V_TFT RELEASE Ver 3.6 Bugs:

[Megahurts]

Hello ME Team,

Thanks for the quick update to fix some "Major" bugs, mostly the license problem.

However,,,, Still some work to be done sorry.

RULER POSITION MARKERS:
Now, the markers on the ruler for X,Y position of the mouse pointer are only visible when the pointer is moving.
When pointer stops moving, the markers disappear. Without an additional display (requested on wish list), digitally showing
where the pointer currently is (within the screen edit area), the Ruler has become almost useless.

(thanks for tightening up the screen edit area on the right and bottom sides though, looks much better)

LAYER LOCKING OBJECTS FROM SELECTION
Layer lock/unlock still needs to exclude objects from pointer interactions like trying to drag-box select multiple objects.
In fact, you can not even do a drag-box selection of foreground objects if pointer starts on a locked object behind them.
(you end up only selecting that locked object, and having the selected layer changed to the layer the locked object is on. NOT Cool!)
Toggling visibility for the locked layer is only way to allow drag-box selections to be done for stacked objects.
(Sorry, but the way the layers switches function are not intuitive to me at all.)

Please consider this function description for the layer switches 'Visible' and 'Lock'. Or explain why I would be wrong in thinking please.

Currently:

Visible- To the eyes only or to the application for object selections? You have it configured for both no mater what 'Lock' is set at. ????

Lock- Objects can not be edited, but can be selected no mater what layer is currently selected. ???? Why?
If you can not edit it, why be able to select it?
If "Locked" objects are included with unlocked objects in multiple object selections, most attempts to edit the collection cause an Error,
sometimes crashing the application.

PROPOSE:

Visible- For the eyes and application based on objects layer 'Lock' state. Locked or invisible objects can not be selected.
Visible - unlocked objects can be selected.
In order to align layer stacked objects, we need to have a state where the back object is visible, but TOTALLY out of interactions
for moving objects in front of it.

Lock- All objects on a 'Locked' layer are excluded from any action at all. NO selecting of objects by click or drag-box.
Clicking on a locked layer object will not change which layer is selected, only selecting unlocked object(s) causes change in what Layer(s) are selected.

This would eliminate errors from happening when locked and unlocked objects are selected together.
Its BASIC, do not allow it to happen in the first place guys.

KEYBOARD CTRL KEY FUNCTION:
Please , please , Oh God Please, add the keyboard Ctrl key modifier to mouse left-click object select/deselect from multiple object selections.
V-TFT has to be the only Windows program on the planet Earth that does not do this. ????? WHY ?????
It needs to work with making Layers selections also, but only with unlocked layers.

Obviously, no way to test if "Live Updates" works yet soooo..........

My first 5 minutes worth of trying it guys, will post anything else I find here, Robert.

(Thanks for the update regardless )

I welcome all bugs for version 3.6.0 to be reported here too, until another version is released that is.

[janko.kaljevic]

Hello Megahurts,

Actually I was not able to reproduce issue with ruler markers disappearing.
When my mouse pointer stops ruler markers are visible and still.
And I agree with you that digital representation of pointer position is good idea. I will put this on our wish list.

To avoid selecting objects on locked layer, you can set it as not visible. This is main feature of visible option for layer.
and when layer is locked it will prevent adding any modifications to it.
When layer is no visible, user will not be able to select any object from it and modify it. (No matter how lock is set)
We will consider not to allow selection of objects on locked layer, thanks for reporting this.

Please try to use Shift instead CTRL, for multi select. It will work just fine.

Best regards.

[Megahurts]

Hi janko,

How are you? Been awhile.

Thanks for pointing out the shift key, did not know that, good to know.
(I could not believe there was not a way to do it. I am relieved there is a way, even if it deviates from Windows Norm.)

On the ruler markers:
Weird thing, it turns out that it does this on my project I started making in Ver. 3.5 first screen only.
I loaded it in to V-TFT, then when I was finally done with the screen and added 2 more, the markers
work fine on them, but still not work on the original screen I made in V3.5?

I have to say, working on a project now, is much better and going between the compiler and V-TFT is better.

I do get some un-expected results with cut/copy - paste an object from time to time though.
I have had V-TFT give an error message saying I can not place (paste) an object on a invisible layer (layer was not selected).
I still get objects selected from invisible layer when selecting ones from a visible layer over the invisible one.
The layers and switches just do not work consistently for me?
I have had objects get pasted on locked-invisible layers despite V-TFT saying it could not be done. (I was not trying to do this)

So I have to believe there are still some bugs with layers operations and controls.

Also, selecting an object from the components drop-down list does not always select it on the screen edit area.

Try this action also:
Select multiple stacked objects on multiple layers (all visible and unlocked), copy them, then paste them on a different screen that has only
one layer.

Maybe the bugs I am seeing are due in part to loading a project from a older version of V-TFT?
I don't know.

But thanks for getting back to us, Robert.

[Megahurts]

Hi,

I have a screen shot showing the lack of ruler markers. The mouse pointer never gets shown from doing this so I have
put red arrows showing about where they should indicate it is, but do not once the mouse has stopped moving.

It seems that some bugs in V-TFT do not happen until there are many objects of different types being used on a screen.
This screen shot shows that I am working with 59 objects on this screen alone, so any chance of easily replicating any
users bug reported may need more than just trying to replicate with a single or few objects reported on.
Just to let you know what I have observed so far.

missing_markers.jpg (73.12 KiB) Viewed 5294 times

missing_markers2.jpg (33.92 KiB) Viewed 5249 times

Additional screen capture showing missing markers on my projects first screen. Mouse pointer is centered on "CALIBRATE" Button, as shown.

Janko, even when my mouse is moving and the markers are visible, they flicker on and off rapidly. Maybe an issue with Intel Graphics adapter?
Or DirectX or OpenGL drawing routines used by V-TFT?

TIP: For all users, do not make the mistake of naming a Screen "Main". This conflicts with the "main:" section of a program.
(Reserved Word. and YES I have done it, more than once actually )

Thanks for checking on us, Robert.

[Megahurts]

Hi ME Team,

Here is another one (bug), that slipped through the cracks:

event_routine_namechange.jpg (24.55 KiB) Viewed 5271 times

1- Changed the name of the Object from RadioButton2_Avg0OnClick() to RadioButton2_Avg1OnClick().
Edited the event called routine name and rebuilt: No problems.

2- Changed the name of Object RadioButton2_Avg4OnClick() to RadioButton2_Avg3OnClick().
Tried the same on it #3, and big problem, #4.
No mater what I try, the event routine name (#4), will not change to reflect editing (#2,#3). Even changing the name in the User Code editing
window (#4), will not take effect, a rebuild of he project causes it to change back to RadioButton2_Avg4OnClick() (#4).

There seems to be a problem with a persistent V-TFT SW Object name variable value that does not get updated under every circumstance.

How do you change the Objects name and the generated event routine name after an event routine has been made?????
(Safely that is)

And if there was code in an event routine, and you change the objects name, you lose that code when it does actually
change the routines name to match.

This is a big problem because it falls under "destructive code treatment or loss" category.

It will probably also be a bug that does not replicate easily, if my luck holds true.

Robert.

[Megahurts]

Hi ME,

Another one that did not get fixed from before:

old_event_routines.jpg (10.79 KiB) Viewed 5250 times

These objects have been deleted that originally declared these routines, but the routine names remain in this list and in the
event file, the routines remained until deleted using the compiler.

Robert.

[timtianyang]

Hi ME,

Another one that did not get fixed from before:

old_event_routines.jpg

These objects have been deleted that originally declared these routines, but the routine names remain in this list and in the
event file, the routines remained until deleted using the compiler.

Robert.

This happened to me too. I have to restart Vtft to refresh names of changed event handlers. not quite user friendly.

[Megahurts]

Hi ME,

Sorry, but got another one for you:

screen_shifting_components.jpg (65.38 KiB) Viewed 5219 times

1. This shows approx. the objects that got pushed to the right. It all should be at the left where #3 points.

2. tried to place a Label (#2), object at location #3 when #1 was in proper position at left (#3).

What I got is as shown except #2 was more to the left and everything shifted to the right #4--> #1.

This happened despite the layers being locked and placing the label was on the last layer that was unlocked #5.
(circles are on layer-0, The box on layer-1)

How do I fix something like this?? Start over?

Also: this screen shows the lack of ruler markers too. So that problem does not pertain to just my other project that had a
screen from a older version of V-TFT. This was a fully fresh project started in V-TFT 3.6.0.

????????????????, Robert.

[Megahurts]

Hi again,

trying to fix the above posted problem and I got this for my troubles:

v-tft_crash_error1.jpg (36.03 KiB) Viewed 5211 times

v-tft_crash_error3.jpg (18.78 KiB) Viewed 5211 times

This happened after I used the "Undo" button in V-TFT to fix the previous problem.
It seemed to have.
But I got the crash when I clicked on the 'Box' Component.

(No, I can not send error reports in full, I have email from my machine disabled for security)

Robert.

[Megahurts]

Hello MikroE,

I noticed this happening the other day:

Place a round box on the screen,
set "Transparent" property to 'true'- sometimes (not always), the 'Gradient Enabled' property switches from 'true' to 'false'.
Now set 'Gradient Enabled' to 'false' (if it did not change),
box interior should now be showing background,
set 'Transparent' property back to 'false',
box interior should show the 'single' solid color set in objects 'Color' property, but does not, it shows the disabled 'Gradient' colors selected in the properties.

Seeing the box on the screen solid color actually be displayed requires toggling 'Gradient Enabled' to 'true' and back to 'false' for it to show correctly.

I don't know if bugs like this also cause the code generated to act the same, but sometimes worry they might. I hope the software
programmer checks on that when bugs are repaired.
[UPDATE]- After some checking, YES it does affect the code V-TFT generates. The 'Transparent' property in code is reversed from what ever
the objects property is set at. See this post about it and suggested work around also:
http://www.mikroe.com/forum/viewtopic.php?f=160&t=54182

Regards, Robert.

[janko.kaljevic]

Hello,

1- Changed the name of the Object from RadioButton2_Avg0OnClick() to RadioButton2_Avg1OnClick().
Edited the event called routine name and rebuilt: No problems.

I was able to reproduce the problem when changing the name of the component (modifying last number in its name) caused new procedures to occur.
We will work on this and we will fix it.

Unfortunately I was not able to loose the code when the component name was changed. Please can you tell me how did you do it.

These objects have been deleted that originally declared these routines, but the routine names remain in this list and in the
event file, the routines remained until deleted using the compiler.

Yes I confirmed this. After deleting a object from screen its events will stay in list. We will work on this matter too.

This happened despite the layers being locked and placing the label was on the last layer that was unlocked #5.
(circles are on layer-0, The box on layer-1)

Please can you send me this project and I will inspect it thoroughly.

Seeing the box on the screen solid color actually be displayed requires toggling 'Gradient Enabled' to 'true' and back to 'false' for it to show correctly.

I was able to confirm this too. And I have forwarded them.1

I apologize for the inconvenience, and thanks for reporting.
We will work to find solutions as soon as possible.

Best regards.

[Megahurts]

Hi Janko,

Thank you for looking into these bugs and sending them to developers.
(on the problem of code loss after renaming a routine, I do not remember exactly how it happened and have not been able to
replicate it also??)

Not sure if the V-TFT project files will help you, but here they are for you to inspect.
(this is the second project files, I have been posting about bugs encountered during work on 2 different projects. The first project
is now too large to upload here.)

cool_sensor_p32mmb.zip

(56.53 KiB) Downloaded 130 times

I have another bug for you to check out also, sorry.

not_vertical_line.jpg (24.17 KiB) Viewed 5146 times

Picture shows a small line placed on a circle object (same layer), and its end point coordinates.
As can be seen, the line should be perfectly vertical (both end points left position @ 6 pixels), but it does not show up that way in V-TFT.

Screen orientation is set to 'portrait' and line width is 'one' (1) as can be seen also.
Image also shows I have zoomed in to the objects locations, but it shows same way at any zoom level.

This object is also in the project file I uploaded for you.

Thanks again, Robert.

[Megahurts]

Hey MikroE,

Here is another bug:

When Ruler is enabled, the 'save screen to bmp' function includes the ruler in the image and cuts off the same amount (rulers width), from the screen's
right and bottom sides.

See image below. I have to wonder?? do you test the software before you release it??

throttle_management_screen.jpg (17.13 KiB) Viewed 5141 times

I'm starting to feel like I should get paid for all the "testing", Robert.

[Megahurts]

Hello MikroE,

Question:
Is it to be assumed that a 1 =True and 0 =False for ALL object properties that use true/false?

If it is so, then there is bug with the box object "Transparent" property code being generated. Here is what I found in code:

true_false_transparent.jpg (47.35 KiB) Viewed 5123 times

Top half of picture is from V-TFT, Bottom half is from mBasic Pro PIC32.
(Project HW Pattern is MMB for PIC32, Compiler selection is mikroBasic Pro for PIC32)

1- Shows Box object in V-TFT and the objects declaration in the driver file.

2- Shows the Visible, Active and Transparent properties in both V-TFT and the driver code file.

As shown in V-TFT (if 1=true and 0=false), the property values should equate to: Visible = 0, Active = 1, Transparent = 1.
But the code V-TFT generates in the driver file is: Visible = 0, Active = 1, Transparent = 0.
Or please give reason why "Transparent" is different (reverse), from all others?
(Transparent property seems to be the only one I have found so far that behaves this way)

I have not checked this for EVERY object but YOU should please.
(I checked another object, "Button", I have in this project and it is set to 'false' in V-TFT, but shows up set to '1' in the code.)

Regards, Robert.

[Megahurts]

Hi whoever,

Not sure these are bugs, but they sure are annoying:

1- Select any object in a project with 2 or more layers and change its name to something custom,
right-click on it and select "Cut",
change layer to paste object to,
paste the object back in to V-TFT,
and now the name has been changed to the V-TFT default objects name and differentiating number (first available numerically).

2- For any components color property, when a color is selected that does not have a name (like Cl_Black, Cl_Red.....),
V-TFT shows a Hex number like: $008000FF
That value requires a LongWord variable to hold and TFT routines have a Word length variable declared?
The generated code will show a different Hex Word value like: 0xF810
How about maintaining a standard for these and show the actual "Word" length value we will be using in programming?

Thanks, Robert.