VC-Dired now lists unregistered files, marked with "?"...

VC-Dired now lists unregistered files, marked with "?"...

[Eric S. Raymond]

...in VC-Dired's non-terse mode.  It's an interesting UI question 
whether they should be listed in terse mode as well.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>

The real point of audits is to instill fear, not to extract revenue;
the IRS aims at winning through intimidation and (thereby) getting
maximum voluntary compliance
	-- Paul Strassel, former IRS Headquarters Agent Wall St. Journal 1980

Re: VC-Dired now lists unregistered files, marked with "?"...

[Dan Nicolaescu]

"Eric S. Raymond" <esr@snark.thyrsus.com> writes:

  > ...in VC-Dired's non-terse mode.  

Great, thank you!

  > It's an interesting UI question whether they should be listed in
  > terse mode as well.

IMHO they should. The unregistered files are always "interesting" for a
user because he might want to register them. The non-terse mode shows
too many files...

BTW, for consistency with PCL-CVS and with the vc-dired display for
other states shouldn't "(Unknown)" be better than just "?" ?

Re: VC-Dired now lists unregistered files, marked with "?"...

[Eric S. Raymond]

Dan Nicolaescu <dann@ics.uci.edu>:
> IMHO they should. The unregistered files are always "interesting" for a
> user because he might want to register them. The non-terse mode shows
> too many files...

On the other hand...I'm working on a survey paper on VCSes right now,
The directory includes a small handful of registered files, and several
subdirectories that are VCS source trees I'm keeping around for reference.
Displaying all of the several thousand unregistered files in those trees
in terse mode just swamps the real information (I tried it that way).

> BTW, for consistency with PCL-CVS and with the vc-dired display for
> other states shouldn't "(Unknown)" be better than just "?" ?

I considered exactly that, but "(unknown)" would actually be wrong in
this instance; the VC state of the file is known.  Alas, "(unregistered)"
doesn't fit in the status field.

See my next post on changing VC's status display.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>

Re: VC-Dired now lists unregistered files, marked with "?"...

[Thien-Thi Nguyen]

() "Eric S. Raymond" <esr@thyrsus.com>
() Fri, 28 Dec 2007 05:26:34 -0500

   Alas, "(unregistered)" doesn't fit in the status field.

i use "alas-no" for that state:
http://www.gnuvola.org/software/personal-elisp/dist/lisp/prog-env/rmangle.el

thi

Re: VC-Dired now lists unregistered files, marked with "?"...

[Dan Nicolaescu]

"Eric S. Raymond" <esr@thyrsus.com> writes:

  > Dan Nicolaescu <dann@ics.uci.edu>:
  > > IMHO they should. The unregistered files are always "interesting" for a
  > > user because he might want to register them. The non-terse mode shows
  > > too many files...
  > 
  > On the other hand...I'm working on a survey paper on VCSes right now,
  > The directory includes a small handful of registered files, and several
  > subdirectories that are VCS source trees I'm keeping around for reference.
  > Displaying all of the several thousand unregistered files in those trees
  > in terse mode just swamps the real information (I tried it that way).

I would think that the above situation is rather unusual, so basing the
default on that is not the best choice.  

IMO users would put files that the VCS should not deal with in
.BLAHignore otherwise using the output VC commands from the command line
is too noisy. So the number of "unknown" files should be rather small.

Re: VC-Dired now lists unregistered files, marked with "?"...

[Richard Stallman]

    On the other hand...I'm working on a survey paper on VCSes right now,
    The directory includes a small handful of registered files, and several
    subdirectories that are VCS source trees I'm keeping around for reference.
    Displaying all of the several thousand unregistered files in those trees
    in terse mode just swamps the real information (I tried it that way).

It sounds like we really want to let the user choose to show or hide
the unregistered files.

Or perhaps a way of sorting that puts the registered files at the top.

Re: VC-Dired now lists unregistered files, marked with "?"...

[Stefan Monnier]

> On the other hand...I'm working on a survey paper on VCSes right now,
> The directory includes a small handful of registered files, and several
> subdirectories that are VCS source trees I'm keeping around for reference.

Then you need to mark those directories as "ignored".


        Stefan

Re: VC-Dired now lists unregistered files, marked with "?"...

[Stefan Monnier]

>     On the other hand...I'm working on a survey paper on VCSes right now,
>     The directory includes a small handful of registered files, and several
>     subdirectories that are VCS source trees I'm keeping around for reference.
>     Displaying all of the several thousand unregistered files in those trees
>     in terse mode just swamps the real information (I tried it that way).

> It sounds like we really want to let the user choose to show or hide
> the unregistered files.

PCL-CVS has never offered such an option and nobody has ever asked
for it.  Just show those files and let the user place them in .fooignore
if they don't want to see them.


        Stefan

Re: VC-Dired now lists unregistered files, marked with "?"...

[Richard Stallman]

    PCL-CVS has never offered such an option and nobody has ever asked
    for it.  Just show those files and let the user place them in .fooignore
    if they don't want to see them.

That sounds rather inconvenient, I must say, if you often look at
new directories.

Re: VC-Dired now lists unregistered files, marked with "?"...

[Eric S. Raymond]

Richard Stallman <rms@gnu.org>:
>     PCL-CVS has never offered such an option and nobody has ever asked
>     for it.  Just show those files and let the user place them in .fooignore
>     if they don't want to see them.
> 
> That sounds rather inconvenient, I must say, if you often look at
> new directories.

I agree.  It's a non-solution.  Better to have three verbosity levels --
terse, terse + unregistered, and everything.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>

Re: VC-Dired now lists unregistered files, marked with "?"...

[Stefan Monnier]

>     PCL-CVS has never offered such an option and nobody has ever asked
>     for it.  Just show those files and let the user place them in .fooignore
>     if they don't want to see them.

> That sounds rather inconvenient, I must say, if you often look at
> new directories.

Huh?  I don't understand the use case you're thinking of.
What new directories?

Most/all VCS will repeatedly warn/complain about un-registered and
non-ignored files, so it's just a matter of normal use to list such files
in the .fooignore file.  This is true for any frontend.


        Stefan

Re: VC-Dired now lists unregistered files, marked with "?"...

[Stefan Monnier]

>> PCL-CVS has never offered such an option and nobody has ever asked
>> for it.  Just show those files and let the user place them in .fooignore
>> if they don't want to see them.
>> 
>> That sounds rather inconvenient, I must say, if you often look at
>> new directories.

> I agree.  It's a non-solution.  Better to have three verbosity levels --
> terse, terse + unregistered, and everything.

PCL-CVS moved away from "levels" and I think it users found it to be
an improvement.  Basically, you just have the default (which is
equivalent to what you call "terse+unregistered") and then you have
commands to affect the current display to:

- place a given file in the .fooignore list (thus bringing you closer
  to just "terse" but this will also work for all other frontends
  rather than only VC and only in terse mode).
- expand a subdir to show all its files and their status
  (i.e. "everything" but with the advantage it can apply to a subdir
  rather than being an all-or-nothing choice (which quickly becomes
  a non-choice when your tree has thousands of files)).
- add a file (with its status) to the display (even if the file doesn't
  exist, so you can "cvs diff -rR1 -rR2 foo" even if "foo" was removed
  between version R2 and the working revision).


-- Stefan