How to submit an example project?

[ROSW6341]

I've updated the example 'graph' project with the following modifications & extensions:

• Renamed to 'Chart'
• Separated Graph & Traces
• Added optional right side Y axis
• C++ class
• Suggest using TFT_eSPI version 2.4.22 or later w/ updated drawFloat()
• Suggest copying these files to the project directory to facilitate 'tweaking'
  on a per project basis

Not familiar with submitting to github. Is a RAR file acceptable & if so, to whom do I send it.

• Bill K [email protected]

[Bodmer]

The simplest way to provide a new example sketch is to zip it and drag-drop here in a message box.

To create a sketch and maintain it (respond to user issues and bug reports) then creating your own repository is the way to go, in that case I can provide a link in the ReadMe.

[ROSW6341]

Thanks for the info. A ZIP of the sketch is attached. There is probably more that could be done to the code but so far it works for me. Hope it is useful to others.
ChartTest-20220217A.zip
PS Feel free to have a look and clean up any obvious &/or blatant errors.
.

[Bodmer]

I loaded and ran it and it is working very well.

Ideally it needs to use the display width and height so the demo automatically fits any user screen.

[ROSW6341]

Glad it worked OK for you. Feel free to edit the code to use the display width & height.
Have done a bit more work on the chart code:

• changed all doubles to floats
• made dp's unsigned (uint8_t)
• changed the map-like code to call a common function that takes floats for the input & input range values but now takes longs for the output range values and returns a long.
• all location values sent to TFT_eSPI is no longer doubles but are now ints

Can merge the changes back into the example (& change the display width & height) and send you an update if you want.

Came across something that might be of interest to you. It is a simple means for rounding floats/doubles based upon the value of dp. (sorry, messed up the original version of the following)

float round(float value, int dp) {
float tens = 10 ^ dp; // -100, -10, 0, 10, 100, ...
return round(value * tens) / tens;
}

Might be better to make 'tens' an int or unsigned, check the sign of 'dp', and swap the mult & div (*, /) sequence accordingly.
Should to work properly for both positive & negative values but has not been tested.

[ROSW6341]

Did some more checking. 10^dp doesn't work properly at least for the esp32. 10^1 returns 11, not 10 for the Arduino ESP32 tools I'm using. Reworked the code and did a fair amount of testing (+/- values of dp and of value). The function as it now stands is:

float round(float value, int dp) {
if (dp == 0)
return round(value);

int tens = 1;

// calculate power of 10 from abs(dp)
// 10, 100, 1000, ...
for (int i = 0; i < abs(dp); i++)
tens *= 10;

if (dp < 0)
return round(value / tens) * tens;

return round(value * tens) / tens;
}