ui: add some parameters to the cli (mostly regarding eink behaviour) by fdlamotte · Pull Request #3687 · meshcore-dev/MeshCore

fdlamotte · 2026-10-26T19:09:04Z

Some get/set cli commands to tweak display behaviour :

ui.show_clock
ui.auto_off_delay
ui.backlight_delay
ui.backlight_on_keys
ui.backlight_on_msgs
ui.auto_off_blank
ui.auto_off_keep_msgs
ui.max_partial

The two first are common to all displays (but second works only if there is enough room).

Other options are for einks, to control backlight and refresh behaviour

ripplebiz · 2026-10-28T02:37:43Z

+  if (memcmp(command, "set ui.show_clock", 17) == 0) {
+    _prefs.ui_show_clock = atoi(&command[17]) ? 1 : 0;
+    savePrefs();
+    sprintf(reply, "> ui.show_clock is now %d", _prefs.ui_show_clock);


Can you make each of the replies briefer, like most (and the ones further down). eg just "> now %d"

ripplebiz · 2026-10-28T02:37:58Z

+  if (memcmp(command, "set ui.auto_off_delay", 21) == 0) {
+    _prefs.ui_auto_off_delay = atoi(&command[21]);
+    savePrefs();
+    sprintf(reply, "> ui.auto_off_delay is now %d seconds", _prefs.ui_auto_off_delay);


same here...

ripplebiz · 2026-10-28T02:45:22Z


  if (_display != NULL) {
    _display->turnOn();
+    _display->setMaxPartialRefresh(_node_prefs->ui_max_partial);


I notice setMaxPartialRefresh() is always called after every turnOn(), which to me indicates that the EPD driver class should have this in its turnOn() impl. And that the ui_max_partial pref is more like config which gets passed to the display instance somehow (at start, and for simplicity could require reboot to apply).

Yes, ideally I would have grabbed the parameter directly from the prefs, but it is not easily accessible so I found this shortcut to add a method and set it at every turnOn from the loop (not to send it more often). Note that at the moment if the user uses auto_off_delay=0, then the screen is never turned on and the parameter is never applied ... so user needs to turn screen off and on manually for the parameter to be applied.

It could also be sent at start, but I personally don't like to reboot to have parameters set since it forces the user to sync the node after reboot. Tell me what behaviour you prefer and I'll implement it this way ...

About passing parameters, turnOff could also have a blank parameter that could be false by default (instead of having a blankScreen method ... but I didn't want to impact all the display drivers by making the change (maybe it can be done in a subtle way ... I'll see for that)

I've modified DisplayDriver so turnOn and turnOff accept optional parameters, on turnOn you can specify partial_refresh_count and on turnOff you can give the precision wether you want to blank screen or not

akalinin-softvelum · 2026-10-28T03:01:05Z

Tested on my T-Echo (dev + this PR, companion USB build). All ui.* get/set work and survive reboot. Things I noticed:

T-Echo and M8 define EINK_SCREENOFF_BLANK, but NodePrefs checks UI_SCREENOFF_BLANK, so blanking is now off by default on these boards.
ui.max_partial is uint8 without a range check (300 → 44, -1 → 255), and the periodic full refresh can no longer be disabled (-1 did that before). 0 gives a full wipe on every frame. On T-Echo each full wipe blocks the loop ~2.5 s, so I'd like a way to turn it off.
Backlight with the touch button: the light now stays on for ui.backlight_delay after release; before, it was only on while touching. Could 0 mean "only while touched" instead of "always off"?
Minor: set ui.auto_off_delay abc silently sets 0 (= never off); the doc doesn't say 0 means never.

IMHO: on T-Echo the previous behaviour was better: no periodic full refresh, backlight only while touching, no blanking. The full refreshes (with the freeze) and the lingering backlight are really annoying in daily use, so I'd prefer them off by default.

fdlamotte · 2026-10-28T12:15:40Z

Thanks very much for the feedback (feedback is always apreciated)

Tested on my T-Echo (dev + this PR, companion USB build). All ui.* get/set work and survive reboot. Things I noticed:

T-Echo and M8 define EINK_SCREENOFF_BLANK, but NodePrefs checks UI_SCREENOFF_BLANK, so blanking is now off by default on these boards.

ui.max_partial is uint8 without a range check (300 → 44, -1 → 255), and the periodic full refresh can no longer be disabled (-1 did that before). 0 gives a full wipe on every frame. On T-Echo each full wipe blocks the loop ~2.5 s, so I'd like a way to turn it off.

Ok, it should be possible to set full refresh to 0, so it's an error on my side ... will make it int32_t so we can have broader range and negative value (meaning disabled)

I will also disable it by default. You're totally right on this one, periodic full refresh is a counter measure to tearing. I was against it when I lived in europe because I didn't see the point (I had no tearing). Now I leave in a tropical country and none of my eink screen work without full refreshes ;) don't know if it is because of temperature or humidity but I need it to get my screens working ... but this should be an opt-in option, not opt-out (and as we have a doc we can explain how to get the best of one's node)

Backlight with the touch button: the light now stays on for ui.backlight_delay after release; before, it was only on while touching. Could 0 mean "only while touched" instead of "always off"?

Yes, it should probably, because user can disable backlight on keys, so there is no point in making 0 as always off (the thing is I've been asked to

Minor: set ui.auto_off_delay abc silently sets 0 (= never off); the doc doesn't say 0 means never.

IMHO: on T-Echo the previous behaviour was better: no periodic full refresh, backlight only while touching, no blanking. The full refreshes (with the freeze) and the lingering backlight are really annoying in daily use, so I'd prefer them off by default.

Yes, this behaviour was the one I set and the one I liked most when I did the UI for techo, so I can only agree with you

We can make defaults on TEcho work as before, so users are not lost with the new settings ...

fdlamotte · 2026-10-28T14:32:05Z

latest commit addresses cli interface responses (as asked by @ripplebiz) and some behaviour discussed with @akalinin-softvelum

max_partial=0 now disable automatic full-refresh (can be triggered), have not dealt with over and underflowed values ... as user can disable it and real value is displayed once set via cmdline
when backlight_delay=0, backlight button still works (and there is no delay)
backlight_delay=-1 turns backlight always on < -1 display is turned off
defaults should be close to historical behaviour

and it seems I made a mess again in my PR when rebasing it to resolve conflict ... but it should disapear once the PR is pulled ;)

… on cmdline

…tial / blanking)

ripplebiz · 2026-10-29T02:13:49Z

latest commit addresses cli interface responses (as asked by @ripplebiz) and some behaviour discussed with @akalinin-softvelum

max_partial=0 now disable automatic full-refresh (can be triggered), have not dealt with over and underflowed values ... as user can disable it and real value is displayed once set via cmdline

when backlight_delay=0, backlight button still works (and there is no delay)

backlight_delay=-1 turns backlight always on < -1 display is turned off

defaults should be close to historical behaviour

and it seems I made a mess again in my PR when rebasing it to resolve conflict ... but it should disapear once the PR is pulled ;)

thanks, looks ok to me.

fdlamotte force-pushed the display_prefs branch 2 times, most recently from f81c71b to 6129cf0 Compare October 27, 2026 11:51

fdlamotte mentioned this pull request Oct 27, 2026

[Feature request] Keep mesh/BLE serviced while the e-ink display refreshes #3701

Open

ripplebiz reviewed Oct 28, 2026

View reviewed changes

fdlamotte force-pushed the display_prefs branch from 6f6a4bd to f9ec686 Compare October 28, 2026 14:57

fdlamotte added 2 commits October 28, 2026 11:06

ui: add some parameters to the cli (mostly regarding eink behaviour)

d4a6e8a

ui: better behaviour of max_partial and backlight_delay, improvements…

d41f20a

… on cmdline

fdlamotte force-pushed the display_prefs branch from f9ec686 to d41f20a Compare October 28, 2026 15:06

DisplayDriver: adding optional parameters for turnOn/turnOff (max_par…

61fed4e

…tial / blanking)

fdlamotte force-pushed the display_prefs branch from eb285a9 to 61fed4e Compare October 28, 2026 16:00

Navigation Menu

Repository navigation

ui: add some parameters to the cli (mostly regarding eink behaviour) - #3687

ui: add some parameters to the cli (mostly regarding eink behaviour)#3687
fdlamotte wants to merge 3 commits into
meshcore-dev:devfrom
fdlamotte:display_prefs

fdlamotte commented Oct 26, 2026

Uh oh!

ripplebiz Oct 28, 2026

Uh oh!

ripplebiz Oct 28, 2026

Uh oh!

ripplebiz Oct 28, 2026

Uh oh!

fdlamotte Oct 28, 2026

Uh oh!

fdlamotte Oct 28, 2026

Uh oh!

akalinin-softvelum commented Oct 28, 2026

Uh oh!

fdlamotte commented Oct 28, 2026

Uh oh!

fdlamotte commented Oct 28, 2026

Uh oh!

ripplebiz commented Oct 29, 2026

Uh oh!

This branch has not been deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

fdlamotte commented Oct 26, 2026

Uh oh!

ripplebiz Oct 28, 2026

Choose a reason for hiding this comment

Uh oh!

ripplebiz Oct 28, 2026

Choose a reason for hiding this comment

Uh oh!

ripplebiz Oct 28, 2026

Choose a reason for hiding this comment

Uh oh!

fdlamotte Oct 28, 2026

Choose a reason for hiding this comment

Uh oh!

fdlamotte Oct 28, 2026

Choose a reason for hiding this comment

Uh oh!

akalinin-softvelum commented Oct 28, 2026

Uh oh!

fdlamotte commented Oct 28, 2026

Uh oh!

fdlamotte commented Oct 28, 2026

Uh oh!

ripplebiz commented Oct 29, 2026

Uh oh!

This branch has not been deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants