re-search-forward hangs

re-search-forward hangs

[Andreas Röhler]

Hi,

emacs -Q, when evaluating from in scratch-buffer


(while (re-search-forward "" nil t 1))

Emacs hangs.

Same with plain

(while (search-forward "" nil t 1))

Expect it terminating.

Are there reasons for this?

Thanks,
Andreas
GNU Emacs 27.0.50 (build 1, i686-pc-linux-gnu, GTK+ Version 3.14.5) of 
2018-05-29

Re: re-search-forward hangs

[Robert Girault]

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

> Hi,
>
> emacs -Q, when evaluating from in scratch-buffer
>
>
> (while (re-search-forward "" nil t 1))
>
> Emacs hangs.
>
> Same with plain
>
> (while (search-forward "" nil t 1))
>
> Expect it terminating.

Why do you expect it to terminate?  

> Are there reasons for this?

The obvious reason is that "" is found anywhere, so this search will
always succeed.  This makes sense to me.  Why do you think this
shouldn't happen?

Re: re-search-forward hangs

[Eli Zaretskii]

> From: Andreas Röhler <andreas.roehler@easy-emacs.de>
> Date: Mon, 30 Jul 2018 14:24:06 +0200
> 
> (while (re-search-forward "" nil t 1))
> 
> Emacs hangs.
> 
> Same with plain
> 
> (while (search-forward "" nil t 1))
> 
> Expect it terminating.
> 
> Are there reasons for this?

Why did you expect it to terminate?

Re: re-search-forward hangs

[Ben Bacarisse]

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

> emacs -Q, when evaluating from in scratch-buffer
>
> (while (re-search-forward "" nil t 1))
>
> Emacs hangs.
>
> Same with plain
>
> (while (search-forward "" nil t 1))
>
> Expect it terminating.

Why do you expect a difference?  I would expect both to loop
indefinitely (and they do in my slightly older Emacs).

> Are there reasons for this?

Both loop because the match always succeeds.  Neither match moves point
so neither can fail.

-- 
Ben.

Re: re-search-forward hangs

[Stefan Monnier]

> emacs -Q, when evaluating from in scratch-buffer
> (while (re-search-forward "" nil t 1))
> Emacs hangs.

It doesn't hang for me.  The command doesn't terminate, of course, but
you can interrupt it with C-g.


        Stefan

Re: re-search-forward hangs

[Andreas Röhler]

On 30.07.2018 16:33, Eli Zaretskii wrote:
>> From: Andreas Röhler <andreas.roehler@easy-emacs.de>
>> Date: Mon, 30 Jul 2018 14:24:06 +0200
>>
>> (while (re-search-forward "" nil t 1))
>>
>> Emacs hangs.
>>
>> Same with plain
>>
>> (while (search-forward "" nil t 1))
>>
>> Expect it terminating.
>>
>> Are there reasons for this?
> 
> Why did you expect it to terminate?
> 

Expect a call of search-forward from  EOB not being successful.

Re: re-search-forward hangs

[Robert Girault]

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

> On 30.07.2018 16:33, Eli Zaretskii wrote:
>>> From: Andreas Röhler <andreas.roehler@easy-emacs.de>
>>> Date: Mon, 30 Jul 2018 14:24:06 +0200
>>>
>>> (while (re-search-forward "" nil t 1))
>>>
>>> Emacs hangs.
>>>
>>> Same with plain
>>>
>>> (while (search-forward "" nil t 1))
>>>
>>> Expect it terminating.
>>>
>>> Are there reasons for this?
>>
>> Why did you expect it to terminate?
>
> Expect a call of search-forward from  EOB not being successful.

I see.  But searching for the empty string seems to be an always
successful operation as long as you have a string.  The empty string is
before or after any character in a string, so I would expect it to be
always found by {re-}search-forward.

Re: re-search-forward hangs

[Andreas Röhler]

On 30.07.2018 21:35, Robert Girault wrote:
> Andreas Röhler <andreas.roehler@easy-emacs.de> writes:
> 
>> On 30.07.2018 16:33, Eli Zaretskii wrote:
>>>> From: Andreas Röhler <andreas.roehler@easy-emacs.de>
>>>> Date: Mon, 30 Jul 2018 14:24:06 +0200
>>>>
>>>> (while (re-search-forward "" nil t 1))
>>>>
>>>> Emacs hangs.
>>>>
>>>> Same with plain
>>>>
>>>> (while (search-forward "" nil t 1))
>>>>
>>>> Expect it terminating.
>>>>
>>>> Are there reasons for this?
>>>
>>> Why did you expect it to terminate?
>>
>> Expect a call of search-forward from  EOB not being successful.
> 
> I see.  But searching for the empty string seems to be an always
> successful operation as long as you have a string.  The empty string is
> before or after any character in a string, so I would expect it to be
> always found by {re-}search-forward.
> 

Hmm, when trying at BOB with some content, changing the times argument

(message "%s" (search-forward "" nil t 1))
(message "%s" (search-forward "" nil t 2))
(message "%s" (search-forward "" nil t 3))

the result remains "43" - looks like the match is not regular.