bug#4259: Cannot indent fortran continue statement in emacs 23

bug#4259: Cannot indent fortran continue statement in emacs 23

[Ushnish Basu]

Hello,

I believe I came across a bug in the latest version of the fortran mode 
in emacs 23. Currently I am using version 23.1.1, but I did not have 
this problem with 23.0.60 from CVS.

Please see the following sample code - I have arranged it as such only 
to illustrate the problem:

       subroutine first
       do 10 i=1,10
  10     continue
       end
       do 20 i=1,100
  20     continue


If I run fortran-indent-line on the first continue (label 10) it works 
fine. However, if I run it on the second one (label 20), I get the error:

fortran-check-end-prog-re: Lisp nesting exceeds 'max-lisp-eval-depth'

presumably because it is after the 'end'

If I increase max-lisp-eval-depth (and max-specpdl-size) to 10000, then 
I get the error

fortran-check-end-prog-re: Variable binding depth exceeds max-specpdl-size

Is this something that has been fixed in CVS?

I look forward to hearing back from you.

Thanks,

Ushnish

bug#4259: Cannot indent fortran continue statement in emacs 23

[Glenn Morris]

Please use M-x report-emacs-bug to report Emacs bugs, rather than
sending mail to addresses in the source files.


Ushnish Basu wrote (on Tue, 25 Aug 2009 at 12:09 -0700):

> Currently I am using version 23.1.1, but I did not have this problem
> with 23.0.60 from CVS.
[...]
>        subroutine first
>        do 10 i=1,10
>   10     continue
>        end
>        do 20 i=1,100
>   20     continue
> 
> If I run fortran-indent-line on the first continue (label 10) it works 
> fine. However, if I run it on the second one (label 20), I get the error:
> 
> fortran-check-end-prog-re: Lisp nesting exceeds 'max-lisp-eval-depth'


The example does not look like valid Fortran, so I'm not too bothered
if fortran-mode cannot indent it (although an infinite loop is not a
nice way to fail).

Do you have a minimum example of a valid Fortran file that fails in
this way?

The problem is another (see eg bug #2106) infinite loop in
end-of-defun, when called from the start of the "20" line. It's not
immediately obvious to me whether this is a problem in fortran-mode or
in end-of-defun.

bug#4259: Cannot indent fortran continue statement in emacs 23

[Ushnish Basu]

Glenn Morris wrote:
> Please use M-x report-emacs-bug to report Emacs bugs, rather than
> sending mail to addresses in the source files.
> 
> 

My apologies - I thought of that after I sent off this email. I have now 
submitted a bug report directly, but I guess this bug was already filed 
as #4259, so the second bug would duplicate.

> Ushnish Basu wrote (on Tue, 25 Aug 2009 at 12:09 -0700):
> 
>> Currently I am using version 23.1.1, but I did not have this problem
>> with 23.0.60 from CVS.

[snip]

> 
> Do you have a minimum example of a valid Fortran file that fails in
> this way?
> 

The bug shows for the following program:

       subroutine first
       do 10 i=1,10
  10     continue
       end
       subroutine second
       do 20 i=1,100
  20     continue
       end

I sent the earlier invalid version just to highlight that it was the 
'end' statement that seemed to be making the difference.

I get the bug even in the current(?) CVS version 23.1.50.1

> The problem is another (see eg bug #2106) infinite loop in
> end-of-defun, when called from the start of the "20" line. It's not
> immediately obvious to me whether this is a problem in fortran-mode or
> in end-of-defun.

Thanks for your help.

Best regards,

Ushnish

bug#4259: Cannot indent fortran continue statement in emacs 23

[Glenn Morris]

Ushnish Basu wrote:

>       subroutine first
>       do 10 i=1,10
>  10     continue
>       end
>       subroutine second
>       do 20 i=1,100
>  20     continue
>       end


Thanks, this looks like a (long-standing but previously hidden)
fortran.el bug with motion commands. I will find a fix.

bug#4259: marked as done (Cannot indent fortran continue statement in emacs 23)

[Emacs bug Tracking System]

[-- Attachment #1: Type: text/plain, Size: 894 bytes --]

Your message dated Sat, 29 Aug 2009 21:40:33 -0400
with message-id <ja8wh23wq6.fsf@fencepost.gnu.org>
and subject line Re: Cannot indent fortran continue statement in emacs 23
has caused the Emacs bug report #4259,
regarding Cannot indent fortran continue statement in emacs 23
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@emacsbugs.donarmstrong.com
immediately.)


-- 
4259: http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=4259
Emacs Bug Tracking System
Contact owner@emacsbugs.donarmstrong.com with problems

[-- Attachment #2: Type: message/rfc822, Size: 3046 bytes --]

From: Ushnish Basu <ubasu@cal.berkeley.edu>
To: "Michael D. Prange" <prange@erl.mit.edu>, Glenn Morris <rgm@gnu.org>
Subject: Cannot indent fortran continue statement in emacs 23
Date: Tue, 25 Aug 2009 12:09:30 -0700
Message-ID: <4A9436EA.3070303@cal.berkeley.edu>

Hello,

I believe I came across a bug in the latest version of the fortran mode 
in emacs 23. Currently I am using version 23.1.1, but I did not have 
this problem with 23.0.60 from CVS.

Please see the following sample code - I have arranged it as such only 
to illustrate the problem:

       subroutine first
       do 10 i=1,10
  10     continue
       end
       do 20 i=1,100
  20     continue


If I run fortran-indent-line on the first continue (label 10) it works 
fine. However, if I run it on the second one (label 20), I get the error:

fortran-check-end-prog-re: Lisp nesting exceeds 'max-lisp-eval-depth'

presumably because it is after the 'end'

If I increase max-lisp-eval-depth (and max-specpdl-size) to 10000, then 
I get the error

fortran-check-end-prog-re: Variable binding depth exceeds max-specpdl-size

Is this something that has been fixed in CVS?

I look forward to hearing back from you.

Thanks,

Ushnish


[-- Attachment #3: Type: message/rfc822, Size: 2078 bytes --]

From: Glenn Morris <rgm@gnu.org>
To: 4259-done@emacsbugs.donarmstrong.com
Subject: Re: Cannot indent fortran continue statement in emacs 23
Date: Sat, 29 Aug 2009 21:40:33 -0400
Message-ID: <ja8wh23wq6.fsf@fencepost.gnu.org>


I believe this is fixed in the current CVS trunk. It would be great if
you are able to test it out.

2009-08-30  Glenn Morris  <rgm at gnu.org>

  * progmodes/fortran.el (fortran-start-prog-re): New constant, extracted
  from fortran-current-defun.
  (fortran-beginning-of-subprogram): Be more precise about finding the
  start, to avoid an infinite loop in end-of-defun.  (Bug#4259)
  (fortran-end-of-subprogram): Simplify.
  (fortran-current-defun): Use fortran-start-prog-re.

bug#4259: closed by Glenn Morris <rgm@gnu.org> (Re: Cannot indent fortran continue statement in emacs 23)

[Ushnish Basu]

Thanks, I tried it out and it works. Thanks for the quick bugfix.

Ushnish

Emacs bug Tracking System wrote:
> This is an automatic notification regarding your bug report
> which was filed against the emacs package:
> 
> #4259: Cannot indent fortran continue statement in emacs 23
> 
> It has been closed by Glenn Morris <rgm@gnu.org>.
> 
> Their explanation is attached below along with your original report.
> If this explanation is unsatisfactory and you have not received a
> better one in a separate message then please contact Glenn Morris <rgm@gnu.org> by
> replying to this email.
> 
> 
> 
> 
> ------------------------------------------------------------------------
> 
> Subject:
> Re: Cannot indent fortran continue statement in emacs 23
> From:
> Glenn Morris <rgm@gnu.org>
> Date:
> Sat, 29 Aug 2009 21:40:33 -0400
> To:
> 4259-done@emacsbugs.donarmstrong.com
> 
> To:
> 4259-done@emacsbugs.donarmstrong.com
> 
> 
> I believe this is fixed in the current CVS trunk. It would be great if
> you are able to test it out.
> 
> 2009-08-30  Glenn Morris  <rgm at gnu.org>
> 
>   * progmodes/fortran.el (fortran-start-prog-re): New constant, extracted
>   from fortran-current-defun.
>   (fortran-beginning-of-subprogram): Be more precise about finding the
>   start, to avoid an infinite loop in end-of-defun.  (Bug#4259)
>   (fortran-end-of-subprogram): Simplify.
>   (fortran-current-defun): Use fortran-start-prog-re.
> 
> 
> ------------------------------------------------------------------------
> 
> Subject:
> Cannot indent fortran continue statement in emacs 23
> From:
> Ushnish Basu <ubasu@cal.berkeley.edu>
> Date:
> Tue, 25 Aug 2009 12:09:30 -0700
> To:
> "Michael D. Prange" <prange@erl.mit.edu>, Glenn Morris <rgm@gnu.org>
> 
> To:
> "Michael D. Prange" <prange@erl.mit.edu>, Glenn Morris <rgm@gnu.org>
> 
> 
> Hello,
> 
> I believe I came across a bug in the latest version of the fortran mode 
> in emacs 23. Currently I am using version 23.1.1, but I did not have 
> this problem with 23.0.60 from CVS.
> 
> Please see the following sample code - I have arranged it as such only 
> to illustrate the problem:
> 
>       subroutine first
>       do 10 i=1,10
>  10     continue
>       end
>       do 20 i=1,100
>  20     continue
> 
> 
> If I run fortran-indent-line on the first continue (label 10) it works 
> fine. However, if I run it on the second one (label 20), I get the error:
> 
> fortran-check-end-prog-re: Lisp nesting exceeds 'max-lisp-eval-depth'
> 
> presumably because it is after the 'end'
> 
> If I increase max-lisp-eval-depth (and max-specpdl-size) to 10000, then 
> I get the error
> 
> fortran-check-end-prog-re: Variable binding depth exceeds max-specpdl-size
> 
> Is this something that has been fixed in CVS?
> 
> I look forward to hearing back from you.
> 
> Thanks,
> 
> Ushnish
>