Org 8 Bug in interactive formula editing mode?

Org 8 Bug in interactive formula editing mode?

[Thorsten Grothe]

Dear list,

I downloaded and installed the latest stable orgmode in my local homedir, typed
make all in a shell and put the following in my .emacs

(package-initialize)
(add-to-list 'load-path "~/.emacs.d/src/org-8.2.7b/lisp")(require 'org-install)

org-version displays the correct org-mode, so everything is fine.

Now I created a table like this:

#+TBLNAME: sec-umsatzsteuer
#+ORGTBL: SEND sec-umsatzsteuer orgtbl-to-latex :skip 2 :splice 
|  Wert | UST/VST | gesamt |
|-------+---------+--------|
| 15000 |    2850 |  17850 |
| 80000 |   15200 |  95200 |
#+TBLFM: $2=round($1*19/100,2)::$3=$1+$2

when I try to edit table formulas interactive with C-c ' emacs jumps to an
empty line at the end of the buffer and not to the formula in the table.

Does anyone have an idea?


Regards
Th. Grothe

Re: Org 8 Bug in interactive formula editing mode?

[Thorsten Grothe]

Dear list,

no one an idea?

Thanks in advance :-)

Regards
Th. Grothe

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Thorsten Grothe <info@th-grothe.de> writes:

> Dear list,
>
> no one an idea?
>
> Thanks in advance :-)
>

Nope - I tried to reproduce it with latest (both master and maint) and
could not, so I suspect an error in your configuration.

Nick

Re: Org 8 Bug in interactive formula editing mode?

[Thorsten Grothe]

Nick,

> Nope - I tried to reproduce it with latest (both master and maint) and
> could not, so I suspect an error in your configuration.

I think I found the problem, it happened after I patched the file with your
patch from the thread above:

#####
diff --git a/lisp/org-table.el b/lisp/org-table.el
index d7ef615..864493e 100644
--- a/lisp/org-table.el
+++ b/lisp/org-table.el
@@ -447,8 +447,8 @@ available parameters."
   "Check if the table has a marking column.
 If yes remove the column and the special lines."
   (let ((special (if maybe-quoted
-		     "^[ \t]*| *\\\\?[\#!$*_^/ ] *|"
-		   "^[ \t]*| *[\#!$*_^/ ] *|"))
+		     "^[ \t]*| *\\\\?[\#!$*_^/] *|"
+		   "^[ \t]*| *[\#!$*_^/] *|"))
 	(ignore  (if maybe-quoted
 		     "^[ \t]*| *\\\\?[!$_^/] *|"
 		   "^[ \t]*| *[!$_^/] *|")))
####

I tried to recover the saved file table.el.orig and everything works fine, so it
is something with the patch, right? What did I wrong ??

Regards
Th. Grothe

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Thorsten Grothe <info@th-grothe.de> writes:

> Nick,
>
>> Nope - I tried to reproduce it with latest (both master and maint) and
>> could not, so I suspect an error in your configuration.
>
> I think I found the problem, it happened after I patched the file with your
> patch from the thread above:
>
> #####
> diff --git a/lisp/org-table.el b/lisp/org-table.el
> index d7ef615..864493e 100644
> --- a/lisp/org-table.el
> +++ b/lisp/org-table.el
> @@ -447,8 +447,8 @@ available parameters."
>    "Check if the table has a marking column.
>  If yes remove the column and the special lines."
>    (let ((special (if maybe-quoted
> -		     "^[ \t]*| *\\\\?[\#!$*_^/ ] *|"
> -		   "^[ \t]*| *[\#!$*_^/ ] *|"))
> +		     "^[ \t]*| *\\\\?[\#!$*_^/] *|"
> +		   "^[ \t]*| *[\#!$*_^/] *|"))
>  	(ignore  (if maybe-quoted
>  		     "^[ \t]*| *\\\\?[!$_^/] *|"
>  		   "^[ \t]*| *[!$_^/] *|")))
> ####
>
> I tried to recover the saved file table.el.orig and everything works fine, so it
> is something with the patch, right? What did I wrong ??
>

I'm running latest maint currently (which includeds the patch) and
I don't have that problem, so I'd say no, there's nothing wrong with
the patch (famous last words).

I don't understand what you did exactly: you installed a new version.
Does that version include the patch? Check org-table.el around line 439:
if the function just has a `lines' argument then it most likely includes
the patch. I hope you did not try to apply the patch on top of that.
If you did, go back, throw away the patch file (it's not useful any
longer), and reinstall.

If the version you installed does *not* have the patch, why would it
have *this* problem? It would exhibit the *different* problem
that you brought up earlier: deleting columns on export.

Nick

Re: Org 8 Bug in interactive formula editing mode?

[Thorsten Grothe]

Nick,

sorry for the confusion, you are right, I was not clear ...

> I don't understand what you did exactly: you installed a new version.
> Does that version include the patch? Check org-table.el around line 439:
> if the function just has a `lines' argument then it most likely includes
> the patch. I hope you did not try to apply the patch on top of that.
> If you did, go back, throw away the patch file (it's not useful any
> longer), and reinstall.

Ok I did the following and found something strange:

First I installed 8.2.7b.zip and installed it in my homedir (without) patch and
I found out the same error: Edit formulas C-c ' throws me to the end of another
line and I cannot see the formula anymore as I wrote above.

After that I installed org-8.2.6, an older version, in my homedir (without
patch) and everything works fine, emacs jumps from field to field in the
formula.

So I was wrong, it has nothing to do with the patch, there is something
different between version org-8.7b and version 8.6.0, which produces this
error I guess.

I will now try to install your version from latest maint, see what happens and
report!!


Regards
Th. Grothe

Re: Org 8 Bug in interactive formula editing mode?

[Thorsten Grothe]

Ok I installed the git-version, the patch is working and export to latex works
fine, but the thing with the formulas still resist, this does not happen with
org-8.2.6 (without patch). I don't know what do do now. Maybe it has something
to do with emacs and not with org?

Thanks again for your endless patience for people like me :-)

Regards
Th. Grothe

Re: Org 8 Bug in interactive formula editing mode?

[Thorsten Grothe]

New test: 

I set up a minimal .emcas only for orgmode like this:

####
(package-initialize)
(add-to-list 'load-path "~/.emacs.d/src/org-mode/lisp")
(require 'org)
(add-to-list 'auto-mode-alist '("\\.org\\'" . org-mode))
(add-hook 'LaTeX-mode-hook 'orgtbl-mode)
(add-hook 'message-mode-hook 'turn-on-orgtbl)
####

My .emacs.d folder contains only the downloaded org files from git, same error :-) 


Regards
Th. Grothe

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Thorsten Grothe <info@th-grothe.de> writes:

> New test: 
>
> I set up a minimal .emcas only for orgmode like this:
>
> ####
> (package-initialize)
> (add-to-list 'load-path "~/.emacs.d/src/org-mode/lisp")
> (require 'org)
> (add-to-list 'auto-mode-alist '("\\.org\\'" . org-mode))
> (add-hook 'LaTeX-mode-hook 'orgtbl-mode)
> (add-hook 'message-mode-hook 'turn-on-orgtbl)
> ####
>
> My .emacs.d folder contains only the downloaded org files from git, same error :-) 
>
>

Have you done a `make' and a `make install' (or if you don't compile
anything, at least a `make autoloads')?

Have you restarted your emacs?

I can't think of anything else right now: it works perfectly here.

Nick

Re: Org 8 Bug in interactive formula editing mode?

[Robert Klein]

[-- Attachment #1: Type: text/plain, Size: 1681 bytes --]

Hi,

I confirm at least some variant of this.

I start emacs using the command line "emacs -Q -l .e2"  with .e2 being:

#+begin_src elisp
  (setq load-path (cons "~/.emacs.d/org-mode/lisp" load-path))
  (add-to-list 'auto-mode-alist '("\\.\\(org\\|org_archive\\|txt\\)$" .
org-mode))
#+end_src

org-mode version is git maint as of this morning.

For my test file (attached) when I open the "first headline" and type
C-c ' in one of the first table rows, I get shown the lines up to and
including "Dan".  Now, as long, as I navigate in formulas concerning the
visible rows, the table is visible in the org files buffer window and
the fields used by the formula are highlighted.

Now, when I go down to the first formula whose referenced field(s) are
not visible, the cursor in the org file jumps to the end of the file
which is then displayed at the upper end of the tabletest.org buffer window.



In "sufficient" long org files the visible cursor jumps to the end of
the org file buffer as soon as you navigate to a line with a formula in
the *Edit Formulae* buffer window.

PS: I've seen this for at least two weeks ago, but forgot about it again.

Best regards
Robert




On 07/21/2014 09:36 PM, Thorsten Grothe wrote:
> New test: 
> 
> I set up a minimal .emcas only for orgmode like this:
> 
> ####
> (package-initialize)
> (add-to-list 'load-path "~/.emacs.d/src/org-mode/lisp")
> (require 'org)
> (add-to-list 'auto-mode-alist '("\\.org\\'" . org-mode))
> (add-hook 'LaTeX-mode-hook 'orgtbl-mode)
> (add-hook 'message-mode-hook 'turn-on-orgtbl)
> ####
> 
> My .emacs.d folder contains only the downloaded org files from git, same error :-) 
> 
> 
> Regards
> Th. Grothe
> 


[-- Attachment #2: tabletest.org --]
[-- Type: application/vnd.lotus-organizer, Size: 1976 bytes --]

Re: Org 8 Bug in interactive formula editing mode?

[Thorsten Grothe]

> Now, when I go down to the first formula whose referenced field(s) are
> not visible, the cursor in the org file jumps to the end of the file
> which is then displayed at the upper end of the tabletest.org buffer window.
> 
> 
> 
> In "sufficient" long org files the visible cursor jumps to the end of
> the org file buffer as soon as you navigate to a line with a formula in
> the *Edit Formulae* buffer window.

I'm glad to hear that I'm not the only one who has this problem. I tested the
file with the same result. Editing formulas works fine until line @4$4 see link
with uploaded screenshot 1:

http://www.zimagez.com/zimage/bildschirmfoto-22072014-100459.php


from @5$4 to the end emacs jumps to the end of buffer 
see link with uploaded screenshot 2:


http://www.zimagez.com/zimage/bildschirmfoto-22072014-100751.php

Regards
Th. Grothe

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Robert Klein <RoKlein@roklein.de> writes:

>
> I confirm at least some variant of this.
>
> I start emacs using the command line "emacs -Q -l .e2"  with .e2 being:
>
> #+begin_src elisp
>   (setq load-path (cons "~/.emacs.d/org-mode/lisp" load-path))
>   (add-to-list 'auto-mode-alist '("\\.\\(org\\|org_archive\\|txt\\)$" .
> org-mode))
> #+end_src
>
> org-mode version is git maint as of this morning.
>
> For my test file (attached) when I open the "first headline" and type
> C-c ' in one of the first table rows, I get shown the lines up to and
> including "Dan".  Now, as long, as I navigate in formulas concerning the
> visible rows, the table is visible in the org files buffer window and
> the fields used by the formula are highlighted.
>

I can't reproduce any of this.

When I open the "first headline" I get the first section ("Lorem ipsum ..."): no table shown
(partial or total).

When I open the second headline "headline", I get the table: all of
it. C-c ' works.

> Now, when I go down to the first formula whose referenced field(s) are
> not visible, the cursor in the org file jumps to the end of the file
> which is then displayed at the upper end of the tabletest.org buffer window.
>
>
>
> In "sufficient" long org files the visible cursor jumps to the end of
> the org file buffer as soon as you navigate to a line with a formula in
> the *Edit Formulae* buffer window.
>
> PS: I've seen this for at least two weeks ago, but forgot about it again.
>

This is from maint:

Org-mode version 8.2.7b (release_8.2.7b-2-g798733 @ /home/nick/src/emacs/org/org-mode/lisp/)

My suggestion would be to `make clean', `make autoloads' and try again:
maybe there is some cruft in your tree that causes problems?

Can anybody else reproduce this?

Nick

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Thorsten Grothe <info@th-grothe.de> writes:

>> Now, when I go down to the first formula whose referenced field(s) are
>> not visible, the cursor in the org file jumps to the end of the file
>> which is then displayed at the upper end of the tabletest.org buffer window.
>> 
>> 
>> 
>> In "sufficient" long org files the visible cursor jumps to the end of
>> the org file buffer as soon as you navigate to a line with a formula in
>> the *Edit Formulae* buffer window.
>
> I'm glad to hear that I'm not the only one who has this problem. I tested the
> file with the same result. Editing formulas works fine until line @4$4 see link
> with uploaded screenshot 1:
>
> http://www.zimagez.com/zimage/bildschirmfoto-22072014-100459.php
>
>
> from @5$4 to the end emacs jumps to the end of buffer 
> see link with uploaded screenshot 2:
>
>
> http://www.zimagez.com/zimage/bildschirmfoto-22072014-100751.php
>

OK, I can see the problem too:

    - make sure you have a sufficiently small frame (with Robert's
      test file, a frame of about 30 lines long should do it)
    - open the first headline
    - open the second headline with the table
    - navigate to somewhere inside the table
    - C-c '
    - navigate among the formulas
    - see cursor jump in the original buffer between a row of the table
      and the end of the buffer

If the frame is bigger than about 50 lines (so that after C-c ', it's
big enough to show the whole table, including the formula and the empty
line underneath it) then I don't see the problem.

Nick

Re: Org 8 Bug in interactive formula editing mode?

[Thorsten Grothe]

[-- Attachment #1: Type: text/plain, Size: 722 bytes --]

> If the frame is bigger than about 50 lines (so that after C-c ', it's
> big enough to show the whole table, including the formula and the empty
> line underneath it) then I don't see the problem.

Thank you very much for your instructions!

yes you are right for the org file. I tested Roberts example with your
instructions and after resizing the formula frame a bit it works but
unfortunately this does not work with my attached tex file. For the first table
it works but for the two other tables it does not work even not if I try to
resize the formula window, so that I can see the whole table and the formula
line, sorry about that :-) I throws me to the end of the buffer again, funny
thing :-)

Regards
Th. Grothe

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: EditSumError.tex --]
[-- Type: text/x-tex, Size: 2724 bytes --]

\documentclass[%
fontsize=11pt,
fleqn,
parskip=half*,
toc=flat,
%DIV=15
]{scrartcl}

% Schriften
\usepackage{kpfonts}		
%\usepackage{libertine}

% Optischer Randausgleich
%\usepackage{microtype}

% Datenbank
\usepackage[pointsonright,forcolorpaper,links,usexkv,
nospacetowork,useforms,allowrandomize,usecustomdesign,
cfg=datenbank,
% nosolutions
answerkey
]{eqexam}
\usepackage{atbdb}
%\vspacewithkeyOn

% Titel
\subject[Rewe]{Rechnungswesen}
\title[USt/VSt]{Umsatzsteuer}
\keywords{}
\author{Th. Grothe}

\begin{document}
\maketitle
\vspace{2cm}
\tableofcontents
\begin{exam}{ReweUS}{\large\bfseries\textcolor{blue}{Umsatzsteuer}}

\section{Umsatzsteuer}
\ID*{ATB_ee538a82_edb3_4526_a4b6_bb16d9ce3f1a}
\begin{problem*}[\auto]
    \begin{parts}
      \item\PTs{6} \ldots die folgenden Geschäftsfälle
        \begin{orgtab}
            \label{sec-umsatzsteuer}	
            #+TBLNAME: sec-umsatzsteuer
            #+ORGTBL: SEND sec-umsatzsteuer orgtbl-to-latex :skip 2 :splice t :lend "\\\\\\SThiderow"
            |  Wert | UST/VST | gesamt |
            |-------+---------+--------|
            | 10000 |    1900 |  11900 |
            | 20000 |    3800 |  23800 |
            #+TBLFM: $2=round($1*19/100,2)::$3=$1+$2
            $
            
        \end{orgtab}
        \begin{solution}\\
            \begin{orgtab}
            	\label{sec-umsatzsteuer-3}        
            	#+TBLNAME: sec-umsatzsteuer-3
                #+ORGTBL: SEND sec-umsatzsteuer-3 orgtbl-to-latex :skip 2 :splice t 
                |    | Kto           |  Soll | Haben |
                |----+---------------+-------+-------|
                |    | AW. Rohstoffe | 10000 |       |
                |    | Vorsteuer     |  1900 |       |
                | an | Vall          |       | 11900 |
                #+TBLFM: @2$3=remote(sec-umsatzsteuer,@2$1)::@3$3=remote(sec-umsatzsteuer,@2$2)::@4$4=remote(sec-umsatzsteuer,@2$3)

                \label{sec-umsatzsteuer-5}        
                #+TBLNAME: sec-umsatzsteuer-5
                #+ORGTBL: SEND sec-umsatzsteuer-5 orgtbl-to-latex :skip 2 :splice t :hline "\\cmidrule(lr){2-2}"
                |    | Kto        |  Soll | Haben |
                |----+------------+-------+-------|
                |    | Bank       | 23800 |       |
                | an | Umsatzerl. |       | 20000 |
                | an | Umsatzst.  |       |  3800 |
                #+TBLFM: @2$3=remote(sec-umsatzsteuer,@3$3)::@3$4=remote(sec-umsatzsteuer,@3$1)::@4$4=remote(sec-umsatzsteuer,@3$2)
                
            \end{orgtab}
        \end{solution}
    \end{parts}
\end{problem*}
\endID

\end{exam}
\end{document}

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Thorsten Grothe <info@th-grothe.de> writes:

>> If the frame is bigger than about 50 lines (so that after C-c ', it's
>> big enough to show the whole table, including the formula and the empty
>> line underneath it) then I don't see the problem.
>
> Thank you very much for your instructions!
>
> yes you are right for the org file. I tested Roberts example with your
> instructions and after resizing the formula frame a bit it works but
> unfortunately this does not work with my attached tex file. For the
> first table it works but for the two other tables it does not work
> even not if I try to resize the formula window, so that I can see the
> whole table and the formula line, sorry about that :-) I throws me to
> the end of the buffer again, funny thing :-)
>

I didn't mean that as a workaround. I meant it as a partial
characterization of the problem that might give somebody in the know an
idea of where to go looking for it.

I believe that you'll have to live with the problem for a while: it
might be easy to fix or it might be difficult but until somebody takes a
deeper look, we won't know for sure.

But IIUC, the problem is a display problem only: you don't see the table
while you are editing the formulas, but once you finish and C-c C-c out
of the formula editing buffer, everything should be updated properly
(well, maybe the table needs to be recalculated - not sure whether that
happens automatically or not - but the formulas *should* be updated to
reflect your editing). Is my understanding correct?

Nick

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Nick Dokos <ndokos@gmail.com> writes:

> Thorsten Grothe <info@th-grothe.de> writes:
>
>>> If the frame is bigger than about 50 lines (so that after C-c ', it's
>>> big enough to show the whole table, including the formula and the empty
>>> line underneath it) then I don't see the problem.
>>
>> Thank you very much for your instructions!
>>
>> yes you are right for the org file. I tested Roberts example with your
>> instructions and after resizing the formula frame a bit it works but
>> unfortunately this does not work with my attached tex file. For the
>> first table it works but for the two other tables it does not work
>> even not if I try to resize the formula window, so that I can see the
>> whole table and the formula line, sorry about that :-) I throws me to
>> the end of the buffer again, funny thing :-)
>>

BTW, I played with your file[fn:1] a bit and I can get around the
problem by resizing the windows so that the file window is more than 100
lines and the formula editing window is about a dozen lines (I can get
about 132 lines on a monitor rotated into portrait mode - even more by
reducing the font size, but my eyes start complaining at around that
point).

So I'm pretty sure that the characerization stands: if the frame
(and the file window) is big enough then the problem is not visible.

Nick

Re: Org 8 Bug in interactive formula editing mode?

[Robert Klein]

I made some quick tests with different org versions; the first one I see
this issue is git commit 30220ffcdcef45c0237f80a2347d4da19877f64a
(release_8.2.6-77-g30220f).

I don't see the patch causing this, yet; have to look further.

Best regards
Robert


On 07/22/2014 09:01 PM, Nick Dokos wrote:
> Nick Dokos <ndokos@gmail.com> writes:
> 
>> Thorsten Grothe <info@th-grothe.de> writes:
>>
>>>> If the frame is bigger than about 50 lines (so that after C-c ', it's
>>>> big enough to show the whole table, including the formula and the empty
>>>> line underneath it) then I don't see the problem.
>>>
>>> Thank you very much for your instructions!
>>>
>>> yes you are right for the org file. I tested Roberts example with your
>>> instructions and after resizing the formula frame a bit it works but
>>> unfortunately this does not work with my attached tex file. For the
>>> first table it works but for the two other tables it does not work
>>> even not if I try to resize the formula window, so that I can see the
>>> whole table and the formula line, sorry about that :-) I throws me to
>>> the end of the buffer again, funny thing :-)
>>>
> 
> BTW, I played with your file[fn:1] a bit and I can get around the
> problem by resizing the windows so that the file window is more than 100
> lines and the formula editing window is about a dozen lines (I can get
> about 132 lines on a monitor rotated into portrait mode - even more by
> reducing the font size, but my eyes start complaining at around that
> point).
> 
> So I'm pretty sure that the characerization stands: if the frame
> (and the file window) is big enough then the problem is not visible.
> 
> Nick
> 
> 
> 

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Robert Klein <roklein@roklein.de> writes:

> I made some quick tests with different org versions; the first one I see
> this issue is git commit 30220ffcdcef45c0237f80a2347d4da19877f64a
> (release_8.2.6-77-g30220f).
>
> I don't see the patch causing this, yet; have to look further.
>

Absolutely right, that looks like the culprit:

,----
| commit 30220ffcdcef45c0237f80a2347d4da19877f64a
| Author: Bastien Guerry <bzg@altern.org>
| Date:   Thu Jun 12 12:04:24 2014 +0200
| 
|     Prefer `set-window-start' over `recenter' in some places.
|     
|     * org-agenda.el (org-agenda-show-1):
|     * org-table.el (org-table-show-reference):   <<<<<<<<<<<<<< this change
|     * org.el (org-fix-ellipsis-at-bol)
|     (org-first-headline-recenter): Use `set-window-start' instead
|     of `recenter'.
|     
|     * org-agenda.el (org-recenter-heading): Delete.
`----

-- 
Nick

Re: Org 8 Bug in interactive formula editing mode?

[Luke Crook]

Nick Dokos <ndokos <at> gmail.com> writes:
> Nope - I tried to reproduce it with latest (both master and maint) and
> could not, so I suspect an error in your configuration.
> 

I experience the same issue.

http://article.gmane.org/gmane.emacs.orgmode/88610


/Luke

Re: Org 8 Bug in interactive formula editing mode?

[Luke Crook]

Nick Dokos <ndokos <at>zgmail.com> writes:
> 
> Can anybody else reproduce this?

/me raises hand.

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Nick Dokos <ndokos@gmail.com> writes:

> Robert Klein <roklein@roklein.de> writes:
>
>> I made some quick tests with different org versions; the first one I see
>> this issue is git commit 30220ffcdcef45c0237f80a2347d4da19877f64a
>> (release_8.2.6-77-g30220f).
>>
>> I don't see the patch causing this, yet; have to look further.
>>
>
> Absolutely right, that looks like the culprit:
>
> ,----
> | commit 30220ffcdcef45c0237f80a2347d4da19877f64a
> | Author: Bastien Guerry <bzg@altern.org>
> | Date:   Thu Jun 12 12:04:24 2014 +0200
> | 
> |     Prefer `set-window-start' over `recenter' in some places.
> |     
> |     * org-agenda.el (org-agenda-show-1):
> |     * org-table.el (org-table-show-reference):   <<<<<<<<<<<<<< this change
> |     * org.el (org-fix-ellipsis-at-bol)
> |     (org-first-headline-recenter): Use `set-window-start' instead
> |     of `recenter'.
> |     
> |     * org-agenda.el (org-recenter-heading): Delete.
> `----

The diff was

,----
| diff --git a/lisp/org-table.el b/lisp/org-table.el
| index 1d6bee0..6d649ab 100644
| --- a/lisp/org-table.el
| +++ b/lisp/org-table.el
| @@ -3863,9 +3863,10 @@ With prefix ARG, apply the new formulas to the table."
|  	(push org-table-current-begin-pos org-show-positions)
|  	(let ((min (apply 'min org-show-positions))
|  	      (max (apply 'max org-show-positions)))
| -	  (goto-char min) (recenter 0)
| +	  (set-window-start (selected-window) (point-min))
|  	  (goto-char max)
| -	  (or (pos-visible-in-window-p max) (recenter -1))))
| +	  (or (pos-visible-in-window-p max)
| +	      (set-window-start (selected-window) (point-max)))))
|        (select-window win))))
|  
|  (defun org-table-force-dataline ()
`----

I suspect that instead of (point-min), it should be just min
and instead of point-max, it should be just max (or possibly
the beginning of the next line if it exists).

Nick

Re: Org 8 Bug in interactive formula editing mode?

[Thorsten Grothe]

> I believe that you'll have to live with the problem for a while: it
> might be easy to fix or it might be difficult but until somebody takes a
> deeper look, we won't know for sure.

Of course I can live with that, we are on open source here and here are many
people like you doing a great big job!! It was only important for me to report
this strange behavior :-)

Regards
Th. Grothe

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Can you please try this patch out and report whether it
works or not and whether it causes any other problems?
I'm not sure if Bastien is around to vet it, but if it
seems to work for you, I can push it and deal with any
other breakage later.


--8<---------------cut here---------------start------------->8---
diff --git a/lisp/org-table.el b/lisp/org-table.el
index bfe396e..f19c027 100644
--- a/lisp/org-table.el
+++ b/lisp/org-table.el
@@ -3859,10 +3859,10 @@ With prefix ARG, apply the new formulas to the table."
 	(push org-table-current-begin-pos org-show-positions)
 	(let ((min (apply 'min org-show-positions))
 	      (max (apply 'max org-show-positions)))
-	  (set-window-start (selected-window) (point-min))
+	  (set-window-start (selected-window) min)
 	  (goto-char max)
 	  (or (pos-visible-in-window-p max)
-	      (set-window-start (selected-window) (point-max)))))
+	      (set-window-start (selected-window) max))))
       (select-window win))))
 
 (defun org-table-force-dataline ()
--8<---------------cut here---------------end--------------->8---

Thanks,
Nick

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Thorsten Grothe <info@th-grothe.de> writes:

>> I believe that you'll have to live with the problem for a while: it
>> might be easy to fix or it might be difficult but until somebody takes a
>> deeper look, we won't know for sure.
>
> Of course I can live with that, we are on open source here and here are many
> people like you doing a great big job!! It was only important for me to report
> this strange behavior :-)
>

For which, many thanks (particularly for insisting: Luke Crook reported
it a few days ago, but afaict, he didn't get any response - I don't even
remember seeing his post).

But as you must have seen elsewhere in the thread, Robert Klein
identified the culprit and I posted a patch that, just possibly, solves
this problem. Let me know what you find out.

Thanks,
Nick

Re: Org 8 Bug in interactive formula editing mode?

[Thorsten Grothe]

> Can you please try this patch out and report whether it
> works or not and whether it causes any other problems?
> I'm not sure if Bastien is around to vet it, but if it
> seems to work for you, I can push it and deal with any
> other breakage later.

Seems to work here, the error is gone, very nice !!!

Regards
Th. Grothe

Re: Org 8 Bug in interactive formula editing mode?

[Nick Dokos]

Thorsten Grothe <info@th-grothe.de> writes:

>> Can you please try this patch out and report whether it
>> works or not and whether it causes any other problems?
>> I'm not sure if Bastien is around to vet it, but if it
>> seems to work for you, I can push it and deal with any
>> other breakage later.
>
> Seems to work here, the error is gone, very nice !!!
>

OK, I pushed the patch to maint and merged it into master.

Thanks to Robert, Thorsten and Luke for all the help!
-- 
Nick

Re: Org 8 Bug in interactive formula editing mode?

[Bastien]

Hi Nick,

Nick Dokos <ndokos@gmail.com> writes:

> OK, I pushed the patch to maint and merged it into master.

Thanks a lot for fixing this too!

-- 
 Bastien