bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

[Andreas Röhler]

(defun end-of-sexp ()
   "Move point to the end of the current sexp.
\[This is an internal function.]"
   (let ((char-syntax (char-syntax (char-after))))
     (if (or (eq char-syntax ?\))
	    (and (eq char-syntax ?\") (in-string-p)))
	(forward-char 1)
       (forward-sexp 1))))

"or" asks if inside a string and calls (forward-char 1).

This must fail with some probability with
triple-quoted-strings as used in Python.

Solution:

When inside a string, jump to (nth 8 (syntax-ppss)) and
call (forward-sexp 1) from there.

bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

[Leo Liu]

On 2013-03-16 15:51 +0800, Andreas Röhler wrote:
> When inside a string, jump to (nth 8 (syntax-ppss)) and
> call (forward-sexp 1) from there.

In my experience (a while ago, I use little python these days) this
cannot reliably goes to the end of the triple-quoted string.

Leo

bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

[Andreas Röhler]

Am 16.03.2013 11:33, schrieb Leo Liu:
> On 2013-03-16 15:51 +0800, Andreas Röhler wrote:
>> When inside a string, jump to (nth 8 (syntax-ppss)) and
>> call (forward-sexp 1) from there.
>
> In my experience (a while ago, I use little python these days) this
> cannot reliably goes to the end of the triple-quoted string.
>
> Leo
>
>
>
>
>

Indeed seems a bug in python-mode.el at this place just now.
Anyway - forward-sexp must go to the end of string from it's beginning,
a string is a balanced expression.


Best,

Andreas

bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

[Stefan Monnier]

>> When inside a string, jump to (nth 8 (syntax-ppss)) and
>> call (forward-sexp 1) from there.

Sounds right, yes.

> In my experience (a while ago, I use little python these days) this
> cannot reliably goes to the end of the triple-quoted string.

That would be a bug in python-mode, then.


        Stefan

bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

[Lars Ingebrigtsen]

Andreas Röhler <andreas.roehler@easy-emacs.de> writes:

> (defun end-of-sexp ()
>   "Move point to the end of the current sexp.
> \[This is an internal function.]"
>   (let ((char-syntax (char-syntax (char-after))))
>     (if (or (eq char-syntax ?\))
> 	    (and (eq char-syntax ?\") (in-string-p)))
> 	(forward-char 1)
>       (forward-sexp 1))))
>
> "or" asks if inside a string and calls (forward-char 1).
>
> This must fail with some probability with
> triple-quoted-strings as used in Python.
>
> Solution:
>
> When inside a string, jump to (nth 8 (syntax-ppss)) and
> call (forward-sexp 1) from there.

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

It's not quite clear to me what the actual problem here is.  Do you have
a test case where thingatpt does the wrong thing in Python mode?

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

bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

[Andreas Röhler]

On 23.08.21 03:09, Lars Ingebrigtsen wrote:
> Andreas Röhler <andreas.roehler@easy-emacs.de> writes:
>
>> (defun end-of-sexp ()
>>    "Move point to the end of the current sexp.
>> \[This is an internal function.]"
>>    (let ((char-syntax (char-syntax (char-after))))
>>      (if (or (eq char-syntax ?\))
>> 	    (and (eq char-syntax ?\") (in-string-p)))
>> 	(forward-char 1)
>>        (forward-sexp 1))))
>>
>> "or" asks if inside a string and calls (forward-char 1).
>>
>> This must fail with some probability with
>> triple-quoted-strings as used in Python.
>>
>> Solution:
>>
>> When inside a string, jump to (nth 8 (syntax-ppss)) and
>> call (forward-sexp 1) from there.
> (I'm going through old bug reports that unfortunately weren't resolved
> at the time.)
>
> It's not quite clear to me what the actual problem here is.

It's only abstract reasoning when looking at the code.

Maybe put the Python code below at the top of some buffer and run the 
test delivered below:

# Python

def main():
     """Some hint"""
     if len(sys.argv) == 1:
         usage()

;; Elisp
(defun forward-sexp-text ()
   (interactive)
   (goto-char 30 )
   (forward-sexp))


>   Do you have
> a test case where thingatpt does the wrong thing in Python mode?
>

bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

[Lars Ingebrigtsen]

[-- Attachment #1: Type: text/plain, Size: 697 bytes --]

Andreas Röhler <andreas.roehler@easy-emacs.de> writes:

> Maybe put the Python code below at the top of some buffer and run the
> test delivered below:
>
> # Python
>
> def main():
>     """Some hint"""
>     if len(sys.argv) == 1:
>         usage()
>
> ;; Elisp
> (defun forward-sexp-text ()
>   (interactive)
>   (goto-char 30 )
>   (forward-sexp))

By that goto-char, do you mean the "S" character?  (It's where point
lands in the buffer included below.)  If so, `forward-sex'
(in Emacs 28) does the correct thing -- it moves to the end of "Some".

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

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: sexp.py --]
[-- Type: text/x-python, Size: 85 bytes --]

# Python

def main():
    """Some hint"""
    if len(sys.argv) == 1:
        usage()

bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

[Andreas Röhler]

On 25.08.21 12:38, Lars Ingebrigtsen wrote:
> Andreas Röhler <andreas.roehler@easy-emacs.de> writes:
>
>> Maybe put the Python code below at the top of some buffer and run the
>> test delivered below:
>>
>> # Python
>>
>> def main():
>>      """Some hint"""
>>      if len(sys.argv) == 1:
>>          usage()
>>
>> ;; Elisp
>> (defun forward-sexp-text ()
>>    (interactive)
>>    (goto-char 30 )
>>    (forward-sexp))
> By that goto-char, do you mean the "S" character?


No, sorry, the 5th double-quote of the triple-quoted-string.

> (It's where point
> lands in the buffer included below.)  If so, `forward-sex'
> (in Emacs 28) does the correct thing -- it moves to the end of "Some".
>

bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

[Lars Ingebrigtsen]

Andreas Röhler <andreas.roehler@easy-emacs.de> writes:

>>> Maybe put the Python code below at the top of some buffer and run the
>>> test delivered below:
>>>
>>> # Python
>>>
>>> def main():
>>>      """Some hint"""
>>>      if len(sys.argv) == 1:
>>>          usage()
>>>
>>> ;; Elisp
>>> (defun forward-sexp-text ()
>>>    (interactive)
>>>    (goto-char 30 )
>>>    (forward-sexp))
>> By that goto-char, do you mean the "S" character?
>
> No, sorry, the 5th double-quote of the triple-quoted-string.

That gives me

forward-sexp: Scan error: "Unbalanced parentheses", 40, 86

which I think is the right thing?

This bug report started with thingatpt.  What is the real thingatpt use
case that doesn't work in triple-quoted strings?

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

bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

[Andreas Röhler]

On 26.08.21 16:05, Lars Ingebrigtsen wrote:
> Andreas Röhler <andreas.roehler@easy-emacs.de> writes:
>
>>>> Maybe put the Python code below at the top of some buffer and run the
>>>> test delivered below:
>>>>
>>>> # Python
>>>>
>>>> def main():
>>>>       """Some hint"""
>>>>       if len(sys.argv) == 1:
>>>>           usage()
>>>>
>>>> ;; Elisp
>>>> (defun forward-sexp-text ()
>>>>     (interactive)
>>>>     (goto-char 30 )
>>>>     (forward-sexp))
>>> By that goto-char, do you mean the "S" character?
>> No, sorry, the 5th double-quote of the triple-quoted-string.
> That gives me
>
> forward-sexp: Scan error: "Unbalanced parentheses", 40, 86
>
> which I think is the right thing?

If inside a string, forward-sexp --or end-of-- should reach the end of 
this string.

BTW end-of-sexp is declared obsolete meanwhile. Below an adapted test.


(defun my-thing-at-point--end-of-sexp-test ()
   (interactive)
   (goto-char (point-max))
   (search-backward "\"" nil t 6)
   ;; the only one ending correctly behind the string
   (thing-at-point--end-of-sexp)
   (goto-char (point-max))
   (search-backward "\"" nil t 6)
   (forward-char 4)
   (thing-at-point--end-of-sexp)
   (goto-char (point-max))
   (search-backward "\"" nil t 3)
   (thing-at-point--end-of-sexp)
   (goto-char (point-max))
   (search-backward "\"" nil t 2)
   (thing-at-point--end-of-sexp))

and here the Python code to test again:


def main():
      """Some hint"""
      if len(sys.argv) == 1:
          usage()



>
> This bug report started with thingatpt.  What is the real thingatpt use
> case that doesn't work in triple-quoted strings?
>

bug#13973: Subject: 24.3; thingatpt.el, end-of-sexp

[Lars Ingebrigtsen]

Andreas Röhler <andreas.roehler@easy-emacs.de> writes:

> BTW end-of-sexp is declared obsolete meanwhile. Below an adapted test.

Thanks.  The odd movements happen in these cases:

"|""Some hint"""

"""Some hint|"""

where | is point.

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