table cell menu does not appear at the menubar

table cell menu does not appear at the menubar

[Tak Ota]

(easy-menu-define table-cell-menu-map table-cell-map "Table cell menu" table-cell-menu)

in lisp/textmodes/table.el used to put a table cell menu item on the
menubar and it appeared only when the point entered a table cell
area.  This is somehow not working anymore.  Does anyone have an idea
why?

-Tak

Re: table cell menu does not appear at the menubar

[Stephen Berman]

On Wed, 02 May 2007 17:22:14 -0700 (PDT) Tak Ota
<Takaaki.Ota@am.sony.com> wrote:

> in lisp/textmodes/table.el used to put a table cell menu item on the
> menubar and it appeared only when the point entered a table cell
> area.  This is somehow not working anymore.  Does anyone have an idea
> why?

It works for me in GNU Emacs 22.1.50.1 (i686-pc-linux-gnu, GTK+
Version 2.10.6) of 2007-04-30.

Steve Berman

Re: table cell menu does not appear at the menubar

[Richard Stallman]

    (easy-menu-define table-cell-menu-map table-cell-map "Table cell menu" table-cell-menu)

    in lisp/textmodes/table.el used to put a table cell menu item on the
    menubar and it appeared only when the point entered a table cell
    area.  This is somehow not working anymore.  Does anyone have an idea
    why?

Can you try building Emacs sources from various dates, and determine
which change broke this?

Re: table cell menu does not appear at the menubar

[Tak Ota]

Thu, 03 May 2007 19:56:14 -0400: Richard Stallman <rms@gnu.org> wrote:

>     (easy-menu-define table-cell-menu-map table-cell-map "Table cell menu" table-cell-menu)
> 
>     in lisp/textmodes/table.el used to put a table cell menu item on the
>     menubar and it appeared only when the point entered a table cell
>     area.  This is somehow not working anymore.  Does anyone have an idea
>     why?
> 
> Can you try building Emacs sources from various dates, and determine
> which change broke this?

I am sorry but I cannot perform this test right now as I am overloaded
with daily work.  I am happy to investigate more when I can spare
time, hopefully soon.

Also the nature of this bug is a bit more elusive than I thought.  If
I kill current emacs and start it freshly the menu starts working
correctly.  It continue to work correctly but eventually stops
(entering the point into a table cell doesn't bring the table menu
item anymore).  I don't exactly know what "eventually" means
technically at this moment.

-Tak

Re: table cell menu does not appear at the menubar

[Richard Stallman]

    Also the nature of this bug is a bit more elusive than I thought.  If
    I kill current emacs and start it freshly the menu starts working
    correctly.  It continue to work correctly but eventually stops
    (entering the point into a table cell doesn't bring the table menu
    item anymore).  I don't exactly know what "eventually" means
    technically at this moment.

If you use GDB to investigate what happens in updating the menu when
it is failing, and simultaneously investigate another new Emacs
process in which this is still working correctly, you can compare
them and see where and why they start to diverge.

I think we should not delay the release to fix this.

Re: table cell menu does not appear at the menubar

[Tak Ota]

I haven't had chance to perform the test under GDB.  I have a few
additional information to this problem.  The problem seems to have
nothing to do with table.el.  It is a GUI problem.

1. The problem may be the build platform specific as it appears to be
   related to the windowing system.  I am building my emacs on Windows
   XP using mingw.

2. When the problem once starts menubar contents do not change
   regardless of the mode where point is at.  The last menubar content
   is stuck.  The last remaining menu responds to mouse event and
   functions correctly.

3. When this problem starts the mouse click on a window does not
   change the focus.  The point, the highlighted window status bar all
   remains the same regardless of mouse click.

-Tak

Sun, 06 May 2007 18:26:43 -0400: Richard Stallman <rms@gnu.org> wrote:

>     Also the nature of this bug is a bit more elusive than I thought.  If
>     I kill current emacs and start it freshly the menu starts working
>     correctly.  It continue to work correctly but eventually stops
>     (entering the point into a table cell doesn't bring the table menu
>     item anymore).  I don't exactly know what "eventually" means
>     technically at this moment.
> 
> If you use GDB to investigate what happens in updating the menu when
> it is failing, and simultaneously investigate another new Emacs
> process in which this is still working correctly, you can compare
> them and see where and why they start to diverge.
> 
> I think we should not delay the release to fix this.
> 

Re: table cell menu does not appear at the menubar

[Lennart Borgman (gmail)]

I have seen some problems with the menus on w32. I am however very 
unsure of what goes wrong. I myself use a slightly patched version of 
Emacs where I use Alt to access the menus from the keyboard. (I use the 
windows keys for Emacs META instead of the Alt key that is the default 
for Emacs META on w32.)

Lately I have began to see some problems I have never seen before. Those 
seems to be related to that I use StickyKeys. There is a bug in XP which 
makes it eat GDI Objects when StickyKeys are on. (It is so irritating 
that if it was easier to file a bug to MS I might even take my own time 
to do that ;-) ) When those are getting low some strange things can happen.

There might be other problems too. I did some changes to the menu 
handling too, since I discovered that trying to use the menus from the 
keyboard might result in changing the buffer instead. As I see it there 
has not been much interest in incorporating my changes in Emacs though 
or investigating the matter more closely. (I do not think my changes to 
the menu code are entirely correct, but at least they work much better 
for me than the standard CVS Emacs.)



Tak Ota wrote:
> I haven't had chance to perform the test under GDB.  I have a few
> additional information to this problem.  The problem seems to have
> nothing to do with table.el.  It is a GUI problem.
> 
> 1. The problem may be the build platform specific as it appears to be
>    related to the windowing system.  I am building my emacs on Windows
>    XP using mingw.
> 
> 2. When the problem once starts menubar contents do not change
>    regardless of the mode where point is at.  The last menubar content
>    is stuck.  The last remaining menu responds to mouse event and
>    functions correctly.
> 
> 3. When this problem starts the mouse click on a window does not
>    change the focus.  The point, the highlighted window status bar all
>    remains the same regardless of mouse click.
> 
> -Tak
> 
> Sun, 06 May 2007 18:26:43 -0400: Richard Stallman <rms@gnu.org> wrote:
> 
>>     Also the nature of this bug is a bit more elusive than I thought.  If
>>     I kill current emacs and start it freshly the menu starts working
>>     correctly.  It continue to work correctly but eventually stops
>>     (entering the point into a table cell doesn't bring the table menu
>>     item anymore).  I don't exactly know what "eventually" means
>>     technically at this moment.
>>
>> If you use GDB to investigate what happens in updating the menu when
>> it is failing, and simultaneously investigate another new Emacs
>> process in which this is still working correctly, you can compare
>> them and see where and why they start to diverge.
>>
>> I think we should not delay the release to fix this.
>>
> 
> 
> 
> _______________________________________________
> Emacs-devel mailing list
> Emacs-devel@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-devel
> 

Re: table cell menu does not appear at the menubar

[Eli Zaretskii]

> Date: Fri, 18 May 2007 00:35:26 +0200
> From: "Lennart Borgman (gmail)" <lennart.borgman@gmail.com>
> Cc: emacs-devel@gnu.org
> 
> I have seen some problems with the menus on w32. I am however very 
> unsure of what goes wrong. I myself use a slightly patched version of 
> Emacs where I use Alt to access the menus from the keyboard.

You aren't saying that you needed to patch Emacs to use Alt for menu
access, are you?  Because I'm quite sure that works in stock Emacs
22.1.

Re: table cell menu does not appear at the menubar

[Lennart Borgman (gmail)]

Eli Zaretskii wrote:
>> Date: Fri, 18 May 2007 00:35:26 +0200
>> From: "Lennart Borgman (gmail)" <lennart.borgman@gmail.com>
>> Cc: emacs-devel@gnu.org
>>
>> I have seen some problems with the menus on w32. I am however very 
>> unsure of what goes wrong. I myself use a slightly patched version of 
>> Emacs where I use Alt to access the menus from the keyboard.
> 
> You aren't saying that you needed to patch Emacs to use Alt for menu
> access, are you?  Because I'm quite sure that works in stock Emacs
> 22.1.

Yes, that is exactly what I say. And I believe I have reported that 
error long ago.

Re: table cell menu does not appear at the menubar

[Eli Zaretskii]

> X-Spam-Status: No, score=0.0 required=5.0 tests=none autolearn=failed 
> 	version=3.1.0
> Date: Fri, 18 May 2007 16:55:42 +0200
> From: "Lennart Borgman (gmail)" <lennart.borgman@gmail.com>
> CC:  Takaaki.Ota@am.sony.com,  emacs-devel@gnu.org
> 
> Eli Zaretskii wrote:
> >> Date: Fri, 18 May 2007 00:35:26 +0200
> >> From: "Lennart Borgman (gmail)" <lennart.borgman@gmail.com>
> >> Cc: emacs-devel@gnu.org
> >>
> >> I have seen some problems with the menus on w32. I am however very 
> >> unsure of what goes wrong. I myself use a slightly patched version of 
> >> Emacs where I use Alt to access the menus from the keyboard.
> > 
> > You aren't saying that you needed to patch Emacs to use Alt for menu
> > access, are you?  Because I'm quite sure that works in stock Emacs
> > 22.1.
> 
> Yes, that is exactly what I say. And I believe I have reported that 
> error long ago.
> 

Doesn't w32-pass-alt-to-system work for you?  It seems to work for me.

Re: table cell menu does not appear at the menubar

[Lennart Borgman (gmail)]

Eli Zaretskii wrote:
>> X-Spam-Status: No, score=0.0 required=5.0 tests=none autolearn=failed 
>> 	version=3.1.0
>> Date: Fri, 18 May 2007 16:55:42 +0200
>> From: "Lennart Borgman (gmail)" <lennart.borgman@gmail.com>
>> CC:  Takaaki.Ota@am.sony.com,  emacs-devel@gnu.org
>>
>> Eli Zaretskii wrote:
>>>> Date: Fri, 18 May 2007 00:35:26 +0200
>>>> From: "Lennart Borgman (gmail)" <lennart.borgman@gmail.com>
>>>> Cc: emacs-devel@gnu.org
>>>>
>>>> I have seen some problems with the menus on w32. I am however very 
>>>> unsure of what goes wrong. I myself use a slightly patched version of 
>>>> Emacs where I use Alt to access the menus from the keyboard.
>>> You aren't saying that you needed to patch Emacs to use Alt for menu
>>> access, are you?  Because I'm quite sure that works in stock Emacs
>>> 22.1.
>> Yes, that is exactly what I say. And I believe I have reported that 
>> error long ago.
>>
> 
> Doesn't w32-pass-alt-to-system work for you?  It seems to work for me.

I am not sure I can reproduce the problem any more, I have to go on some 
keyboard training for a while.

I think I told about this long ago. It works most of the time, but not 
always. On a fast pc you may not be able to see the problem. However 
earlier when I had a slow pc it happened very often that instead of 
opening the menus Emacs sent the key I typed after Alt to the keymaps in 
use.

I do not remember now if it was necessary to have StickyKeys enabled or 
not, but I always have that.

AFAICS it was a fault in the logic for the menu access.

It may have been fixed since I wrote my patch, but I have not seen 
anything that makes me believe that.

Maybe I should also say again that I am not confident that my patch is 
entirely correct. It required some queing of the input and the logic 
involved there in those parts of Emacs is not very easy. (And at that 
time with the slow machine and my inexperience with Emacs it was very 
time consuming to try to find the different parts of it.)

I do not believe I have had problem directly related to it, but I am 
unsure because of the GDI objects bug in XP.

Re: table cell menu does not appear at the menubar

[Jason Rumney]

Eli Zaretskii wrote:
>> Date: Fri, 18 May 2007 00:35:26 +0200
>> From: "Lennart Borgman (gmail)" <lennart.borgman@gmail.com>
>> Cc: emacs-devel@gnu.org
>>
>> I have seen some problems with the menus on w32. I am however very 
>> unsure of what goes wrong. I myself use a slightly patched version of 
>> Emacs where I use Alt to access the menus from the keyboard.
>>     
>
> You aren't saying that you needed to patch Emacs to use Alt for menu
> access, are you?  Because I'm quite sure that works in stock Emacs
> 22.1.
>   

I think that what requires Lennart's patch is not the use of the Alt key
for menus, but the use of the Flag key for meta to free up the Alt key
for its alternate use.

Re: table cell menu does not appear at the menubar

[Lennart Borgman (gmail)]

Jason Rumney wrote:
> Eli Zaretskii wrote:
>>> Date: Fri, 18 May 2007 00:35:26 +0200
>>> From: "Lennart Borgman (gmail)" <lennart.borgman@gmail.com>
>>> Cc: emacs-devel@gnu.org
>>>
>>> I have seen some problems with the menus on w32. I am however very 
>>> unsure of what goes wrong. I myself use a slightly patched version of 
>>> Emacs where I use Alt to access the menus from the keyboard.
>>>     
>> You aren't saying that you needed to patch Emacs to use Alt for menu
>> access, are you?  Because I'm quite sure that works in stock Emacs
>> 22.1.
>>   
> 
> I think that what requires Lennart's patch is not the use of the Alt key
> for menus, but the use of the Flag key for meta to free up the Alt key
> for its alternate use.

It is two different patches.

Re: table cell menu does not appear at the menubar

[Richard Stallman]

    2. When the problem once starts menubar contents do not change
       regardless of the mode where point is at.  The last menubar content
       is stuck.  The last remaining menu responds to mouse event and
       functions correctly.

menu_bar_items is the function that is supposed to make a list of
all the keymaps to be processed.  So the problem is either there
or in or within a function that calls it.

When the problem happens, can you see if and when menu_bar_items
gets called?  And when it calls get_local_map, what args does it use
and what value does it get?

Re: table cell menu does not appear at the menubar

[Tak Ota]

Since I freshly got the completely source from CVS Friday (05/18/07) I
have built it and been running it under GDB.  So far I haven't seen
the problem yet.  Has anyone changed something that might have fixed
this stuck menu problem?

-Tak

Fri, 18 May 2007 19:08:57 -0400: Richard Stallman <rms@gnu.org> wrote:

>     2. When the problem once starts menubar contents do not change
>        regardless of the mode where point is at.  The last menubar content
>        is stuck.  The last remaining menu responds to mouse event and
>        functions correctly.
> 
> menu_bar_items is the function that is supposed to make a list of
> all the keymaps to be processed.  So the problem is either there
> or in or within a function that calls it.
> 
> When the problem happens, can you see if and when menu_bar_items
> gets called?  And when it calls get_local_map, what args does it use
> and what value does it get?
> 

Re: table cell menu does not appear at the menubar

[Lennart Borgman (gmail)]

I have no idea about that, but the problem I mentioned earlier with 
menus on w32 is a race  problem/sync problem (or whatever the correct 
term is).

It might be difficult to catch such problems with the debugger of 
course. I inserted trace output to the console to catch it.


Tak Ota wrote:
> Since I freshly got the completely source from CVS Friday (05/18/07) I
> have built it and been running it under GDB.  So far I haven't seen
> the problem yet.  Has anyone changed something that might have fixed
> this stuck menu problem?
> 
> -Tak
> 
> Fri, 18 May 2007 19:08:57 -0400: Richard Stallman <rms@gnu.org> wrote:
> 
>>     2. When the problem once starts menubar contents do not change
>>        regardless of the mode where point is at.  The last menubar content
>>        is stuck.  The last remaining menu responds to mouse event and
>>        functions correctly.
>>
>> menu_bar_items is the function that is supposed to make a list of
>> all the keymaps to be processed.  So the problem is either there
>> or in or within a function that calls it.
>>
>> When the problem happens, can you see if and when menu_bar_items
>> gets called?  And when it calls get_local_map, what args does it use
>> and what value does it get?

Re: table cell menu does not appear at the menubar

[Tak Ota]

Thanks for the tip.  If I don't see the problem under debugger for
another few days I'll start investigating in other ways.

-Tak

Tue, 22 May 2007 00:36:24 +0200: "Lennart Borgman (gmail)" <lennart.borgman@gmail.com> wrote:

> I have no idea about that, but the problem I mentioned earlier with 
> menus on w32 is a race  problem/sync problem (or whatever the correct 
> term is).
> 
> It might be difficult to catch such problems with the debugger of 
> course. I inserted trace output to the console to catch it.
> 
> 
> Tak Ota wrote:
> > Since I freshly got the completely source from CVS Friday (05/18/07) I
> > have built it and been running it under GDB.  So far I haven't seen
> > the problem yet.  Has anyone changed something that might have fixed
> > this stuck menu problem?
> > 
> > -Tak
> > 
> > Fri, 18 May 2007 19:08:57 -0400: Richard Stallman <rms@gnu.org> wrote:
> > 
> >>     2. When the problem once starts menubar contents do not change
> >>        regardless of the mode where point is at.  The last menubar content
> >>        is stuck.  The last remaining menu responds to mouse event and
> >>        functions correctly.
> >>
> >> menu_bar_items is the function that is supposed to make a list of
> >> all the keymaps to be processed.  So the problem is either there
> >> or in or within a function that calls it.
> >>
> >> When the problem happens, can you see if and when menu_bar_items
> >> gets called?  And when it calls get_local_map, what args does it use
> >> and what value does it get?
> 
> 

Re: table cell menu does not appear at the menubar

[Richard Stallman]

    Since I freshly got the completely source from CVS Friday (05/18/07) I
    have built it and been running it under GDB.  So far I haven't seen
    the problem yet.  Has anyone changed something that might have fixed
    this stuck menu problem?

How often does it usually happen?  Did you often go 4 days without
seeing it?

Re: table cell menu does not appear at the menubar

[Tak Ota]

It is quite random.  It happened in a few hours.  It happened after a
few days.  I don't remember if it happened after 4 days.  I'll keep
collecting more data.

-Tak

Tue, 22 May 2007 10:52:27 -0400: Richard Stallman <rms@gnu.org> wrote:

>     Since I freshly got the completely source from CVS Friday (05/18/07) I
>     have built it and been running it under GDB.  So far I haven't seen
>     the problem yet.  Has anyone changed something that might have fixed
>     this stuck menu problem?
> 
> How often does it usually happen?  Did you often go 4 days without
> seeing it?
> 

Re: table cell menu does not appear at the menubar

[martin rudalics]

[-- Attachment #1: Type: text/plain, Size: 1240 bytes --]

 > It is quite random.  It happened in a few hours.  It happened after a
 > few days.  I don't remember if it happened after 4 days.  I'll keep
 > collecting more data.

I didn't verify this but considering the recent thread on
`sgml-mode-point-entered' I believe something similar might occur here.
Look at the following two functions:

(defun table--point-entered-cell-function (&optional old-point new-point)
   "Point has entered a cell.
Refresh the menu bar."
   (unless table-cell-entered-state
     (setq table-cell-entered-state t)
     (setq table-mode-indicator t)
     (force-mode-line-update)
     (table--warn-incompatibility)
     (run-hooks 'table-point-entered-cell-hook)))

(defun table--point-left-cell-function (&optional old-point new-point)
   "Point has left a cell.
Refresh the menu bar."
   (when table-cell-entered-state
     (setq table-cell-entered-state nil)
     (setq table-mode-indicator nil)
     (force-mode-line-update)
     (run-hooks 'table-point-left-cell-hook)))

Both do a `force-mode-line-update' which might recursively toggle
`table-cell-entered-state' for preceding cells on the same line.  In
that case the bug should get caught by Yamamoto's fix.  I would
nevertheless apply the attached patch.

[-- Attachment #2: table.patch --]
[-- Type: text/plain, Size: 2332 bytes --]

*** table.el	Tue Jan 23 06:41:12 2007
--- table.el	Mon May 28 14:33:00 2007
***************
*** 6,12 ****
  ;; Keywords: wp, convenience
  ;; Author: Takaaki Ota <Takaaki.Ota@am.sony.com>
  ;; Created: Sat Jul 08 2000 13:28:45 (PST)
! ;; Revised: Wed Jan 03 2007 13:23:46 (PST)

  ;; This file is part of GNU Emacs.

--- 6,12 ----
  ;; Keywords: wp, convenience
  ;; Author: Takaaki Ota <Takaaki.Ota@am.sony.com>
  ;; Created: Sat Jul 08 2000 13:28:45 (PST)
! ;; Revised: Mon Mai 28 2007 14:33:01 ((MEZ) - Mitteleurop. Sommerzeit)

  ;; This file is part of GNU Emacs.

***************
*** 5333,5353 ****
  (defun table--point-entered-cell-function (&optional old-point new-point)
    "Point has entered a cell.
  Refresh the menu bar."
!   (unless table-cell-entered-state
!     (setq table-cell-entered-state t)
!     (setq table-mode-indicator t)
!     (force-mode-line-update)
!     (table--warn-incompatibility)
!     (run-hooks 'table-point-entered-cell-hook)))

  (defun table--point-left-cell-function (&optional old-point new-point)
    "Point has left a cell.
  Refresh the menu bar."
!   (when table-cell-entered-state
!     (setq table-cell-entered-state nil)
!     (setq table-mode-indicator nil)
!     (force-mode-line-update)
!     (run-hooks 'table-point-left-cell-hook)))

  (defun table--warn-incompatibility ()
    "If called from interactive operation warn the know incompatibilities.
--- 5333,5355 ----
  (defun table--point-entered-cell-function (&optional old-point new-point)
    "Point has entered a cell.
  Refresh the menu bar."
!   (let ((inhibit-point-motion-hooks t))
!     (unless table-cell-entered-state
!       (setq table-cell-entered-state t)
!       (setq table-mode-indicator t)
!       (force-mode-line-update)
!       (table--warn-incompatibility)
!       (run-hooks 'table-point-entered-cell-hook))))

  (defun table--point-left-cell-function (&optional old-point new-point)
    "Point has left a cell.
  Refresh the menu bar."
!   (let ((inhibit-point-motion-hooks t))
!     (when table-cell-entered-state
!       (setq table-cell-entered-state nil)
!       (setq table-mode-indicator nil)
!       (force-mode-line-update)
!       (run-hooks 'table-point-left-cell-hook))))

  (defun table--warn-incompatibility ()
    "If called from interactive operation warn the know incompatibilities.

[-- Attachment #3: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: table cell menu does not appear at the menubar

[Richard Stallman]

Your table.el fix seems safe, so please install it in Emacs 22.

Re: table cell menu does not appear at the menubar

[martin rudalics]

> Your table.el fix seems safe, so please install it in Emacs 22.

Installed in EMACS_22_BASE.