"^^ " appears as "^^^"

[korbinian90]

Firstly, great work! The layout works generally really well!

One bug that I had is that sometimes typing "^^ " is written as "^^^". It seems to depend on the window I'm typing in. It happens every time in the Telegram app on windows (that's where this debug log is from).

Layout: neo
symmetricalLevel3Modifiers: 0
returnKeyAsMod3R: 0
tabKeyAsMod4L: 0
capsLockEnabled: 1
shiftLockEnabled: 0
level4LockEnabled: 0
qwertzForShortcuts: 1
swapLeftCtrlAndLeftAlt: 0
swapLeftCtrlLeftAltAndLeftWin: 0
supportLevels5and6: 1
capsLockAsEscape: 0
mod3RAsReturn: 0
mod4LAsTab: 0
debugWindow: 1

key down sc 43 vk 0xbf 0x0 0 (M3 right)

key down sc 20 vk 0x54 0x0 0 (A-Z)
injected sc 20 vk 0xdc 0x10 0
injected sc 0 vk 0x20 0x10 0
key up sc 20 vk 0x54 0x80 0 (A-Z)

key down sc 20 vk 0x54 0x0 0 (A-Z)
injected sc 20 vk 0xdc 0x10 0
injected sc 0 vk 0x20 0x10 0
key up sc 43 vk 0xbf 0x80 0 (M3 right)
key up sc 20 vk 0x54 0x80 0 (A-Z)

key down sc 57 vk 0x20 0x0 0
key up sc 57 vk 0x20 0x80 0

[korbinian90]

Hm, it might be because of the weird behaviour of this key "^". Without neo2-llkh, I have to press twice that ^^ appears.

And I found a workaround for me for Telegram. If I press "^ ", then "^^" appears.

[MaxGyver83]

That's strange.

When you type ^ as Mod3+T, neo-llkh sends first ^ and then space because the QWERTZ ^ is a dead key and only after typing/sending space, it becomes visible (as you already know from QWERTZ). When sending space, neo-llkh.exe only uses the virtual key code (=0x20) of the space bar and leaves the scan code empty (= 0). That's this line in the debugging output:

injected sc 0 vk 0x20 0x10 0

Usually that's not a problem, but maybe it is for Chrome. I made a test version that also sets the scan code of the space bar (only when you enter ^). Could you test it, please?
→ https://maximilian-schillinger.de/neo-llkh-caret-text.exe

[korbinian90]

Thanks for compiling it!!
Strangely, it is exactly the same behaviour, no improvement.
But it seem to be restricted to only Telegram.
I would guess that some applications just perform themselvers input manipulation for certain keys...

Einstellungen aus C:\Users\korbi\Desktop\settings.ini:
Layout: neo
symmetricalLevel3Modifiers: 0
returnKeyAsMod3R: 0
tabKeyAsMod4L: 0
capsLockEnabled: 0
shiftLockEnabled: 0
level4LockEnabled: 0
qwertzForShortcuts: 1
swapLeftCtrlAndLeftAlt: 0
swapLeftCtrlLeftAltAndLeftWin: 0
supportLevels5and6: 0
capsLockAsEscape: 0
mod3RAsReturn: 0
mod4LAsTab: 0
debugWindow: 1

key down sc 43 vk 0xbf 0x0 0 (M3 right)

key down sc 20 vk 0x54 0x0 0 (T)
injected sc 20 vk 0xdc 0x10 0
injected sc 57 vk 0x20 0x10 0
key up sc 20 vk 0x54 0x80 0 (T)
key up sc 43 vk 0xbf 0x80 0 (M3 right)

key down sc 43 vk 0xbf 0x0 0 (M3 right)

key down sc 20 vk 0x54 0x0 0 (T)
injected sc 20 vk 0xdc 0x10 0
injected sc 57 vk 0x20 0x10 0
key up sc 20 vk 0x54 0x80 0 (T)
key up sc 43 vk 0xbf 0x80 0 (M3 right)

[thomasgm1971]

hello, the nodead-caret works flawlessly in notepad. in gvim theres a very strange behaviour, sometimes 10 keypresses and not a single caret, sometimes 3 carets per single keypress. in google chrome the character works fine in the address bar, in any edit field in the page area its like gvim. im writing in gmail at the moment, im not able to get a single caret in my message. in the start menu's search field every keypress yields 2 carets and a space. on my computer its not just telegram, its a bigger problem with that character. have a nice day! thomas Am Sa., 15. Aug. 2020 um 15:57 Uhr schrieb Korbinian Eckstein < notifications@github.com>:

…

Thanks for compiling it!! Strangely, it is exactly the same behaviour, no improvement. But it seem to be restricted to only Telegram. I would guess that some applications just perform themselvers input manipulation for certain keys... Einstellungen aus C:\Users\korbi\Desktop\settings.ini: Layout: neo symmetricalLevel3Modifiers: 0 returnKeyAsMod3R: 0 tabKeyAsMod4L: 0 capsLockEnabled: 0 shiftLockEnabled: 0 level4LockEnabled: 0 qwertzForShortcuts: 1 swapLeftCtrlAndLeftAlt: 0 swapLeftCtrlLeftAltAndLeftWin: 0 supportLevels5and6: 0 capsLockAsEscape: 0 mod3RAsReturn: 0 mod4LAsTab: 0 debugWindow: 1 key down sc 43 vk 0xbf 0x0 0 (M3 right) key down sc 20 vk 0x54 0x0 0 (T) injected sc 20 vk 0xdc 0x10 0 injected sc 57 vk 0x20 0x10 0 key up sc 20 vk 0x54 0x80 0 (T) key up sc 43 vk 0xbf 0x80 0 (M3 right) key down sc 43 vk 0xbf 0x0 0 (M3 right) key down sc 20 vk 0x54 0x0 0 (T) injected sc 20 vk 0xdc 0x10 0 injected sc 57 vk 0x20 0x10 0 key up sc 20 vk 0x54 0x80 0 (T) key up sc 43 vk 0xbf 0x80 0 (M3 right) — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#12 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ANPMBUHEQIQWPJ73EDURYXLSA2H4TANCNFSM4P6ENVSQ> .

[MaxGyver83]

@korbinian90 & @thomasgm1971 : Could you check if this issue still occurs in version v0.5.3? It might have been fixed by 51fdd3a.

[thomasgm1971]

hi max, thank you for your great work! pr #26 and pr #12 are gone, everything works as it should. the "problem" is... i tested the behaviour before updating, and the reported malfunctions were gone yet in 0.5.2 if not earlier! my message to pr #12 is from the middle of August, i think it was the 0.3 release still. since then i circumvented the problem, and never again tried typing "^" and "`" in vim and in web formulars. have a nice week, thomas Am So., 31. Jan. 2021 um 15:38 Uhr schrieb Max Schillinger <notifications@github.com>:

…

@korbinian90 & @thomasgm1971 : Could you check if this issue still occurs in version v0.5.3? It might have been fixed by 51fdd3a. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[korbinian90]

I just tested the new version 0.5.3 and I couldn't reproduce the problem. Seems to work perfectly!
Good job and thank you!