Lost PyMakr in VSCode

[Paul Thornton]

@kgunnar Ralph is busy working on the fix now. Hopefully, it won't be much longer. I believe others have confirmed manually rolling back and installing VSCode 1.30 is a temp fix. but VScode seems to ignore its option telling it to not auto update at times so you may lose it again.

[combaindeft]

@paul-thornton said in Lost PyMakr in VSCode:

@kgunnar Ralph is busy working on the fix now. Hopefully, it won't be much longer. I believe others have confirmed manually rolling back and installing VSCode 1.30 is a temp fix. but VScode seems to ignore its option telling it to not auto update at times so you may lose it again.

Jupp.

• A) 1st make sure to disable auto-update ... or it'll sneak it in again :-)

In setting, search for: update.channel ... and Choose None or Manual

• B) You can download (rollback) here: https://code.visualstudio.com/updates/v1_30

[curtis.hendrix]

@combaindeft Just a side note: make sure to change the update.channel setting before uninstalling 1.31. I got burned after installing 3.10, since the updater runs before the settings get saved. So 3.10 saw the old update.channel setting, so it updated to 1.31.

[combaindeft]

@curtis-hendrix Yeah 🤓 *fixed-post*

[Ralph]

v1.0.6 has just been released to fix these issues. Let me know if any of you are still facing any after updating.

[Martinnn]

@ralph 1.0.6 just installed, is broken (no command bar, no connection) in VSCode 1.30

[Ralph]

@martinnn Make sure you run vscode 1.31. The precompiled serialport lib's are now rebuild for that version specifically.

[Martinnn]

@ralph OK, updated VSCode the second time today. Works with pymakr 1.0.6.

[Ralph]

@martinnn Good to hear!

Also, sorry for the delay on this new version guys. We'll have to work on our rebuild process a bit to keep pymakr compatible with vscode.

[jand]

@ralph: Today I suddenly couldn't connect anymore to the extension board (3.1) with a lopy4 through the COM13 port I always use. Next I even lost Pymakr completely in VSCode.

Reading the posts in this topic, I realized I needed to upgrade VSCode to the latest version, which I did. I now have:
Version: 1.31.1 (system setup)
Commit: 1b8e8302e405050205e69b59abb3559592bb9e60
Date: 2019-02-12T02:20:54.427Z
Electron: 3.1.2
Chrome: 66.0.3359.181
Node.js: 10.2.0
V8: 6.6.346.32
OS: Windows_NT x64 6.1.7601

And I have Pymakr 1.0.6.

But I cannot connect to the lopy4 anymore: > Failed to connect (Error: Port is not open). Click here to try again.

The VSCode command 'List serial ports' yields the following:
Found 2 serialports
COM5 (Dell) (copied to clipboard)
COM3 (Dell)

But I cannot connect to the lopy4. It doesn't work with COM5, not with COM3, and not with COM13.
When I connect through COM5, the console starts producing a continuous stream like this:
$GPGGA,,,,,,0,00,0.5,,M,,,,*00
$GPGSV,1,1,00,,,,,,,,,,,,,,,,79
$GPGLL,,,,,235944.000,V,N77
$GPGGA,,,,,,0,0,,,M,,,,*1B
$GPGSA,A,1,,,,,,,,,,,,,,,*1E
$GPGST,235944.000,,,,0,,,74
$GPRMC,235944.000,V,,,,,,,,,,N40
$GPGSV,1,1,00,,,,,,,,,,,,,,,,79
$GPGLL,,,,,235946.000,V,N75
etc.....

What to do?

[Ralph]

Hi @jand, looks like your device is not detected by the driver anymore. If it would be, it would definitely show up in the 'list serial ports' list. Did you restart your computer and try again? Happens on my mac sometimes if you unplug the device when it's connected, it doesn't show up again until you reboot.

I don't know what COM5 has attached, but that doesn't look like something any Pycom device would output :) Perhaps @iwahdan can confirm that.

[robert-hh]

@jand said in Lost PyMakr in VSCode:

$GPGGA,,,,,,0,00,0.5,,M,,,,*00
$GPGSV,1,1,00,,,,,,,,,,,,,,,,79
$GPGLL,,,,,235944.000,V,N77
$GPGGA,,,,,,0,0,,,M,,,,*1B
$GPGSA,A,1,,,,,,,,,,,,,,,*1E
$GPGST,235944.000,,,,0,,,74
$GPRMC,235944.000,V,,,,,,,,,,N40
$GPGSV,1,1,00,,,,,,,,,,,,,,,,79
$GPGLL,,,,,235946.000,V,N75

These look like sequences from a GPS module. Do you by change have one attached?

[crumble]

@ralph said in Lost PyMakr in VSCode:

I don't know what COM5 has attached, but that doesn't look like something any Pycom device would output :) Perhaps @iwahdan can confirm that.

This is the output of NMEA GPS sentences. Can be a pytrack ;)

@jand If VSCode works like Atom, click in the window of the output and hit ctrl-c a couple of times. This shall bring you into the REPL.

There may be a setting which hits the break key as soon the pytrack is connected.

PS: Your sentences show a lot of garbage. Check the way you read them. It seems that you drop a lot of chars and mix up multiple sentences.

[Paul Thornton]

This post is deleted!

[jand]

@robert-hh No, I do not have a GPS module attached. Just an expansion board 3.1 with a Lopy4.

[jand]

@ralph I did restart the computer, but that does not change anything. Connecting to the Lopy from VS Code through the wifi still works. But through USB doesn't.

Desperate to understand the problem, I now installed VS code with the pymakr extension (1.0.6) on a brand new laptop. VS code then shows the following configuration:
Version: 1.31.1 (system setup)
Commit: 1b8e8302e405050205e69b59abb3559592bb9e60
Date: 2019-02-12T02:20:54.427Z
Electron: 3.1.2
Chrome: 66.0.3359.181
Node.js: 10.2.0
V8: 6.6.346.32
OS: Windows_NT x64 10.0.17134

But here again I cannot connect through usb. The symptoms are now different. Pymakr lists the following serial devices:
Found 1 serialport
COM3 (Microsoft) (copied to clipboard)

But when setting the address to COM3 in the global settings, and then connecting does not give any response: no command prompt signalling a connection, and even no error message. No response at all!.

I should give up and try atom instead? It is not the first time that pymakr stopped working after a VS code update.

[Paul Thornton]

@jand do you have autoconnect set to true in the global config? Try turning it to false if so. Had a similar issue to this resolved by that.

[jand]

@paul-thornton and @Ralph I realized that my problems started when I started using a newly bought expansion board V3.1. So I now tested both the old expansion board V2.1A, and the new expansion board V3.1. And yes, this confirms that there is a different performance between both. But I also get unexpected results for the old expansion board V2.1A.

I am doing these tests with two laptops: laptop ‘A’ that runs Windows7, and a brand new laptop ‘B’ that runs Windows 10. Both have VS Code 1.31.1 and Pymakr 1.0.6 installed. And I use the same lopy4 for all tests. See my previous posts for the details on the software configurations (as obtained with the ‘about’ button in VS code). These are the results of the tests:

New laptop ‘B’ with expansion board V2.1A:

List serial ports yields:
Found 1 serialport
COM4 (FTDI) (copied to clipboard)

The Autoconnect works (on COM4).
When I disconnect and next do a manual connect (after setting the COM4 port in the address field in pymakr.conf), I still get the reply “>>> Autoconnect: Connecting on COM4...”. And what is even more strange: it does no matter whether I set the address to COM4 or COM3 or COM13 in pymakr.conf! It gets autoconnected. Now and then I get “> Failed to connect (Error: Port is not open). Click here to try again.” Trying again usually gives a successful autoconnect.

New laptop ‘B’ with expansion board V3.1:

List serial ports yields:
Found 1 serialport
COM3 (Microsoft) (copied to clipboard)

The Autoconnect does not work.
When doing a manual connect (after setting the COM port in the address field in pymakr.conf to COM3), nothing happens: no >>> prompt but also no error message. I cannot connect to the board.

Old laptop ‘A’ with expansion board V2.1A:

List serial ports yields:
Found 3 serialports
COM13 (FTDI) (copied to clipboard)
COM5 (Dell)
COM3 (Dell)

The Autoconnect does not work (nothing happens when I connect the board with the pymakr console open).

When I disconnect and next do a manual connect (after setting the COM port in the address field in pymakr.conf to COM13), I still get the reply Connecting on COM13....”. But I don’t get a ‘>>>’ prompt, no connection is established. When I now press the reset button on the lopy4, it gets connected:
ets Jun 8 2016 00:22:57

rst:0x1 (POWERON_RESET),boot:0x12 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
mode:DIO, clock div:1
load:0x3fff8028,len:8
load:0x3fff8030,len:1728
load:0x4009fa00,len:0
load:0x4009fa00,len:14612
entry 0x400a059c
Pycom MicroPython 1.18.2.r1 [v1.8.6-849-e0fb68e] on 2018-11-26; LoPy4 with ESP32
Type "help()" for more information.

So I get a ‘>>>’ prompt. But when I enter a command ‘or just an ‘enter’ at the command prompt, nothing happens: the connection seems to be lost immediately

When I do a manual connect (after setting the COM port in the address field in pymakr.conf to COM3), I get the following result:
Connecting on COM3...

Failed to connect (Error: Port is not open). Click here to try again.

When setting the address to COM5, I get a continuous stream of garbage:
Connecting on COM5...
$GPGGA,,,,,,0,00,0.5,,M,,,,*00
$GPGSV,1,1,00,,,,,,,,,,,,,,,,79
$GPGLL,,,,,235944.000,V,N77
$GPGGA,,,,,,0,0,,,M,,,,*1B
$GPGSA,A,1,,,,,,,,,,,,,,,*1E
$GPGST,235944.000,,,,0,,,74
$GPRMC,235944.000,V,,,,,,,,,,N40
(and this keeps on producing a continuous stream of garbage)

Old laptop ‘A’ with expansion board V3.1:

List serial ports yields:
Found 2 serialports
COM5 (Dell) (copied to clipboard)
COM3 (Dell)

The Autoconnect does not work.

When do a manual connect (after setting the COM port in the address field in pymakr.conf to COM3), I get the following result:
Connecting on COM3...

Failed to connect (Error: Port is not open). Click here to try again.

When setting the address to COM5, I get a continuous stream of garbage:
Connecting on COM5...
$GPGGA,,,,,,0,00,0.5,,M,,,,*00
$GPGSV,1,1,00,,,,,,,,,,,,,,,,79
$GPGLL,,,,,235944.000,V,N77
$GPGGA,,,,,,0,0,,,M,,,,*1B
$GPGSA,A,1,,,,,,,,,,,,,,,*1E
$GPGST,235944.000,,,,0,,,74
$GPRMC,235944.000,V,,,,,,,,,,N40
(and this keeps on producing a continuous stream of garbage)

When setting the address to COM13 (which I know is wrong), I get:
Connecting on COM13...

Failed to connect (Error: Port is not open). Click here to try again.

Since I bought two expansion boards V3.1 at the same time, I also did the tests with the second one. It produces exactly the same results as the first one.

In summary: with the new V3.1 board, I can never establish a connection through USB. With the old V2.1A board I can connect on laptop B: I can connect because the autoconnect works. But with the old V2.1A board I cannot connect on laptop A. And results are quite inconsistent to say the least.

Finally, I also tried the suggestion of Paul Thornton (to set the autoconnect to false in the global config). That does not improve anything. The only effect is that in case of the old V2.1A board on laptop B, it no longer autoconnects, but still allows a manual connect when the address is correctly set to COM4.

[Paul Thornton]

@jand said in Lost PyMakr in VSCode:

Have you modified that LoPy's Firmware at all?

I'm at a loss as to how you're seeing GPS data with no GPS connected. even if the serial connection was noisy and spewing random junk out. there's no way that data would reliably look like GPS data. and it happening more than once in a lifetime are just so astronomically low...

Are these work laptops? Is there any chance a work IT department could have somehow attached an internal gps chip to the laptop(s) that just happens to expose it self as a com port somehow and never mentioned it?.

[jand]

@paul-thornton Yes, these are work laptops. I don't think the IT department has attached an internal chip. But I cannot know that for sure.
I have not modified the LoPy4's firmware since I bought it a few weeks ago.
The random junk is not the only strange thing. Other things are very puzzling too.

I worked with old laptop 'A' with expansion board V2.1A and with an older Lopy4 and a lopy quite intensively over the past year or so (in VS Code with pymakr). It always worked except for one interruption for a few days that was also due to an upgrade of VS code that required an update in pymakr as well).

So i am not sure anymore if the trouble is due to the January update of VS Code or due to the new extension board & Lopy4.