OpenCPN Version 4.8.0 Released

[transmitterdan]

Originally Posted by CarCode

Using the same ID for different checkboxes results in a crash here.
See "ID_MAGSHOWCHECKBOX" in options.cpp at line 3937:

    //  "Mag Heading" checkbox
    pCBTrueShow =
        new wxCheckBox(panelUnits, ID_MAGSHOWCHECKBOX, _("Show true"));
    unitsSizer->Add(pCBTrueShow, 0, wxALL, group_item_spacing);
    pCBMagShow =
        new wxCheckBox(panelUnits, ID_MAGSHOWCHECKBOX, _("Show magnetic bearings and headings"));
    unitsSizer->Add(pCBMagShow, 0, wxALL, group_item_spacing);

Gerhard

Gerhard,

Can you tell me the exact steps to recreate the crash?

What version of wxWidgets is your version of O built against?

I ran this code in debug mode in Windows and it did not cause a crash or even an assert. I don't see why it should cause a crash. The only requirement for IDs to be unique is when an event handler needs to know which item triggered an event. But the events generated by these check boxes are not "handled" by O. Probably they could all be changed to wxID_ANY.

[did-g]

Originally Posted by nkiesel

I just tried 4.8.0 and have serious problems with mark displays. Maybe a configuration setting, but it's unnerving that this is even possible, Problem is that - as I zoom in (using either mouse or keyboard), marks disappear when charts are swapped to higher resolution charts. This is with ENC charts and "chart quilting" enabled.

I played around with "" but I have to crank it up all the way to "5" before this no longer happens (which results in a very cluttered display).

My ideal scenario would be that marks _never_ disappear when zooming in.

Did you try the reverse? ie -2 or -3.

If you uncheck 'Reduced detail at small scale' these marks will never disappear, even when zooming out... It's ok but for the last chart.

Maybe changes in quilt charts switching logic makes it worse in 4.8 vs previous versions but a chart update can also trigger it.

It seems to me they are better at following SCAMIN IMO recommendations, so may it could be used in quilt logic.

I haven't downloaded Netherlands charts recently though, they were very creative in their use of SCAMIN.

[CarCode]

Originally Posted by transmitterdan

Gerhard,

Can you tell me the exact steps to recreate the crash?

What version of wxWidgets is your version of O built against?

I ran this code in debug mode in Windows and it did not cause a crash or even an assert. I don't see why it should cause a crash. The only requirement for IDs to be unique is when an event handler needs to know which item triggered an event. But the events generated by these check boxes are not "handled" by O. Probably they could all be changed to wxID_ANY.

If you will like to read the code you will see that anywhere else the correct IDs are used for g_bShowTrue or g_bShowMag and pCBTrueShow or CBMagShow (ID_TRUESHOWCHECKBOX or ID_MAGSHOWCHECKBOX) but at this place they are wrong. This has absolutely nothing to do with a wxWidgets version but is simply a typo while writing code.

Gerhard

[transmitterdan]

Originally Posted by CarCode

If you will like to read the code you will see that anywhere else the correct IDs are used for g_bShowTrue or g_bShowMag and pCBTrueShow or CBMagShow (ID_TRUESHOWCHECKBOX or ID_MAGSHOWCHECKBOX) but at this place they are wrong. This has absolutely nothing to do with a wxWidgets version but is simply a typo while writing code.

Gerhard

Yes, it is a typo. But I think it has no ill effect. The ID values are no where used by the program.

I am trying to understand how it can cause a crash. Can you tell me exactly what steps you used to cause the crash?

[Gilletarom]

Hello All,

PC Lenovo W10.

I had a lot of trouble getting the tides. And as much to get the opening of a tide window. Am I alone?

Gilletarom

[verkerkbr]

Originally Posted by Gilletarom

Hello All,

PC Lenovo W10.

I had a lot of trouble getting the tides. And as much to get the opening of a tide window. Am I alone?

Gilletarom

Bonjour Gilletarom,

I have this mentioned before. It works, but you have to find the right spot.

This proves to be very difficult. Sometimes I have to click the tide bar more then 5 or 6 times to open the window.

I have proposed to make a small square in the tidal bar, where you click to open the tidal graph.

I still don't know where exactly the spot is to open the Window.

Regards from the Pays Bas,




Bram

[cagney]

Originally Posted by verkerkbr

Bonjour Gilletarom,

I have this mentioned before. It works, but you have to find the right spot.

This proves to be very difficult. Sometimes I have to click the tide bar more then 5 or 6 times to open the window.

I have proposed to make a small square in the tidal bar, where you click to open the tidal graph.

I still don't know where exactly the spot is to open the Window.

Regards from the Pays Bas,




Bram

I have previously had the same problem, but carefully testing on 4.8.0, I can't see any problem using Kubuntu 16.04 / 64 bit.
I can right click anywhere on the vertical tidal icon and open the window. I can even click outside the vertical column, on the sides, within half the width, and get the tidal graph window. It is a bit more sensitive at the very top and bottom.
Perhaps this is just a problem on Win?10?

Thomas

[P_Dub]

Thomas
I am with Windows 10 pro version and have no problems with that (tide and current bars). Further more my problem with the ship (png format) is OK now.
Providing you put only one in UserIcons folder.
Bye
P_Dub

[nkiesel]

Originally Posted by transmitterdan

Was this behavior different in 4.6.x?

I just build from commit 86cef240 (Beta 4.5.307) and that does not behave that way (i.e. once a mark is shown, it's always shown when zooming in further).

I will try git bisect to find out which version changed the behavior.

[Gilletarom]

Hello Dave,

Last night, it took me at least 20 clicks right for the tide window to open.

This morning, new test concerning the display of the tides. In one case out of two, with right clicc, it is the general menu that opens, in one case out of two, it is the window of the tide that opens.

And this is a phenomenon that I had already mentioned in an earlier version.

Gilletarom

[Gilletarom]

Hello,

Last night I tried to use the Microsoft Surface touch pad in W10. And I plugged the cable out of the VHF into the USB port of the tablet.
All worked well, information about AIS was available.
Then I tested the tides display. This worked much faster than on the Lenovo PC. And the opening of the tide window is correct

[CarCode]

Originally Posted by transmitterdan

Yes, it is a typo. But I think it has no ill effect.

Thank you to finally admit it is a typo. Next step is to admit this typo might have no effect to you with a stupid compiler but for others with a smart compiler it has.

Gerhard

[transmitterdan]

Originally Posted by CarCode

Thank you to finally admit it is a typo. Next step is to admit this typo might have no effect to you with a stupid compiler but for others with a smart compiler it has.

Gerhard

I have submitted a PR. I have no idea how a "smart" compiler can have an issue with a value that is no where used in the program. Where is the ID value for these wxWindow objects used? I could find no code that reads these values.

[transmitterdan]

Originally Posted by CarCode

Thank you to finally admit it is a typo. Next step is to admit this typo might have no effect to you with a stupid compiler but for others with a smart compiler it has.

Gerhard

I don't know what smart compiler you are using. But this typo has existed for 7 months.

[ejs]

Gerhard,

We are also still waiting for the exact steps used to re-create the crash you mentioned.

Eric