Re: recalculate tables

[Michael Welle <mwe012008@gmx.net>]

Hello,

Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:

> Hello,
>
> Michael Welle <mwe012008@gmx.net> writes:
>
>> I have a few of those, that worked in the past. So I think the behaviour
>> had changed some time in the past year or so.
>
> Quite possible. I made some invasive changes in "org-table.el" a while
> ago.
SCM revealed that I 'fixed' one table with an empty header around
Christmas last year. But I can't remember if I found the problem a few
weeks earlier or a few months.


>> I liked the old behaviour (but it's not super important for me)
>
> The problem is that I don't know what is "the old behaviour".  The fact
> that column formulas do not apply in table headers has been there for
> ages.
Yes, Russell suggested something like that. That puzzles me a bit. The
table, that I mentioned above is updated every three months. Even if the
formulas would not have been applied three or four times before I found
out about it, I wouldn't have used Org version 7.8 (the one that
Russell uses) or earlier back at that time, I think.

If the change is somewhere in, let's say, org-table.el, one thing I can
vaguely imagine is that an older org-table.el was lying around and got
loaded. I experienced something like that with ob-sh.el and ob-shell.el.
But on the other hand, I have trouble to understand how that could
happen (other than with ob-sh the file name stayed the same).


> What is that behaviour?
I have a handful of tables like the following:

|---+---+----|
| 3 | 4 |    |
|---+---+----|
#+TBLFM: $3=$2-$1

I don't add data to them very frequently. But since they all have the
third column filled, I think recalculating had worked in tables formated
like that before.

If you give me a week or so I will try to find the change set that
introduced the current behaviour.

Regards
hmw