Posts made by dissy

@madsrassmussen Yes, I noticed this odd behavior starting in 2.0.3 and it seems to be a different odd in 2.0.4, but neither work as expected.

When letting the web terminal do the upload, the named file under 'apps' is zero bytes and the file contents end up in 'main.py'

In 2.0.3 I could download main.py to my PC, rename it, and use web terminal file manager to upload it into the apps folder, and all was well.

In 2.0.4 however, it almost seems like the device 'My Apps' menu/list is unaware of this file, even though it is listed. When selected the device runs the .py file listed after the one I selected instead.

I've also had the variable undefined errors but I think that is a different issue. Maybe not?
One example is copy/pasting the mpy code to a windows computer into say notepad.
The device requires unix style line endings, not windows line endings. It also requires no blank lines to exist.
(This is why I would copy the existing 'main.py', as it was already formatted correctly.)

Otherwise, most times the upload just doesn't work at all. I too have not found a pattern or submitted a bug report due to this lack of info. I wasn't even sure it wasn't just me :)
Hopefully someone more knowledgeable can chime in

@peter None needed. It is public domain

@peter Yes, that is perfectly OK with me. I'm glad you found it useful! Creating a module sounds like a wonderful idea

@m5stack
In the module "Display", the block "Module Display config" generates the following upython code and error message:

module_display = M5.addDisplay({"module_display":{enabled:True, width:1920, height:480, refresh_rate:60, scale_w:0, scale_h:0, pixel_clock:7425000}})

NameError: name 'enabled' isn't defined

The correct python syntax that should be generated, with strings double-quoted, is:

module_display = M5.addDisplay({"module_display":{"enabled":True, "width":1920, "height":480, "refresh_rate":60, "scale_w":0, "scale_h":0, "pixel_clock":7425000}})

Manually entering the above adds in the second display without errors, but putting that code in an "execute mpy code" block instead makes the UI editor unusable

@membersound I am quite literally about to order a display module myself, so I'm not speaking from experience, but I did some research this weekend before deciding to order.

By all indications, each display is independent.
For instance, with the m5unified library, the core2 uses " M5.Lcd.print() " for the built-in and " M5.Display.print() " for the display module.
It seems at first glance all the methods are the same (print, drawRect, setFont, etc)

So you'll almost certainly need to make some code changes.
But if your goal is to "mirror" the displays, I'd think it would just be a matter of having two of every command? As in copy/paste and in the second one change "Lcd" to "Display"

(and now on to the long shipping wait ahead ;)

@colobro309 That isn't a bug, that's just how mathematics works.
And/or are logical functions, they only take true/false as inputs and returns a true/false.
Add/subtract are arithmetic functions, they only take numbers as inputs and returns a number.

You will need to convert your string or number into a true/false value first.
The "IF" block can do this as it returns true/false.
Eg "If number=0 return false else return true"

@britontour said in UiFlow 2.0 discuss(how-to, bug, feature request or sometings):

I see "PRIVATE" on the Develop screen of the S3, even though I configured it as Public.

Before it connects it will read "private" in yellow, and show "user account" as "none"
After it successfully connects this will change and should show "public" in green and your email address in "user account"

What server connection does it need? I may need to get IP or ports whitelisted at my office.

https (port 443) to host uiflow2.m5stack.com

@m5stack @IAMLIUBO
UIflow 2.0.2 for cardputer, always reports "private"

File /fs/system/cardputer/develop/public.jpeg
The background color was changed from yellow to green, but the text still reads 'private' instead of 'public'

@britontour If m5burner registered the device into uiflow, and it shows having a MAC and Token value (partially masked out), then it should be re-paired properly.
I'm unsure why your device can not connect to the server to report itself online.
I've just verified the process still works (I don't have a core s3 but used a cardputer which is stamp s3 based)

I did just discover a bug in the 2.0.2 "develop" screen. It may be cardputer specific, but it incorrectly displays "private" in both modes. It's just a display bug though, it isn't effecting my connection

@iamliubo Feature request for UIFlow 2.0.2 on Cardputer

On the main menu, can the characters on the arrow key buttons also be assigned to perform navigation?
IE so pressing right (Function + arrow) or pressing forward-slash (the same button without holding Function) both navigate right, and the same for left / comma?
It would also be nice to have that functionality in the App.list menu with the up/down arrows (semi-colon and period respectively)

This way screens that only respond to arrow keys can be navigated without having to hold the Function key.

(Note: The new behavior makes perfect sense for screens using both cursor navigation as well as text-input, eg the wifi settings, where holding Function for navigation should remain required)

@bigal2301 Unfortunately there isn't a built-in keyboard buffer handler block.

There is a keypress event handler which is good, but it seems designed so you can't both get a key code and the key string at the same time. This makes it tricky to handle 'special' keys.

I just threw together an example called "Example Keyboard Buffer" in the project zone.

It handles enter with a function, handles the delete key, and filters out tab/escape.
Much of the code is comments explaining things as well as a small scroll buffer on screen.

@waittherack Make sure to also add an "execute (eg import)" block to with:
from hardware import sdcard

I also had the FileIO option under Software turned on, but not sure if that's needed just to reference a file by path.

@britontour You may wish to completely start over with the pairing process.

1. Put your device in programming mode and use m5burner to do an 'erase'
2. In uiflow2, click the button to select a device to connect to, and hover the mouse over your device icon. In the top right of that icon (when hovered over) should be a square you can click to edit the device info. In there should be an 'unregister device' button to fully delete it.
3. Back in m5burner, reflash uiflow2 to the device and configure for your wifi.

Be sure to keep the 'public' switch on this time.
When the flash is done, keep the usb cable attached, hit reset on your device to get into the uiflow menu, and open uiflow2 in your browser again.
Click the button to select a device. In a few seconds it should show in the 'public' list with a green dot. Be sure to select it here and not under the usb devices list. Unplug the device from USB.

You should now be able to turn the device on/off and the device button in uiflow2 will go green/orange to reflect it is paired properly.

@invxblrkl Not available yet.

There are touch count and x/y reading blocks however, so it can be done the long manual way. I just had my own go last week at making touch buttons this way.
Check if touch count went up, and if so check if x/y is within the area you want a button.
I used rect objects on screen to get the coordinates.

I'd assume a switch can be done that way by changing the rect background color, or an image you flip between two pngs.

The official way to add comments only adds them to UIFlow (they are not exported into the python code)
You can right-click any block and select "Add comment", which enables a question mark icon on the block. Clicking that icon lets you add a comment in a bubble.
(I use those comments to describe why I did something specifically in UIflow/blockly.)

The unofficial way is the code block you show. Those comments DO get exported into the python code.
(I use those comments for things like describing variables usage and (re)configuration, stuff people would still need to know when using the micropython code directly.)

It would however be wonderful to have one official "Comment block" block that could be set to work either way, and set to start expanded or collapsed, etc.
I think it would be wise to retain "Insert into python code" or not as an option, as each type has advantages in different circumstances.

@ajb2k3 My intent was making it known to the devs, in the hopes the problem could be fixed.

Bug fix request
UIFlow2 "http" block is completely non-functional, failing to load the 'requests' library. There is a typo with an errant 'u' that should not be there.

ImportError: no module named 'urequests'