Re: Debugging calendar gremlin in emacs 22.0.50.1

Re: Debugging calendar gremlin in emacs 22.0.50.1

[Glenn Morris]

ken wrote:

> After working already dozens of times, "M-x diary" suddenly fails with
> error message "Wrong type argument: sequencep, t".

The general way to get more information in such cases is to use M-x
toggle-debug-on-error to turn on the debugger.

Also check if the error occurs with an emacs started with -q
--no-site-file.

Re: Debugging calendar gremlin in emacs 22.0.50.1

[ken]

Glenn Morris wrote:
> ken wrote:
> 
>> After working already dozens of times, "M-x diary" suddenly fails with
>> error message "Wrong type argument: sequencep, t".
> 
> The general way to get more information in such cases is to use M-x
> toggle-debug-on-error to turn on the debugger.
> 
> Also check if the error occurs with an emacs started with -q
> --no-site-file.

Thanks for that.  I vaguely remember those.  Now they're in my personal
documentation.

After invoking "emacs -q --no-site-file" it loads fine.  Then I visit
~/.emacs and evaluate the following that are needed for "M-x diary":

(setq diary-file "~/diary.emacs")
(setq number-of-diary-entries 14)
(setq calendar-latitude 40.4)
(setq calendar-longitude -51.9)

Everything's still good.  So then I run "M-x diary" and get this:

The local variables list in diary.emacs
contains values that may not be safe (*).

Do you really want to apply it?  You can type
y - to apply the local variables list.
n - to ignore....
! - to apply... and permanently mark these...

  * auto-fill-mode: off
    fill-column: 999999

The reason for this (I'm guessing) is that I have this at the top of
~/diary.emacs:

;-*- mode: lisp; auto-fill-mode: off; fill-column: 999999 -*-

I hit y and emacs loads what it thinks is the diary.

~/diary.emacs contains:

---------------------------------------------------------
;-*- mode: lisp; auto-fill-mode: off; fill-column: 999999 -*-

%%(diary-anniversary 9 18 1970) %d years ago it was 9/18/1970

%%(diary-anniversary 9 19 1911) %d years ago it was 9/19/1911

%%(diary-anniversary 9 20 1934) %d years ago it was 9/20/1934

%%(diary-anniversary 9 20 1885) %d years ago it was 9/20/1885

%%(diary-anniversary 9 20 1878) $d years ago it was 9/20/1878

;end of ~/diary.emacs
---------------------------------------------------------


And this is exactly the output of the output I get from the
above-invoked "M-x diary" except that the first line (beginning ";-*-
mode... is not displayed, and the rest of the lines are single-spaced
(see below)--  but if I make this buffer (entitled "Diary for Tuesday,
September 19, 2006") the current buffer, do "C-x h M-w" and paste it
somewhere, it'll show that first line.  Quite weird.

Even more weird:  If I then do "C-x 5 f ~/diary.emacs", it does nothing
the first time.  I can go make another pot of coffee, come back, and
there's no new frame.  If I then run exactly the same command a second
time, emacs then opens a new frame entitled "Diary for Tuesday,
September 19, 2006" (i.e., displayed in the modeline) and which has
exactly the same contents as the Diary buffer, i.e.:

-------------------------------------------------------

%%(diary-anniversary 9 18 1970) %d years ago it was 9/18/1970
%%(diary-anniversary 9 19 1911) %d years ago it was 9/19/1911
%%(diary-anniversary 9 20 1934) %d years ago it was 9/20/1934
%%(diary-anniversary 9 20 1885) %d years ago it was 9/20/1885
%%(diary-anniversary 9 20 1878) $d years ago it was 9/20/1878
----------------------------------------------------------

Next (in this new frame) I do "C-x f ~/diary.emacs RETURN": Nothing
happens.  Doing it a second time, nothing happens.  (I'm guessing that
emacs thinks it has already loaded the file, but a lot here is
guessing.)  So I do "C-x v ~/diary.emacs" and the file is loaded with
its (correct) name in the modeline.

Anybody understand what's going on?

                                *

Thinking again back to the sudden appearance of the gremlin, I did do a
kernel upgrade.  Though I really don't see how it could be the cause of
the problem, that was the only thing that changed (that I can think of,
that I know of) between the time I wasn't getting the error message
"Wrong type argument: sequencep, t" and when I started getting it.  The
kernel used now is 2.6.11.4-21.14-default.

Re: Debugging calendar gremlin in emacs 22.0.50.1

[Peter Dyballa]

Am 19.09.2006 um 14:19 schrieb ken:

> Everything's still good.  So then I run "M-x diary" and get this:
>
> The local variables list in diary.emacs
> contains values that may not be safe (*).
>
> Do you really want to apply it?  You can type
> y - to apply the local variables list.
> n - to ignore....
> ! - to apply... and permanently mark these...
>
>   * auto-fill-mode: off
>     fill-column: 999999
>
> The reason for this (I'm guessing) is that I have this at the top of
> ~/diary.emacs:
>
> ;-*- mode: lisp; auto-fill-mode: off; fill-column: 999999 -*-
>
> I hit y and emacs loads what it thinks is the diary.


This has to do with a security feature introduced early this summer:  
safe-local-variable-values. If you answer ``!´´ then the name will be  
recorded in your customisation and you won't be asked again to  
confirm that this local variable is safe. (I had to type ! a few  
times to make coding-system in different contexts a safe local  
variable.)


>
> ~/diary.emacs contains:
>
> ---------------------------------------------------------
> ;-*- mode: lisp; auto-fill-mode: off; fill-column: 999999 -*-

I think the convention is to start such a line with ;;;. Then I think  
it's needed that some space is left between the comment and the  
expression, i.e. the -*- ...; -*- part.

>
> %%(diary-anniversary 9 18 1970) %d years ago it was 9/18/1970
>
> %%(diary-anniversary 9 19 1911) %d years ago it was 9/19/1911
>
> %%(diary-anniversary 9 20 1934) %d years ago it was 9/20/1934
>
> %%(diary-anniversary 9 20 1885) %d years ago it was 9/20/1885
>
> %%(diary-anniversary 9 20 1878) $d years ago it was 9/20/1878
>
> ;end of ~/diary.emacs

At this semicolon, could be, the set of local variables ends ... thy  
syntax is:

	-*- <some keyword>: <some other keyword>; <potentially other pair>; -*-

You also can put at the end:

;; Local variables:
;; mode: lisp
;; auto-fill-mode: off
;; fill-column: 999999
;; End:


>
> Next (in this new frame) I do "C-x f ~/diary.emacs RETURN": Nothing
> happens.

This is correct: you set fill-column. (But the other input, i.e.  
``~ / d i a r y . e m a c s´´ and the final RET should do something ...)



--
Greetings

   Pete

"We have to expect it, otherwise we would be surprised."

Re: Debugging calendar gremlin in emacs 22.0.50.1

[ken]

Peter Dyballa wrote:
> 
> I think the convention is to start such a line with ;;;. Then I think
> it's needed that some space is left between the comment and the
> expression, i.e. the -*- ...; -*- part.

It's nice that emacs works with the incorrect syntax also.  :)


-- 
"It is not knowable how long that conflict would last, it could last,
you know, six days, six weeks.  I doubt six months."
	--Secretary of Defense Donald Rumsfeld, 2/7/03

Re: Debugging calendar gremlin in emacs 22.0.50.1

[Glenn Morris]

ken wrote:

> And this is exactly the output of the output I get from the
> above-invoked "M-x diary" except that the first line (beginning
> ";-*- mode... is not displayed, and the rest of the lines are
> single-spaced (see below)-- but if I make this buffer (entitled
> "Diary for Tuesday, September 19, 2006") the current buffer, do "C-x
> h M-w" and paste it somewhere, it'll show that first line. Quite
> weird.

I don't think there's anything weird here. The standard ("simple")
diary display uses invisible text to hide parts of the diary buffer.
So visiting the diary file and displaying the diary buffer give
different results. Copying and pasting from the simple diary can also
be confusing. CVS emacs displays a warning header about this by
default.

> Even more weird: If I then do "C-x 5 f ~/diary.emacs", it does
> nothing the first time. I can go make another pot of coffee, come
> back, and there's no new frame. If I then run exactly the same
> command a second time, emacs then opens a new frame entitled "Diary
> for Tuesday, September 19, 2006" (i.e., displayed in the modeline)
> and which has exactly the same contents as the Diary buffer, i.e.:

The default simple diary display visits the diary file in a buffer, so
it's best to kill it if you want to actually load the raw file (as it
were).

> Thinking again back to the sudden appearance of the gremlin, I did
> do a kernel upgrade. Though I really don't see how it could be the
> cause of the problem, that was the only thing that changed (that I
> can think of, that I know of) between the time I wasn't getting the
> error message "Wrong type argument: sequencep, t" and when I started
> getting it. The kernel used now is 2.6.11.4-21.14-default.

I don't think the kernel has anything to do with this error, which is
unrelated to all the stuff you describe above. So where's the
backtrace from this sequencep error?

You probably want to update and rebuild your CVS before investing much
time in debugging this.

Re: Debugging calendar gremlin in emacs 22.0.50.1

[Glenn Morris]

Replies to the list, rather than me, please.


ken wrote (on Tue, 19 Sep 2006 at 19:10 -0400):

> Here's my test diary-file:
> ---------------------------------------------------------
> ;-*- mode: lisp; auto-fill-mode: off; fill-column: 999999 -*-
>
> %%(diary-anniversary 9 18 1970) %d years ago it was 9/18/1970
[...]

> I run M-x diary and this is the output:

> %%(diary-anniversary 9 18 1970) %d years ago it was 9/18/1970
[...]

> I didn't think that is what it's supposed to do.


Without knowing what you thought it was supposed to do, it's hard to
comment. If you meant, why are the %d's etc not being expanded, such
things only happen if you use fancy-diary-display, and from the sound
of it you aren't.


> > unrelated to all the stuff you describe above. So where's the
> > backtrace from this sequencep error?
>
> Oh, yeah. the toggle-debug-on-error helped find the problem. It was
> a diary-related thing in my .emacs.

Re: Debugging calendar gremlin in emacs 22.0.50.1

[ken]

Glenn Morris wrote:
> 
> Replies to the list, rather than me, please.

Accident, sorry.  It seems that posts sent out by the list have no
Reply-To field in them.


> ....
> 
> Without knowing what you thought it was supposed to do, it's hard to
> comment. If you meant, why are the %d's etc not being expanded, such
> things only happen if you use fancy-diary-display, and from the sound
> of it you aren't.

I've always used fancy-diary-display so, yes, I was assuming that the
%d's would be expanded with their values with simple-diary-display and
other functions like diary-sunrise-sunset and diary-phases-of-moon would
likewise be evaluated and their output displayed.  I guess that by
thinking the ordinary language meanings of "simple" and "fancy" I
misunderstood.  So... it's one of those cases of "That's not a bug, it's
a *feature*!"

I'm curious now... why would anyone want to use the simple diary
display...?  And why is it the default?


-- 
"It is not knowable how long that conflict would last, it could last,
you know, six days, six weeks.  I doubt six months."
	--Secretary of Defense Donald Rumsfeld, 2/7/03

Re: Debugging calendar gremlin in emacs 22.0.50.1

[Edward M. Reingold]

>>>>> "k" == ken  <gebser@speakeasy.net> writes:

    k> I'm curious now... why would anyone want to use the simple diary
    k> display...?  And why is it the default?

Nobody in his right mind would.  Stallman forced me to write it that way; I
originally had the fancy display as the default (indeed only) display
mechanism.

-- 

Professor Edward M. Reingold                Email: reingold@iit.edu
Department of Computer Science              Voice: (312) 567-3309
Illinois Institute of Technology            Assistant: (312) 567-5152
Stuart Building                             Fax:   (312) 567-5067
10 West 31st Street, Suite 236
Chicago, IL  60616-3729  U.S.A.

Re: Debugging calendar gremlin in emacs 22.0.50.1

[ken]

Edward M. Reingold wrote:
>>>>>> "k" == ken  <gebser@speakeasy.net> writes:
> 
>     k> I'm curious now... why would anyone want to use the simple diary
>     k> display...?  And why is it the default?
> 
> Nobody in his right mind would.  Stallman forced me to write it that way; I
> originally had the fancy display as the default (indeed only) display
> mechanism.

Okay, that makes sense now.  I believe it's a requirement for a Genius
Grant that one exhibit a small percentage of nerdique kookiness.  &^]


-- 
"It is not knowable how long that conflict would last, it could last,
you know, six days, six weeks.  I doubt six months."
	--Secretary of Defense Donald Rumsfeld, 2/7/03

Re: Debugging calendar gremlin in emacs 22.0.50.1

[ken]

Glenn Morris wrote:
> ken wrote:
> 
>> And this is exactly the output of the output I get from the
>> above-invoked "M-x diary" except that the first line (beginning
>> ";-*- mode... is not displayed, and the rest of the lines are
>> single-spaced (see below)-- but if I make this buffer (entitled
>> "Diary for Tuesday, September 19, 2006") the current buffer, do "C-x
>> h M-w" and paste it somewhere, it'll show that first line. Quite
>> weird.
> 
> I don't think there's anything weird here. The standard ("simple")
> diary display uses invisible text to hide parts of the diary buffer.
> So visiting the diary file and displaying the diary buffer give
> different results. Copying and pasting from the simple diary can also
> be confusing. CVS emacs displays a warning header about this by
> default.

I guess my post was too long to get the point across.  Here's my test
diary-file:

---------------------------------------------------------
;-*- mode: lisp; auto-fill-mode: off; fill-column: 999999 -*-

%%(diary-anniversary 9 18 1970) %d years ago it was 9/18/1970

%%(diary-anniversary 9 19 1911) %d years ago it was 9/19/1911

%%(diary-anniversary 9 20 1934) %d years ago it was 9/20/1934

%%(diary-anniversary 9 20 1885) %d years ago it was 9/20/1885

%%(diary-anniversary 9 20 1878) $d years ago it was 9/20/1878

;end of ~/diary.emacs
---------------------------------------------------------

I run M-x diary and this is the output:

-------------------------------------------------------

%%(diary-anniversary 9 18 1970) %d years ago it was 9/18/1970
%%(diary-anniversary 9 19 1911) %d years ago it was 9/19/1911
%%(diary-anniversary 9 20 1934) %d years ago it was 9/20/1934
%%(diary-anniversary 9 20 1885) %d years ago it was 9/20/1885
%%(diary-anniversary 9 20 1878) $d years ago it was 9/20/1878
----------------------------------------------------------

I didn't think that is what it's supposed to do.



> 
>> ....
> 
> I don't think the kernel has anything to do with this error, which is
> unrelated to all the stuff you describe above. So where's the
> backtrace from this sequencep error?

Oh, yeah. the toggle-debug-on-error helped find the problem.  It was a
diary-related thing in my .emacs.  Out of urgency I just commented them
all out for now, so I don't know yet which statement caused the error,
but it's gone now.


> 
> ....

Debugging calendar gremlin in emacs 22.0.50.1

[ken]

After working already dozens of times, "M-x diary" suddenly fails with
error message "Wrong type argument: sequencep, t".  The only changes
made were to add more entries in the diary file.  To see if this was the
cause of the problem, I trimmed the diary file to just one entry:

%%(diary-float 11 4 -1) <p>(American) Thanksgiving Day.</p>

but the behavior is the same.

Running "M-x load-file" on ..../diary-lib.el yeilds the same error.


How to figure this one out?


tia

-- 
One is not superior merely because one
sees the world as odious.
                -- Chateaubriand (1768-1848)