bug#9959: 24.0.90; `count-words' command from a narrowed buffer

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Dani Moncayo]

Hi,

Try the `count-words' command from a narrowed buffer (without active region).

The message displayed in the echo area is "Buffer has X lines, Y
words, and Z characters.", where the values X, Y and Z are actually
computed for the narrowed portion (not for the entire buffer).

So, this is inaccurate.  The message should begin with something like
"Buffer (narrowed part) has ...".


In GNU Emacs 24.0.90.1 (i386-mingw-nt6.1.7601)
 of 2011-10-27 on DANI-PC
Windowing system distributor `Microsoft Corp.', version 6.1.7601
configured using `configure --with-gcc (4.5)'

-- 
Dani Moncayo

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Juri Linkov]

> Try the `count-words' command from a narrowed buffer (without active region).
>
> The message displayed in the echo area is "Buffer has X lines, Y
> words, and Z characters.", where the values X, Y and Z are actually
> computed for the narrowed portion (not for the entire buffer).
>
> So, this is inaccurate.  The message should begin with something like
> "Buffer (narrowed part) has ...".

Like `C-x =' displays information about a narrowed buffer
in angle brackets `<...>', `count-words' could do the same
and display both widened and narrowed counts.

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Drew Adams]

> > The message should begin with something like
> > "Buffer (narrowed part) has ...".
> 
> Like `C-x =' displays information about a narrowed buffer
> in angle brackets `<...>', `count-words' could do the same
> and display both widened and narrowed counts.

FWIW, in Emacs generally, "buffer (narrowed part)" is referred to as the buffer
"restriction".

But I'm not sure that users unfamiliar with the term would understand it
immediately in a prompt.  Same potential problem with an unexplained convention
such as `<...>'.  I don't have a better suggestion, however.  So far, I'd lean
toward, "Buffer restriction has 112 lines, 673 words, 6074 chars"

Or maybe both counts could be shown, making it a bit clearer what the info is?
"Buffer (restriction): 1135 (112) lines, 6894 (673) words, 70452 (6074) chars".
And in that case, maybe "Buffer (narrowed)..." would be better.  Another
possibility (shorter):

 Narrowed/total: 112/1135 lines, 673/6894 words, 6074/70452 chars

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Juri Linkov]

>  Narrowed/total: 112/1135 lines, 673/6894 words, 6074/70452 chars

This format is nice because it's compact.  Perhaps information about the
region should be added too when the region is active:

Region/narrowed/total: 5/112/1135 lines, 8/673/6894 words, 45/6074/70452 chars

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Drew Adams]

> >  Narrowed/total: 112/1135 lines, 673/6894 words, 6074/70452 chars
> 
> This format is nice because it's compact.  Perhaps 
> information about the region should be added too when
> the region is active:
> 
> Region/narrowed/total: 5/112/1135 lines, 8/673/6894 words, 
> 45/6074/70452 chars

I assume you mean only in the case where the buffer is narrowed and you have an
active region within the restriction.

That would be OK.  But keep in mind that your example uses a small region and
restriction.  In general, the message could be a bit long.  With the same buffer
size, the message could have up to an additional 12 chars.

(The length wouldn't be a problem for me, because my minibuffer frame is as wide
as my screen, and I don't mind if a message wraps.  Just mentioning the
possibility.)

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Kevin Rodgers]

On 11/6/11 8:34 AM, Drew Adams wrote:

> Or maybe both counts could be shown, making it a bit clearer what the info is?
> "Buffer (restriction): 1135 (112) lines, 6894 (673) words, 70452 (6074) chars".
> And in that case, maybe "Buffer (narrowed)..." would be better.  Another
> possibility (shorter):
>
>   Narrowed/total: 112/1135 lines, 673/6894 words, 6074/70452 chars

If I've narrowed a very large buffer to a small region, I am not interested in
the counts for the entire buffer nor willing to wait for them.

-- 
Kevin Rodgers
Denver, Colorado, USA

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Drew Adams]

> If I've narrowed a very large buffer to a small region, I am 
> not interested in the counts for the entire buffer nor
> willing to wait for them.

Yes, that's the cost of such combining.

Possible solutions include (a) adding separate commands for individual pieces of
the combination and (b) finding a way (e.g. prefix arg) to specify interactively
to this command that you want only one or two pieces of info, not everything.

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Drew Adams]

> > If I've narrowed a very large buffer to a small region, I am 
> > not interested in the counts for the entire buffer nor
> > willing to wait for them.
> 
> Yes, that's the cost of such combining.
> 
> Possible solutions include (a) adding separate commands for 
> individual pieces of
> the combination and (b) finding a way (e.g. prefix arg) to 
> specify interactively
> to this command that you want only one or two pieces of info, 
> not everything.

Another possibility: Do as we do with `apropos' commands.  We have both a prefix
arg and a user option, `apropos-do-all', that affect the behavior, with the less
complete, and faster, behavior as the default.

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Juri Linkov]

> Another possibility: Do as we do with `apropos' commands.  We have both a prefix
> arg and a user option, `apropos-do-all', that affect the behavior, with the less
> complete, and faster, behavior as the default.

Or do as we do with `what-cursor-position' (`C-x =') that prints info in
the echo area, but with prefix argument shows full details in *Help* buffer.

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Drew Adams]

> > Another possibility: Do as we do with `apropos' commands.  
> > We have both a prefix arg and a user option,
> > `apropos-do-all', that affect the behavior, with the less
> > complete, and faster, behavior as the default.
> 
> Or

It's not an "or" - the two are orthogonal (except for key binding
choice/conflict).  I was talking about how much info is shown.  You're talking
about where the info is displayed.

The `apropos' approach is good because it lets users define a personal
preference that might be for seeing more info (=t), yet it also lets users
change the behavior on the fly using a prefix arg.

> do as we do with `what-cursor-position' (`C-x =') that 
> prints info in the echo area, but with prefix argument shows
> full details in *Help* buffer.

I don't see the output of even the full info as being something voluminous,
necessitating *Help*.  But perhaps you had something more in mind.  The info
we've discussed so far should be OK for the minibuffer.

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Chong Yidong]

Kevin Rodgers <kevin.d.rodgers@gmail.com> writes:

> On 11/6/11 8:34 AM, Drew Adams wrote:
>
>> Or maybe both counts could be shown, making it a bit clearer what the info is?
>> "Buffer (restriction): 1135 (112) lines, 6894 (673) words, 70452
>> (6074) chars".
>> And in that case, maybe "Buffer (narrowed)..." would be better.  Another
>> possibility (shorter):
>>
>>   Narrowed/total: 112/1135 lines, 673/6894 words, 6074/70452 chars
>
> If I've narrowed a very large buffer to a small region, I am not interested in
> the counts for the entire buffer nor willing to wait for them.

Agreed.  I committed the simplest version, i.e. reporting "Narrowed part
of buffer" in the message.  I'll leave this open in case as a wishlist
item; if anyone wants to implement the more complicated reporting with
C-u, feel free.

bug#9959: 24.0.90; `count-words' command from a narrowed buffer

[Lars Ingebrigtsen]

Dani Moncayo <dmoncayo@gmail.com> writes:

> Try the `count-words' command from a narrowed buffer (without active region).
>
> The message displayed in the echo area is "Buffer has X lines, Y
> words, and Z characters.", where the values X, Y and Z are actually
> computed for the narrowed portion (not for the entire buffer).
>
> So, this is inaccurate.  The message should begin with something like
> "Buffer (narrowed part) has ...".

(I'm going through old bug reports that unfortunately weren't resolved
at the time.)

Chong Yidong fixed this at the time.  The discussion then went on to
whether `C-u M-x count-words' should report data for both the
narrowed-to and total buffer.  That sounds reasonable to me, so I've now
done that in Emacs 29 (although the message is kinda wide).

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no