Wireshark-dev: Re: [Wireshark-dev] QCustomPlot 2.0.0
From: Anders Broman <[email protected]>
Date: Tue, 12 Mar 2019 14:52:15 +0000

Hi,

Qcustomplot 2.0.1 is out and it may have addressed some of the problems Roland saw https://www.qcustomplot.com/release/2.0.1/changelog.txt

( - added missing Q_DECL_OVERRIDE declarations, thus preventing warnings some compiler settings)

 

It looks to me like requiring Qt >= 5.10 is not realistic as this time ( Ubuntu 18.04 has 5.9 ) so moving to Qt Charts is not possible yet?)

So it seems like it would be a good thing to move to Qcustomplot 2.0.1. It would be good if someone who knows what they are doing took a look 😊

 

https://code.wireshark.org/review/#/c/17980/

 

Best regards

Andes

From: Wireshark-dev <[email protected]> On Behalf Of Alexis La Goutte
Sent: den 10 februari 2018 17:01
To: Developer support list for Wireshark <[email protected]>
Subject: Re: [Wireshark-dev] QCustomPlot 2.0.0

 

Qt Charts look nice (and it look available from Qt 5.6...)

Not really time for the moment to try to finish the port to qcustomsplot

and we have already try to push some patch on upstream but it is very long !

 

 

 

On Sat, Feb 10, 2018 at 3:03 PM, Roland Knall <[email protected]> wrote:

;-) My main concern though will be performance. We have to figure out, if QCustomPlot or Qt Charts is better in regards to performance, some initial tests do not show a clear winner, but all in due time.

 

My main concern is just, that we cannot integrate QCustomPlot 2.0 without modifications.

 

cheers

Roland

 

On Sat, Feb 10, 2018 at 12:52 PM, Martin Mathieson via Wireshark-dev <[email protected]> wrote:

Given how good the Qt Charts screenshots look, and how unimportant the change to control the ticks on my axis is to me, I will leave this.

Martin

 

On Sat, Feb 10, 2018 at 10:42 AM, Roland Knall <[email protected]> wrote:

I've tried to port Alexis patch to the current master and use QCustomPlot 2.0.0 final with it.

 

Ran into quite a few issues, mostly virtual methods not properly marked for override. This leads to a compilation dropout with -Werror. 

 

So, the question I am having is, is it worth implementing QCustomPlot 2.0.0 with all necessary changes, sending them upstream and supporting them in the meantime, or much rather wait for the upgrade 'til WS 3.0 and switch that to Qt 5.10 and use QtCharts instead? It depends of course on what version of Qt Charts is needed and if those support all necessary things for us, but in the long run seems to be the better method.

 

Does not mean we should remove QCustomPlot right away or at all, but keep it in the 2.x and go with Qt Charts in 3.0. As said, depends on the version of Qt we choose for 3.0

 

cheers

Roland

 

On Sat, Feb 10, 2018 at 11:23 AM, Martin Mathieson via Wireshark-dev <[email protected]> wrote:

Thanks Alexis, I had not seen your earlier draft..

Nor had I diffed the earlier QCustomPlot release against

 what we were using to realise that we were patching it.  Can already see one place where I did things differently to you :)

Martin

 

On Sat, Feb 10, 2018 at 9:02 AM, Alexis La Goutte <[email protected]> wrote:

 

Hi,

I have already try on 2016 (when there is qcustomplot 2.0 beta) (published my old draft > https://code.wireshark.org/review/#/c/17980/ )

But there is some issue (some change on function and need to reapply some wireshark patch if i remenber..)

But qcustomsplot 2.0 is compatible with Qt 4.8 (to Qt 5.9)

Cheers

 

On Sat, Feb 10, 2018 at 8:43 AM, Roland Knall <[email protected]> wrote:

Hi

It would be a good thing for 2.6. So gar it should work out. Just a short request, if you add it, please check it Qt4.8 is still supported for it (requirement for 2.6) and if you add it, we should move it to a new directory.

I’ll do the latter in a separate patch before I think.

Regards
Roland


> Am 10.02.2018 um 00:03 schrieb Martin Mathieson via Wireshark-dev <[email protected]>:
>
> Hi,
>
> This was released September last year, has anyone else played with it?
>
> Locally, I got all our uses of it compiling (but mostly not tested).  The reason I tried it out was that for the LTE RLC Graph, I wanted to use the QCPAxisTickerFixed class, to force it to not subdivide the Sequence Number Y axis (which are whole numbers).  This works nicely now for me.
>
> Unless anyone knows of any problems with it, I will work through the changes to our other users of it over the next few days and post a WIP patch - I'm probably not in a good position to test all of them thoroughly.
>
> Regards,
> Martin

> ___________________________________________________________________________
> Sent via:    Wireshark-dev mailing list <[email protected]>
> Archives:    https://www.wireshark.org/lists/wireshark-dev
> Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
>             mailto:[email protected]?subject=unsubscribe
___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <[email protected]>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
             mailto:[email protected]?subject=unsubscribe

 


___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <[email protected]>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
             mailto:[email protected]?subject=unsubscribe

 


___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <[email protected]>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
             mailto:[email protected]?subject=unsubscribe

 


___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <[email protected]>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
             mailto:[email protected]?subject=unsubscribe

 


___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <[email protected]>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
             mailto:[email protected]?subject=unsubscribe

 


___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <[email protected]>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
             mailto:[email protected]?subject=unsubscribe

 

Attachment: smime.p7s
Description: S/MIME cryptographic signature