Re: [PATCH] bad table formula recorded in some cases

[tbanelwebmin <tbanelwebmin@free.fr>]

Ok, Timothy, fair enough

Le 21/07/2021 à 17:07, Timothy a écrit :
> Hi Thierry,
>
> tbanelwebmin <tbanelwebmin@free.fr> writes:
>> I don't know the intention. But the answer may lie in the comment 4
>> lines above:
>>    ;; Don't overwrite TBLFM, we might use text properties to
>>    ;; store stuff.
>>
>> In this case, the intention would be to keep the original "#+TBLFM:"
>> instead of inserting a fresh new one.
>>
>> But we are in the else branch of (if (looking-at ...)), which means
>> there was no "#+TBLFM:". And no text properties to save. Therefore we
>> may safely remove this (match-string 2).
> Thank you for looking into this, I'm reassured by your inference that
> this change is safe to make. I'm not really one of the main
> contribution-acceptors/pushers though, so I'd rather leave this for
> someone like Nicolas to sign off on.
>
> Would you mind bumping this thread in a few weeks if nothing happens?
>
> Hope that's not too much of an inconvenience,
>
> Timothy.
>
>> Le 21/07/2021 à 12:50, Timothy a écrit :
>>> Hi Thierry,
>>>
>>> Thanks for this! Looking at the change you suggest, do you know why the
>>> (match-string 2) bit might have been added in the first place? I'm just
>>> wondering if there might be some edge-case adversely affected by this ---
>>> hence trading one bug for another :P
>>>
>>> --
>>> Timothy
>>>
>>> tbanelwebmin <tbanelwebmin@free.fr> writes:
>>>
>>>> Small bug, small fix.
>>>>
>>>> Suppose we have a table embedded in a begin-end block.
>>>>
>>>> #+begin: aaa :param value
>>>> | a | b |
>>>> | a | b |
>>>> #+end:
>>>>
>>>> Suppose we want to add a formula, with C-c =
>>>> We end up with an incorrect result:
>>>>
>>>> #+begin: aaa :param value
>>>> | a | 33 |
>>>> | a |  b |
>>>>  :param value $2=33
>>>> #+end:
>>>>
>>>> The fix: in org-table.el, line 2177, change
>>>>   (insert (or (match-string 2) "#+TBLFM:")))
>>>> to
>>>>   (insert "#+TBLFM:"))
>>>>
>>>> Then we get the correct result:
>>>>
>>>> #+begin: aaa :param value
>>>> | a | 33 |
>>>> | a |  b |
>>>> #+TBLFM: $2=33
>>>> #+end:
>>>>
>>>> Why? Because (match-string 2) is supposed to refer to the (looking-at)
>>>> instruction 7 lines above. But (match-string 2) is in the else branch,
>>>> which means that (looking-at) failed. Therefore (match-string 2) returns
>>>> garbage.
>>>>
>>>> Thanks to Uwe Brauer for pointing to this bug.