bug#10148: 24.0.91; "pending" regexp Isearch

bug#10148: 24.0.91; "pending" regexp Isearch

[Dani Moncayo]

Hi,

from "emacs -Q", in the *scratch* buffer, type `C-< C-M-s i s *'.

At that point, the Isearch prompt is "Pending regexp I-search: is*".
What does "Pending" mean in this context?  I've been unable to find
out the answer in the manual, and I can't figure it out either (the
search seems pretty complete to me).

So please, add an explanation of this "pending" thing to the manual.

TIA.


In GNU Emacs 24.0.91.1 (i386-mingw-nt6.1.7601)
 of 2011-11-25 on DANI-PC
Windowing system distributor `Microsoft Corp.', version 6.1.7601
configured using `configure --with-gcc (4.6) --cflags -fno-omit-frame-pointer'

-- 
Dani Moncayo

bug#10148: 24.0.91; "pending" regexp Isearch

[Juri Linkov]

> from "emacs -Q", in the *scratch* buffer, type `C-< C-M-s i s *'.
>
> At that point, the Isearch prompt is "Pending regexp I-search: is*".
> What does "Pending" mean in this context?  I've been unable to find
> out the answer in the manual, and I can't figure it out either (the
> search seems pretty complete to me).

It helps to understand what does this mean after putting a single letter "i"
in front of the word "is", e.g. by putting this first line
in the *scratch* buffer:

;; i This buffer is for notes you don't want to save, and for Lisp evaluation.

and typing the same key sequence `C-< C-M-s i s *'.

"Pending" is necessary for Isearch to go back to the first shorter match.

bug#10148: 24.0.91; "pending" regexp Isearch

[Dani Moncayo]

> It helps to understand what does this mean after putting a single letter "i"
> in front of the word "is", e.g. by putting this first line
> in the *scratch* buffer:
>
> ;; i This buffer is for notes you don't want to save, and for Lisp evaluation.
>
> and typing the same key sequence `C-< C-M-s i s *'.
>
> "Pending" is necessary for Isearch to go back to the first shorter match.

So, correct me if I'm wrong: "pending" here means "this match has been
found by re-starting the search from the original position".

If that is so, I don't understand the "pending" adjective here.
According to [1], "pending" = "about to happen or waiting to happen",
which is misleading here, no?.

What about changing "pending" for something like "restarted"?

In any case, please describe this flag in chapter 15 of the manual.

--- Footnotes ---

[1] http://dictionary.cambridge.org/dictionary/british/pending_1?q=pending

--
Dani Moncayo

bug#10148: 24.0.91; "pending" regexp Isearch

[Juri Linkov]

> If that is so, I don't understand the "pending" adjective here.
> According to [1], "pending" = "about to happen or waiting to happen",
> which is misleading here, no?.
>
> What about changing "pending" for something like "restarted"?

Actually in isearch.el, the name of the variable that corresponds
to the prompt "pending ", is `isearch-adjusted'.

It seems the prompt string "adjusted" is more appropriate here
in this case.

bug#10148: 24.0.91; "pending" regexp Isearch

[Dani Moncayo]

>> If that is so, I don't understand the "pending" adjective here.
>> According to [1], "pending" = "about to happen or waiting to happen",
>> which is misleading here, no?.
>>
>> What about changing "pending" for something like "restarted"?
>
> Actually in isearch.el, the name of the variable that corresponds
> to the prompt "pending ", is `isearch-adjusted'.
>
> It seems the prompt string "adjusted" is more appropriate here
> in this case.

From my limited knowledge of the English language, "adjusted" is not
very helpful for the user to figure out what's going on: that the
search has been restarted from the original position.


-- 
Dani Moncayo

bug#10148: 24.0.91; "pending" regexp Isearch

[Juri Linkov]

> From my limited knowledge of the English language, "adjusted" is not
> very helpful for the user to figure out what's going on: that the
> search has been restarted from the original position.

The search is not restarted - its stack is not empty, i.e. DEL still
can discard last input, etc.  In this case, the search just adjusts
its position to another match.

bug#10148: 24.0.91; "pending" regexp Isearch

[Dani Moncayo]

On Wed, Nov 30, 2011 at 11:38, Juri Linkov <juri@jurta.org> wrote:
>> From my limited knowledge of the English language, "adjusted" is not
>> very helpful for the user to figure out what's going on: that the
>> search has been restarted from the original position.
>
> The search is not restarted - its stack is not empty, i.e. DEL still
> can discard last input, etc.  In this case, the search just adjusts
> its position to another match.

You're right, so I take back my suggestion of "restarted".

I don't have a better suggestion right now, but IMO, "adjusted" is too
general.  It conveys little information.

-- 
Dani Moncayo

bug#10148: 24.0.91; "pending" regexp Isearch

[Dani Moncayo]

On Wed, Nov 30, 2011 at 11:55, Dani Moncayo <dmoncayo@gmail.com> wrote:
> On Wed, Nov 30, 2011 at 11:38, Juri Linkov <juri@jurta.org> wrote:
>>> From my limited knowledge of the English language, "adjusted" is not
>>> very helpful for the user to figure out what's going on: that the
>>> search has been restarted from the original position.
>>
>> The search is not restarted - its stack is not empty, i.e. DEL still
>> can discard last input, etc.  In this case, the search just adjusts
>> its position to another match.
>
> You're right, so I take back my suggestion of "restarted".
>
> I don't have a better suggestion right now, but IMO, "adjusted" is too
> general.  It conveys little information.

Well, on second thought, "adjusted" doesn't sound too bad :-).

So I'm OK with it, but please explain also its meaning in the Emacs manual.

Thank you.

-- 
Dani Moncayo

bug#10148: 24.0.91; "pending" regexp Isearch

[Stefan Monnier]

> and typing the same key sequence `C-< C-M-s i s *'.
> "Pending" is necessary for Isearch to go back to the first shorter match.

I don't see why it would be "necessary".

> Well, on second thought, "adjusted" doesn't sound too bad :-).

Another option is to not say anything at all.  I don't have a strong
opinion on this, but I never found "pending" useful and I don't think
I'd have felt differently if it had said "adjusted" instead.


        Stefan

bug#10148: 24.0.91; "pending" regexp Isearch

[Dani Moncayo]

On Wed, Nov 30, 2011 at 15:08, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
>> and typing the same key sequence `C-< C-M-s i s *'.
>> "Pending" is necessary for Isearch to go back to the first shorter match.
>
> I don't see why it would be "necessary".
>
>> Well, on second thought, "adjusted" doesn't sound too bad :-).
>
> Another option is to not say anything at all.  I don't have a strong
> opinion on this, but I never found "pending" useful and I don't think
> I'd have felt differently if it had said "adjusted" instead.

+1

-- 
Dani Moncayo

bug#10148: 24.0.91; "pending" regexp Isearch

[Drew Adams]

> Another option is to not say anything at all.  I don't have a strong
> opinion on this, but I never found "pending" useful and I don't think
> I'd have felt differently if it had said "adjusted" instead.

+1

Or use some text that says what it means?  You say it moves to another search
hit ("adjusts its position to another match").

Is that other search hit always the next/previous one (depending on direction)?
Or is it at least some simple, consistent movement that is easy to name?

If so, maybe just say that you are moving to the "[next]" match or some such.
Dunno - it's not clear to me what is involved.  But it's also not clear to me
from either "pending" or "adjusted".  And if you say nothing at all but move to
another search hit, that too might be confusing.  HTH.

bug#10148: 24.0.91; "pending" regexp Isearch

[Juri Linkov]

>> and typing the same key sequence `C-< C-M-s i s *'.
>> "Pending" is necessary for Isearch to go back to the first shorter match.
>
> I don't see why it would be "necessary".

`isearch-fallback' [ab]uses `isearch-adjusted' to be able to
move point to `isearch-barrier' because when `isearch-adjusted' is t,
then `isearch-search-and-update' doesn't move point to another position
not intended by `isearch-fallback'.  So this is just a technical detail
of implementation that doesn't need a message.

>> Well, on second thought, "adjusted" doesn't sound too bad :-).
>
> Another option is to not say anything at all.  I don't have a strong
> opinion on this, but I never found "pending" useful and I don't think
> I'd have felt differently if it had said "adjusted" instead.

The message "pending" is not necessary for `isearch-fallback',
but it helps when using `isearch-toggle-regexp', `isearch-toggle-word'
and `isearch-toggle-case-fold' because they doesn't repeat the search
immediately and require an action from the user to find the next match
(like another `C-s').

bug#10148: 24.0.91; "pending" regexp Isearch

[Stefan Monnier]

> The message "pending" is not necessary for `isearch-fallback',
> but it helps when using `isearch-toggle-regexp', `isearch-toggle-word'
> and `isearch-toggle-case-fold' because they doesn't repeat the search
> immediately and require an action from the user to find the next match
> (like another `C-s').

In this context "pending" makes sense, yes.  So the problem is that
"pending" is also used for the other case where it does not make sense.


        Stefan

bug#10148: 24.0.91; "pending" regexp Isearch

[Juri Linkov]

>> The message "pending" is not necessary for `isearch-fallback',
>> but it helps when using `isearch-toggle-regexp', `isearch-toggle-word'
>> and `isearch-toggle-case-fold' because they doesn't repeat the search
>> immediately and require an action from the user to find the next match
>> (like another `C-s').
>
> In this context "pending" makes sense, yes.  So the problem is that
> "pending" is also used for the other case where it does not make sense.

This is still doubtful because when using an incomplete regexp like
`C-M-s i s \ |', "pending" still makes sense since Isearch is waiting
for the user to type an alternative search string.

Actually this situation is already described in (info "(emacs) Regexp Search"):

     In some cases, adding characters to the regexp in an incremental
  regexp search can make the cursor move back and start again.  For
  example, if you have searched for `foo' and you add `\|bar', the cursor
  backs up in case the first `bar' precedes the first `foo'.

So it seems that Dani just wants to add a phrase like

  and `Pending' appears in the search prompt

to the end of the first sentence cited above in the Info manual.

bug#10148: 24.0.91; "pending" regexp Isearch

[Eli Zaretskii]

> From: Juri Linkov <juri@jurta.org>
> Date: Thu, 01 Dec 2011 09:37:34 +0200
> Cc: 10148@debbugs.gnu.org
> 
> [...] when using an incomplete regexp like
> `C-M-s i s \ |', "pending" still makes sense since Isearch is waiting
> for the user to type an alternative search string.

AFAIK, "pending" doesn't mean "I'm waiting for more data".  It means
"data already arrived, and is waiting to be processed".

If we want to say "waiting for more data", let's just say that.

(Apologies if I said something stupid, because I didn't track this
thread.)

bug#10148: 24.0.91; "pending" regexp Isearch

[Drew Adams]

>      In some cases, adding characters to the regexp in an incremental
>   regexp search can make the cursor move back and start again.  For
>   example, if you have searched for `foo' and you add 
> `\|bar', the cursor
>   backs up in case the first `bar' precedes the first `foo'.

Question: How does this relate to the Isearch error messages that indicate
invalid and incomplete regexp patterns, and are thus more specific:
"\\`Premature \\|\\`Unmatched \\|\\`Invalid "?

bug#10148: 24.0.91; "pending" regexp Isearch

[Stefan Monnier]

>> In this context "pending" makes sense, yes.  So the problem is that
>> "pending" is also used for the other case where it does not make sense.
> This is still doubtful because when using an incomplete regexp like
> `C-M-s i s \ |', "pending" still makes sense since Isearch is waiting
> for the user to type an alternative search string.

No, it's not: it has already moved point to the start of the search
where "is\\|" indeed is a valid match.

It's very different from the "toggle case sensitivity" where point is
not updated to reflect the new search criterion yet, hence it is "pending".


        Stefan

bug#10148: 24.0.91; "pending" regexp Isearch

[Lars Ingebrigtsen]

Juri Linkov <juri@jurta.org> writes:

>> In this context "pending" makes sense, yes.  So the problem is that
>> "pending" is also used for the other case where it does not make sense.
>
> This is still doubtful because when using an incomplete regexp like
> `C-M-s i s \ |', "pending" still makes sense since Isearch is waiting
> for the user to type an alternative search string.
>
> Actually this situation is already described in (info "(emacs) Regexp Search"):
>
>      In some cases, adding characters to the regexp in an incremental
>   regexp search can make the cursor move back and start again.  For
>   example, if you have searched for `foo' and you add `\|bar', the cursor
>   backs up in case the first `bar' precedes the first `foo'.
>
> So it seems that Dani just wants to add a phrase like
>
>   and `Pending' appears in the search prompt
>
> to the end of the first sentence cited above in the Info manual.

This was ten years ago, and "pending" is still somewhat confusing, but
I've now added the text to the manual in Emacs 29.

I agree that "pending" isn't very informative here, but I don't really
have a suggestion for a clearer wording.  In the "is*" case, we want to
say "we've recalculated (and possibly added more) matches", which is
unusual when isearching, so it's probably nice that we get a message for
that.  But something like "Recalculated" instead of "Pending" might be
nice?

In the \| case, we're probably waiting for more data, so "Pending" is
fine there, but I don't think we really can separate between the two
cases, can we?  "foo\|" is a valid regexp; it's just very odd.

So I dunno.  Perhaps just leaving it as is is the best solution.  Any
opinions?

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

bug#10148: 24.0.91; "pending" regexp Isearch

[Drew Adams]

Could someone please answer the question
I posed in message #50?

  Question: How does this relate to the Isearch
  error messages that indicate invalid and
  incomplete regexp patterns, and are thus more
  specific: "\\`Premature \\|\\`Unmatched \\|\\`Invalid "?

https://debbugs.gnu.org/cgi/bugreport.cgi?bug=10148#50

bug#10148: 24.0.91; "pending" regexp Isearch

[Juri Linkov]

> Could someone please answer the question
> I posed in message #50?
>
>   Question: How does this relate to the Isearch
>   error messages that indicate invalid and
>   incomplete regexp patterns, and are thus more
>   specific: "\\`Premature \\|\\`Unmatched \\|\\`Invalid "?
>
> https://debbugs.gnu.org/cgi/bugreport.cgi?bug=10148#50

This is unrelated to these error messages.

bug#10148: 24.0.91; "pending" regexp Isearch

[Drew Adams]

> > Could someone please answer the question
> > I posed in message #50?
> This is unrelated to these error messages.

Thanks.

bug#10148: 24.0.91; "pending" regexp Isearch

[Juri Linkov]

> This was ten years ago, and "pending" is still somewhat confusing

Do you really see that "pending"? ;-)  We removed it recently in bug#52356.

bug#10148: 24.0.91; "pending" regexp Isearch

[Lars Ingebrigtsen]

Juri Linkov <juri@jurta.org> writes:

>> This was ten years ago, and "pending" is still somewhat confusing
>
> Do you really see that "pending"? ;-)  We removed it recently in bug#52356.

The test case on the current trunk still says "Pending" -- emacs -Q, C-u
C-s i s * -> pending.

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

bug#10148: 24.0.91; "pending" regexp Isearch

[Juri Linkov]

[-- Attachment #1: Type: text/plain, Size: 384 bytes --]

>>> This was ten years ago, and "pending" is still somewhat confusing
>>
>> Do you really see that "pending"? ;-)  We removed it recently in bug#52356.
>
> The test case on the current trunk still says "Pending" -- emacs -Q, C-u
> C-s i s * -> pending.

If no one wants "Pending" anywhere, then this patch removes its
remaining occurrences (bug#52356 removed other two occurrences):


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: isearch-fallback.patch --]
[-- Type: text/x-diff, Size: 848 bytes --]

diff --git a/lisp/isearch.el b/lisp/isearch.el
index be0227b6e7..9f848dd229 100644
--- a/lisp/isearch.el
+++ b/lisp/isearch.el
@@ -2858,7 +2911,7 @@ isearch-fallback
 		 allow-invalid))
     (if to-barrier
 	(progn (goto-char isearch-barrier)
-	       (setq isearch-adjusted t))
+	       (setq isearch-adjusted 'fallback))
       (let* ((stack isearch-cmds)
 	     (previous (cdr stack))	; lookbelow in the stack
 	     (frame (car stack)))
@@ -2893,7 +2946,7 @@ isearch-fallback
 	    (goto-char (if isearch-forward
 			   (max last-other-end isearch-barrier)
 			 (min last-other-end isearch-barrier)))
-	    (setq isearch-adjusted t)))))))
+	    (setq isearch-adjusted 'fallback)))))))
 
 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 ;; scrolling within Isearch mode.  Alan Mackenzie (acm@muc.de), 2003/2/24

bug#10148: 24.0.91; "pending" regexp Isearch

[Lars Ingebrigtsen]

Juri Linkov <juri@jurta.org> writes:

> If no one wants "Pending" anywhere, then this patch removes its
> remaining occurrences (bug#52356 removed other two occurrences):

I haven't tried the patch -- what will the message be after applying it?
(I think it would be nice to say something about having adjusted, but
the word "pending" isn't ideal.)

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

bug#10148: 24.0.91; "pending" regexp Isearch

[Juri Linkov]

>> If no one wants "Pending" anywhere, then this patch removes its
>> remaining occurrences (bug#52356 removed other two occurrences):
>
> I haven't tried the patch -- what will the message be after applying it?
> (I think it would be nice to say something about having adjusted, but
> the word "pending" isn't ideal.)

It will just remove the word "pending" from the message.
We could use different words to show in this regexp case,
and another word after toggling isearch mode.
But I don't see how using another word would address complaints
about polluting the search message with useless words.

bug#10148: 24.0.91; "pending" regexp Isearch

[Lars Ingebrigtsen]

Juri Linkov <juri@jurta.org> writes:

> It will just remove the word "pending" from the message.
> We could use different words to show in this regexp case,
> and another word after toggling isearch mode.
> But I don't see how using another word would address complaints
> about polluting the search message with useless words.

Like I said, I think notifying the user about "something changed" is
useful -- but the word "pending" is just misleading, and I think that's
what everybody complained about.  So changing it to "Recalculated" or
"Restarted" or "Amended" or something along those lines would be an
improvement.

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

bug#10148: 24.0.91; "pending" regexp Isearch

[Stefan Monnier via Bug reports for GNU Emacs, the Swiss army knife of text editors]

>> It will just remove the word "pending" from the message.
>> We could use different words to show in this regexp case,
>> and another word after toggling isearch mode.
>> But I don't see how using another word would address complaints
>> about polluting the search message with useless words.
>
> Like I said, I think notifying the user about "something changed" is
> useful -- but the word "pending" is just misleading, and I think that's
> what everybody complained about.  So changing it to "Recalculated" or
> "Restarted" or "Amended" or something along those lines would be an
> improvement.

IIUC the "pending" is used to indicate that the search state hasn't been
fully updated to reflect the new search pattern.
Maybe the better fix is to fully update the search state?

I'm not sure which part of the search state hasn't been fully
updated, tho.


        Stefan