crash in new ledger

[Jack]

On 7/9/22 02:28, thb--- via KMyMoney-devel wrote:
> Jack via KMyMoney-devel schrieb am 08.07.2022 18:33 (GMT +02:00):
>
> On 2022.07.08 03:11, thb--- via KMyMoney-devel wrote:
>>> Jack via KMyMoney-devel schrieb am 07.07.2022 19:16 (GMT +02:00):
>>>
>>>> I got several of these over the past days, but didn't report yet, as
>>>> I wasn't sure I was on absolute latest.  I just recompiled from git
>>>> master head, and get this (consistently) when saving an investment
>>>> transaction, where I had changed the security.  Let me know if there
>>>> is any other information which might be helpful.
>>> Thanks for reporting. Despite all the details I have a problem to see
>>> where this happens. Mainly, because the backtrace does not (always)
>>> match my version of current master :(  It would be great, if you
>>> could simply specify the commit hash you are on in the future.
>>>
>>> I am on 462fb8d2df6 and investtransactioneditor.cpp:1212 shows this
>>> for me
>> I am also on this - kmymoney -v shows "kmymoney 5.1.80-462fb8d2d" and
>> here is a new BT
>>
> [... bt removed ...]
>
>>> The problem though, as Jeremy already pointed out, is a div by zero
>>> but on (my) line 1212 there is no division at all. The only one in
>>> the vicinity is three lines further down and shows
>>>
>>>    d->stockSplit.setPrice(d->stockSplit.value() /
>>> d->stockSplit.shares());
>>>
>>> which only gets used for a stock split transaction and price mode set
>>> to PricePerTransaction.
>> Well, the BT does now point to this line, and I do have
>> PricePerTransaction set, but this is a dividend transaction, not a
>> split.  However, is this for a stock split, or if the transaction is
>> NOT a stock split?  Is there a  wrong test in that "if"?
> I should have written for "non stock split transactions". My fault and
> the test in that "if" is perfectly OK.
>
> The problem is/was, that a dividend transaction should have zero in the
> shares attribute of the transaction. In that case, the division cannot work.
I'll have to reconfirm I have the right version, but it looks like it, 
and I still crash.  The inner loop is if shares is zero, then divide by 
shares (depending on PriceMode).  Am I still misreading it?  (version 
shows kmymoney 5.1.80-352f01a87.
> I fixed another problem in that area, because one could edit a transaction,
> make it a "buy", add shares, switch to dividend and save the transaction.
> This would cause the number of shares erroneously being added to the account.
> That one is also fixed with those two commits.
>
> https://invent.kde.org/office/kmymoney/-/commit/0b0410a0ef51613824f2ddaf462efb207da478e0
> https://invent.kde.org/office/kmymoney/-/commit/69de26b2640717373a40dd4191101d480ebe69ba
>
> Thanks for all your help.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kmymoney-devel/attachments/20220709/580440ce/attachment.htm>