segfault in new ledger code

[Thomas Baumgart]

On Freitag, 10. Januar 2020 23:24:24 CET Jack wrote:

> On 2020.01.10 16:48, Thomas Baumgart wrote:
> > On Freitag, 10. Januar 2020 22:06:56 CET Jack wrote:
> > > On 2020.01.10 15:34, Jack wrote:
> >>> Per your suggestion in the Wayland thread, I just recompiled with  
> >>> UNFINISHEDFEATURES enabled.  Selecting the new ledger gives me an  
> >>> immediate segfault, from both 5.0 and master, and whether or not I  
> >>> have already selected an account.  I'm pretty sure I can easily  
> >>> enough get a backtrace, but before I do, what other information  
> >>> would you want to help debug?
> 
> >> Sorry - I forgot on important point - these segfaults are all using  
> >> X11.  I do have Wayland available, but it's not very stable, so I'll  
> >> test with Wayland a bit later.  I figure unless the crash is  
> >> specifically X related, it would likely crash under wayland also.
> 
> > Too bad that it crashes for you. Can you try to do this with an empty  
> > file just to make sure it is not related to your data? I expect it to  
> > be your data since it does not crash for me.
> 
> OK, the good news is that I see no crash or flickering or any other odd  
> video effects under wayland in the new ledger.  I'm also assuming  
> "UNFINISHED" is applicable here, as the tab order is odd, and even goes  
> outside the ledger, and a new transaction isn't actually saved.

I know. 

> In terms of the flickering, I assume you are using different graphic  
> widgets from the old ledger/transaction form.  Is there any hope of  
> fixing the issue in the old code, or does this just accelerate the need  
> to complete the new ledger?

Not sure about the former, but the latter becomes important. The 5.0 code
coming from our version 4 code base uses a combination of widgets for the
date entry. The unfinished new ledger code uses the standard Qt date entry
widget. At least we now know where to look for modifications and that we
do a plan B if necessary.

Thanks for testing. The results somehow confirm my expectations.

> I'm still not sure how common Wayland  
> actually is, but I seem to get the feeling that it's being made the  
> default display system enough that it will have to be addressed.

If more and more users are switching to it we should be prepared.

> I didn't do a full walkthrough of all views and windows, since I  
> haven't figured out why wayland is unstable for me and seems to crash  
> within five or ten minutes of use (or even just sitting) and after a  
> wayland session, I need to reboot because just going back to an X11  
> session doesn't work correctly.  I'll be happy to do further testing as  
> needed, but nothing very extensive seems likely, as I won't know if any  
> crash is because of wayland itself (or bad configuration on my part) or  
> related to KMM, though simply observing video behavior seems not to be  
> a problem.
> 
> Jack
> 

-- 

Regards

Thomas Baumgart

https://www.signal.org/       Signal, the better WhatsApp
-------------------------------------------------------------
"Hey! I could use Tex!" and I've only gotten to use it a little
so far but it's so far superiour to MS Words "be everywhere do
everything"-ness it's not even funny... (lordSauron)
-------------------------------------------------------------
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 868 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/kmymoney-devel/attachments/20200111/df2a42fe/attachment.sig>