bug#12986: Pressing backspace during isearch exits case-sensitive and word modes

bug#12986: Pressing backspace during isearch exits case-sensitive and word modes

[Kelly Dean]

Using 24.2, type:
onto to
Then do:
C-a C-s to
It highlights the first occurrence of "to" as the current match, and the second as the next match.
Then do: M-s w
Now, if you type a space, it continues in word mode as expected, but if instead you backspace over the o, it exits word mode. It shouldn't do that.
This same problem applies to case-sensitivity mode (M-c) too.

bug#12986: Pressing backspace during isearch exits case-sensitive and word modes

[Juri Linkov]

> Using 24.2, type:
> onto to
> Then do:
> C-a C-s to
> It highlights the first occurrence of "to" as the current match,
> and the second as the next match.
> Then do: M-s w
> Now, if you type a space, it continues in word mode as expected, but
> if instead you backspace over the o, it exits word mode. It shouldn't do that.
> This same problem applies to case-sensitivity mode (M-c) too.

<backspace> (isearch-delete-char) restores the previous search state.
If you want just to remove the last character from the search string,
you can type `C-M-w' (isearch-del-char).

bug#12986: Pressing backspace during isearch exits case-sensitive and word modes

[Juri Linkov]

[please keep your replies Cc'd to 12986@debbugs.gnu.org]

>> <backspace> (isearch-delete-char) restores the previous search state.
>> If you want just to remove the last character from the search string,
>> you can type `C-M-w' (isearch-del-char).

> Sorry for the mistaken bug report. But it still seems to be a UI
> inconsistency: although the described behavior happens for word mode and
> case-insensitivity mode, it doesn't happen for regex mode, which it what
> led me to believe it was a bug for the first two. If it's supposed to
> happen for the first two, then shouldn't it happen for all three?

Good point.  `isearch-delete-char' (that uses `isearch-pop-state')
restores word mode but not regexp mode.  This looks like
unaccountable inconsistency.  Perhaps it should restore
regexp mode as well.

Case-insensitivity is already restored when you type <backspace>
but you can't see this because it has no indication in the prompt.
Maybe it should momentarily flash a case sensitivity indicator
when you type <backspace> like it does when you type `M-c'
(but only when <backspace> changes the state of case-insensitivity).

The recently added `isearch-lax-whitespace' could be saved/restored too.

For customizability a new user option could be added to define a list
of search states that the user wants to keep on the isearch stack.

bug#12986: Pressing backspace during isearch exits case-sensitive and word modes

[Juri Linkov]

>> But it still seems to be a UI inconsistency: although the described
>> behavior happens for word mode and case-insensitivity mode, it
>> doesn't happen for regex mode, which it what led me to believe it was
>> a bug for the first two. If it's supposed to happen for the first two,
>> then shouldn't it happen for all three?
>
> Good point.  `isearch-delete-char' (that uses `isearch-pop-state')
> restores word mode but not regexp mode.  This looks like
> unaccountable inconsistency.  Perhaps it should restore
> regexp mode as well.
>
> Case-insensitivity is already restored when you type <backspace>
> but you can't see this because it has no indication in the prompt.
> Maybe it should momentarily flash a case sensitivity indicator
> when you type <backspace> like it does when you type `M-c'
> (but only when <backspace> changes the state of case-insensitivity).
>
> The recently added `isearch-lax-whitespace' could be saved/restored too.
>
> For customizability a new user option could be added to define a list
> of search states that the user wants to keep on the isearch stack.

This patch adds a new user option `isearch-keep-stack-variables'
intended to accompany another option `isearch-keep-mode-variables'
added in bug#11378.  It adds a new option but doesn't change
the default behavior - its default value is `isearch-regexp'
that means to not keep regexp mode when pressing backspace
(other search variables will be supported later as well).

=== modified file 'lisp/isearch.el'
--- lisp/isearch.el	2013-05-30 23:50:36 +0000
+++ lisp/isearch.el	2013-06-02 09:45:01 +0000
@@ -153,6 +153,17 @@ (defcustom isearch-hide-immediately t
   :type 'boolean
   :group 'isearch)
 
+(defcustom isearch-keep-stack-variables '(isearch-regexp)
+  "A set of search variables to keep and not to restore from the search stack."
+  :type '(set (const :tag "Regexp search" isearch-regexp)
+	      (const :tag "Case folding" isearch-case-fold-search)
+	      (const :tag "Invisible text" isearch-invisible)
+	      (const :tag "Filters" isearch-filter-predicates)
+	      (const :tag "Lax whitespace" isearch-lax-whitespace)
+	      (const :tag "Regexp lax whitespace" isearch-regexp-lax-whitespace))
+  :version "24.4"
+  :group 'isearch)
+
 (defcustom isearch-resume-in-command-history nil
   "If non-nil, `isearch-resume' commands are added to the command history.
 This allows you to resume earlier Isearch sessions through the
@@ -1109,6 +1143,7 @@ (cl-defstruct (isearch--state
                  (case-fold-search isearch-case-fold-search)
                  (pop-fun (if isearch-push-state-function
                               (funcall isearch-push-state-function)))
+		 (regexp isearch-regexp)
 		 (filter-predicates isearch-filter-predicates))))
   (string :read-only t)
   (message :read-only t)
@@ -1122,6 +1157,7 @@ (cl-defstruct (isearch--state
   (barrier :read-only t)
   (case-fold-search :read-only t)
   (pop-fun :read-only t)
+  (regexp :read-only t)
   (filter-predicates :read-only t))
 
 (defun isearch--set-state (cmd)
@@ -1136,6 +1172,8 @@ (defun isearch--set-state (cmd)
 	isearch-barrier (isearch--state-barrier cmd)
 	isearch-case-fold-search (isearch--state-case-fold-search cmd)
 	isearch-filter-predicates (isearch--state-filter-predicates cmd))
+  (unless (memq 'isearch-regexp isearch-keep-stack-variables)
+    (setq isearch-regexp (isearch--state-regexp cmd)))
   (if (functionp (isearch--state-pop-fun cmd))
       (funcall (isearch--state-pop-fun cmd) cmd))
   (goto-char (isearch--state-point cmd)))

bug#12986: Pressing backspace during isearch exits case-sensitive and word modes

[Lars Ingebrigtsen]

Juri Linkov <juri@jurta.org> writes:

>> Good point.  `isearch-delete-char' (that uses `isearch-pop-state')
>> restores word mode but not regexp mode.  This looks like
>> unaccountable inconsistency.  Perhaps it should restore
>> regexp mode as well.

[...]

> This patch adds a new user option `isearch-keep-stack-variables'
> intended to accompany another option `isearch-keep-mode-variables'
> added in bug#11378.  It adds a new option but doesn't change
> the default behavior - its default value is `isearch-regexp'
> that means to not keep regexp mode when pressing backspace
> (other search variables will be supported later as well).

I respun the patch for Emacs 28, but looking at it, I guess it doesn't
support actually restoring the modes it's discussing?

diff --git a/lisp/isearch.el b/lisp/isearch.el
index 0879f948cf..e9a3a2e921 100644
--- a/lisp/isearch.el
+++ b/lisp/isearch.el
@@ -167,6 +167,17 @@ isearch-hide-immediately
 the search, with the exception of the last successful match, if any."
   :type 'boolean)
 
+(defcustom isearch-keep-stack-variables '(isearch-regexp)
+  "A set of search variables to keep and not to restore from the search stack."
+  :type '(set (const :tag "Regexp search" isearch-regexp)
+	      (const :tag "Case folding" isearch-case-fold-search)
+	      (const :tag "Invisible text" isearch-invisible)
+	      (const :tag "Filters" isearch-filter-predicates)
+	      (const :tag "Lax whitespace" isearch-lax-whitespace)
+	      (const :tag "Regexp lax whitespace" isearch-regexp-lax-whitespace))
+  :version "28.1"
+  :group 'isearch)
+
 (defcustom isearch-resume-in-command-history nil
   "If non-nil, `isearch-resume' commands are added to the command history.
 This allows you to resume earlier Isearch sessions through the
@@ -1515,6 +1526,7 @@ isearch-update-from-string-properties
                  (case-fold-search isearch-case-fold-search)
                  (pop-fun (if isearch-push-state-function
                               (funcall isearch-push-state-function)))
+		 (regexp isearch-regexp)
                  (match-data isearch-match-data))))
   (string nil :read-only t)
   (message nil :read-only t)
@@ -1528,6 +1540,7 @@ isearch-update-from-string-properties
   (barrier nil :read-only t)
   (case-fold-search nil :read-only t)
   (pop-fun nil :read-only t)
+  (regexp :read-only t)
   (match-data nil :read-only t))
 
 (defun isearch--set-state (cmd)
@@ -1542,6 +1555,8 @@ isearch--set-state
 	isearch-barrier (isearch--state-barrier cmd)
 	isearch-case-fold-search (isearch--state-case-fold-search cmd)
 	isearch-match-data (isearch--state-match-data cmd))
+  (unless (memq 'isearch-regexp isearch-keep-stack-variables)
+    (setq isearch-regexp (isearch--state-regexp cmd)))
   (if (functionp (isearch--state-pop-fun cmd))
       (funcall (isearch--state-pop-fun cmd) cmd))
   (goto-char (isearch--state-point cmd)))

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

bug#12986: Pressing backspace during isearch exits case-sensitive and word modes

[Lars Ingebrigtsen]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> I respun the patch for Emacs 28, but looking at it, I guess it doesn't
> support actually restoring the modes it's discussing?

But I guess since the companion patch in bug#11378 was never applied,
it's a moot issue.  Juri, is this something that should still be worked
on, or has this patch (and the one in bug#11378) been abandoned?

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

bug#12986: Pressing backspace during isearch exits case-sensitive and word modes

[Juri Linkov]

>> I respun the patch for Emacs 28, but looking at it, I guess it doesn't
>> support actually restoring the modes it's discussing?
>
> But I guess since the companion patch in bug#11378 was never applied,
> it's a moot issue.  Juri, is this something that should still be worked
> on, or has this patch (and the one in bug#11378) been abandoned?

This feature is currently under discussion with Augusto.

bug#12986: Pressing backspace during isearch exits case-sensitive and word modes

[Lars Ingebrigtsen]

Juri Linkov <juri@jurta.org> writes:

>>> I respun the patch for Emacs 28, but looking at it, I guess it doesn't
>>> support actually restoring the modes it's discussing?
>>
>> But I guess since the companion patch in bug#11378 was never applied,
>> it's a moot issue.  Juri, is this something that should still be worked
>> on, or has this patch (and the one in bug#11378) been abandoned?
>
> This feature is currently under discussion with Augusto.

This was a year ago -- was there any further progress?

bug#12986: Pressing backspace during isearch exits case-sensitive and word modes

[Stefan Kangas]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Juri Linkov <juri@jurta.org> writes:
>
>>>> I respun the patch for Emacs 28, but looking at it, I guess it doesn't
>>>> support actually restoring the modes it's discussing?
>>>
>>> But I guess since the companion patch in bug#11378 was never applied,
>>> it's a moot issue.  Juri, is this something that should still be worked
>>> on, or has this patch (and the one in bug#11378) been abandoned?
>>
>> This feature is currently under discussion with Augusto.
>
> This was a year ago -- was there any further progress?

Another year has passed -- has there been any progress here?