
> D> just removing all relevant context from posts of mine (which you have
> D> now done two times already) in order to repeat your unsubstantiated
> D> claim is starting to look silly.
> 
> Yes, I have been noticing loyalty problems with Eli for quite some
> time now.
> 
> It seems he may not grasp fully the goals of the movement.
> 
> He might need reeducation.
> 
> We will have an assistant talk with him. 

I take this to be a criticism of the way I carried on.  It has been
pointed out to me already that there was no reason for me to get as
nasty as I did.  It would have been quite sufficient to let the
readers judge the opposite opinions expressed in our original posts
without further quibbling.

Sorry for that.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum
Email: David.Kastrup@t-online.de

Re: dired-do-hardlink should update link count we see

[Richard Stallman]

    The claim is that Dired never queries the filesystem after any operation, 
    it just manipulates the buffer contents.

Some dired operations relist a specific file to update the line
based on the current facts.  See dired-relist-file.
I will make dired-do-hardlink do that too.