LabWindows/CVI Idea Exchange

Post an idea

Add full C++ support

I add this new idea based on the info on this thread.

I know that this idea is almost the same, but the basic reason is quite different.

I think that C++ offers much more flexibility for the development and it is a much more powerful approach for the future.

Now CVI uses CLANG/LLVM compiler, that is a C/C++ compiler.

But, as written by Luis from NI

Supporting C++ entails far more than just the compiler. For example: the standard C++ library (including templates, etc...), the APIs of all the standard CVI libraries, the source editor (auto-indentation, coloring, etc...), code-generation from the various wizards, memory runtime checking, debugging, etc... so, even though the compiler does support it, for all of CVI to support C++ is still an extremely large undertaking that has to be weighed against other, competing work that's going into CVI.

View All (1)

Add a tool for static code analysis

View All (4)

Check for MISRA C compliance

View All (3)

Before CVI2012 a conditional breakpoint could be toggled between enabled and disabled with a single mouse click. The behaviour changed in that version so that a mouse click now deletes the breakpoint.

I understand that this is more consistent with the behaviour of an unconditional breakpoint, which is set and deleted with the mouse click; I understand also that Ctrl+F9 shortcut key has been introduced to toggle breakpoint state, but in my opinion this new arrangement is less user friendly than before.

The risk is that once you set a conditional breakpoint, possibily with a complicated condition and/or pass count set to match some particular operating condition, you can delete it with a single click while you would maybe simply disabling it.

The same applies if you press F9, so I suggest additionally that when F9 key is applied to a conditional breakpoint it disables the breakpoint instead of deleting it.

View All (2)

This suggestion comes from this discussion: basically, it seems strange to me that I may have a disabled quit button and the close control (the red cross in the upper right corner of the panel) which is active at the same time.

It seems to me that the close control should honour the state of an existing quit button declared in Close control panel attribute. At present the only solution is to test the state of the button within the button callback itself.

If this idea is not accepted, at least a note should be added to the documentation, e.g. in ATTR_CLOSE_CTRL help page, specifying that the callback is fired regardless the state of the control.

View All (2)

As can be seen here, "Save changes" items in Environment options are global to the CVI IDE.

I suggest these options are made project-specific instead. The rationale behind this suggestion is simple: suppose I am developing project A and I enable autosave modifications before compiling and running the project; at the same time I may need to look at the code for project B already deployed to customers, and I want to prevent modifications to the code without notice. At present, while I am testing the application A I set these options to "Always"; if next I come and see the code for project B, I may accidentally make (and save!) modifications that change the code of the deployed app without notice

If these options were per-project I would be granted I am not making mistakes.

Several CVI functions exist that permit to add / install functions in the code. At present, we need to manually create the function prototype, and the fastest way I found is this one: open the function panel for e.g. InstallPanelCallback function, select Event Function parameter, right-click the field or press F1 to show the help, select the function prototype dragging with the mouse, Ctrl-C to copy it, close the function panel and paste the prototype in the code.

It would be *very* handy if a way existed to speed up the process: a "Generate Prototype" item could be added to Code menu that takes the function name set by the user in Event Fuction field and generates the appropriate function prototype in the source code.

There is a huge set of functions that could benefit with this addition, following is an initial list but I am sure I'm missing some more functions:

In the User Interface Library:

All Install*Callback
PostDeferredCall
PostDeferredCallToThread
RegisterWinMsgCallback

In the Utility Library:

CmtScheduleThreadPoolFunction
CmtScheduleThreadPoolFunctionAdv
CmtInstallThreadPoolCallback
CmtInstallTSQCallback

In RS232 Library:

InstallComCallback

In VISA library:

viInstallHandler

In DAQMx Library:

DAQmxRegister*Event functions

In TCP Library:

All Register* functions

In Network Variable Library:

All CNVCreate* function

In the ActiveX Library:

CA_RegisterEventCallback

In the Asynchronous Timer instrument driver:

NewAsyncTimer
NewAsyncTimerWithPriority

In the Programmer's Toobox:

InstallWinMsgCallback
PostDelayedCall
PostDeferredCallToThreadAndWait
ChainCtrlcallback

And don't forget all sorting functions that require a comparison function to operate:

qsort
SortTreeItems
SortTableCells
HeapSort
List*Sort fuctions for lists

Zoom Normal key-control in Graph Tool

Hi,

CVI already has a way to check for duplicate instances of an application through CheckForDuplicateAppInstance(..). My suggestion would be to extend this functionality so that the duplicate application instance could send a message to the original application instance. This is common for when the programmer wants to pass the command line arguments from the duplicate instance to the original instance (before exiting) so that the original instance can act on them. Normally this is done through some inter-process communication protocol like DDE, but there's an opportunity to simplify this for the programmer. Probably the easiest implementation would be an application-level callback that will be called when the duplicate instance is started, which would accept a pointer to some data that could be set by the second instance.

Thanks!

GetTableIndex function is missing

sometimes i use the table object in CVI but there is one function that is missing and i think it's really usefull, the one that give you the ring index directly.

now i'm doing it in this way, maybe there is a better way to do it but i've not spent too much time to do it.

/// HIFN This function give you the item ring index.
/// HIRET an integer value containing the function result.
/// HIPAR panel/panel handle where the control is located.
/// HIPAR control/which control i have to analyze.
/// HIPAR column/in what column i have to look.
/// HIPAR row/in what row i have to look.
static int GetTableIndex(int panel, int control, int column, int row)
{
	int Result = {0};
	long IndexLength = {0};
	char *IndexTemp = {0};
	
	GetTableCellValLength(panel, control, MakePoint (column, row), &IndexLength);
	IndexTemp = calloc(IndexLength + 1, sizeof(char));
	GetTableCellAttribute (panel, control, MakePoint (column, row), ATTR_CTRL_VAL, IndexTemp);
	GetTableCellRingIndexFromValue (panel, control, 0, MakePoint (column, row), &Result, IndexTemp);
	free(IndexTemp);
	IndexTemp = 0;
	return Result;
}

Most user settings are kept in either the project or the workspace file. Unfortunately, this is not true for building a function tree (Options/Generate Function Tree). Starting CVI the previous entries are lost. Actually this is an old issue but may be not too old to be improved...

So I suggest to save the information about Instrument name, Function prefix, Default qualifier, and Output function tree in the appropriate configuration file (prj or cws)

Thanks

Improved error reporting

View All (1)

Hello,

building upon my earlier (but difficult to implement) suggestion and the forum discussion on event data here I can provide a hopefully improved suggestion.

The issues addressed by this idea:

receiving an EVENT_DATA_CHANGED event for a numeric control does not mean that the numeric value has changed
receiving an EVENT_DATA_CHANGED event for a numeric control does not provide information if the value has been increased or decreased
receiving an EVENT_DATA_CHANGED event for a numeric control does not provide information if the 'up' or 'down' arrow button has been pressed (which is not necessarily the same as the second issue)

Idea: Provide eventData1 and eventData2 information.

For example, eventData1 could tell about the numeric value:

eventData1 = -1: numeric value has been decreased
eventData1 = 0: numeric value is unchanged
eventData1 = 1: numeric value has been increased

eventData2 could tell about the increment/decrement arrow, e.g.:

eventData2 = -1: decrement arrow has been pressed by mouse
eventData2 = -2: decrement arrow has been pressed on keyboard
eventData2= 1: increment arrow has been pressed by mouse
eventData2= 2: increment arrow has been pressed on keyboard

Since so far no eventdata are provided this suggestion should not break backward compatibility.

View All (3)

Option to have popups always on top

View All (3)

Strip chart trace color control

I used to be able to change strip chart colors "on the fly".

This facility was available pre Version 8.5 - see http://forums.ni.com/t5/LabWindows-CVI/Changing-color-in-a-Strip-Chart/m-p/2292246/highlight/true#M60735

This would allow me to change the color for future points and NOT affect what was already plotted. This is the way that ATTR_TRACE_COLOR used to work.

As the above discussion thread says, this feature was a bug, but a very useful one at that!

Could we have a control that worked in this way ATTR_TRACE_POINT_COLOR which would only work for points plotted from thereon, the strip chart still scrolling as it does at present.

I believe that this could be done if I used a graph control, but that would involve explicity redrawing the graph for every plot point to achieve a scrolling display.

LabWindows/CVI Idea Exchange

Add official support for GIT

Support newer versions of C runtime on real-time systems

Add full C++ support

Add a tool for static code analysis

Check for MISRA C compliance

Restore toggle conditional breakpoint state with mouse click

Make panel close control honour the state of a quit button (if any)

Improve the generation of source code browsing information

Make "Save changes" environment options project-specific

Generate function prototypes for programmatically installed functions

joystick support must be available and instrument drivers should be provided

Zoom Normal key-control in Graph Tool

Improved functionality for duplicate application instances

GetTableIndex function is missing

Add Transparency Attribute for LabWindows/CVI Panels

Remember values for generating a function tree

Improved error reporting

Provide eventdata for numeric controls and EVENT_DATA_CHANGED

Option to have popups always on top

Strip chart trace color control