Bugs with Input Command and AHK_H v1.1.28.00 H008

Post AHK_H specific scripts & libraries and discuss the usage and development of HotKeyIt's fork/branch
robodesign
Posts: 349
Joined: 30 Sep 2017, 03:59
Facebook: marius.sucan
GitHub: mariussucan
Location: Romania
Contact:

Bugs with Input Command and AHK_H v1.1.28.00 H008

07 Apr 2018, 17:34

Hello!

Some time ago I reported a bug about Input command not being able to detect the correct letters/keys after a dead key was pressed in Metro/Win10 apps: e.g., `a = à or 'e = é ... and others. I am reffering to AHK-L. The bug was confirmed and fixed some time ago, please see this post:

https://autohotkey.com/boards/viewtopic ... 14&t=42668

However, the problem seems to be still present in AHK_H. Can you please double-check if the specific code/fix was merged into AHK_H ?

Thank you very much .

Best regards, Marius.
Last edited by robodesign on 09 Apr 2018, 16:15, edited 1 time in total.
-------------------------
KeyPress OSD v4: GitHub or forum. (presentation video)
My home page.
User avatar
Drugwash
Posts: 560
Joined: 29 May 2014, 21:07
Location: Ploieşti, Romania

Re: Bugs with Input Command and AHK_H v1.1.28.00 H008

09 Apr 2018, 05:48

A small correction: what you're referring to is not classic AHK but AHK_L. AHK Classic (AKA Basic) is the ANSI-only version 1.0 developed by Chris. ;)

The fix in this commit appears to have been merged here. There has been a subsequent fix to that in AHK_L, following a bug report from fincs, merged in this AHK_H commit.

However, I have noticed a small discrepancy that may or may not affect the behavior in certain context: the code is using the MOD_RSHIFT constant (I'm assuming it's VK_RSHIFT) for certain dead keys, while MSDN states that in functions where left and right modifiers cannot be differentiated, the value for left key is (or should be) used. For safety I think it should always use MOD_LSHIFT (or VK_LSHIFT where appropriate).

As of now, AHK_H has not yet been updated to AHK_L 1.1.28.02 which fixes some issue with WinActivate and restoring active window, which may or may not be related to to the erroneous behavior you reported to me earlier in MM.

You also failed to mention the layout(s) that exhibit the erroneous behavior in regard to dead keys; this may be important since I've already noticed upon testing that some layouts (such as English US International) may return 1 instead of -1 in ToUnicodeEx() for some dead keys even in XP, which screws up the entire detection system. MapVirtualKeyEx() does set the high bit for those some dead keys but its usage adds to the code complexity. Unless I did something completely wrong, that is.
[EDIT] Upon further testing I noticed that MapVirtualKeyEx() does not detect as such the dead keys formed using Shift or AltGr because it doesn't take into account the keyboard state. This makes any alternate detection method unreliable if based on it.
At this point I don't know other reliable method for detecting dead keys in such uncommon layouts, except maybe for directly querying the tagKbdLayer tables in the corresponding DLLs.
[/EDIT]

Hopefully the developers - Lexikos, HotkeyIt, Fincs - would be so kind to share some details on this issue and maybe we could find a fix.
Last edited by Drugwash on 13 Apr 2018, 03:14, edited 1 time in total.
I've deleted my CloudMe account because of GDPR - the now legal base for privacy invasion and data theft.
just me
Posts: 5397
Joined: 02 Oct 2013, 08:51
Location: Germany

Re: Bugs with Input Command and AHK_H v1.1.28.00 H008

09 Apr 2018, 10:36

A small correction, he's not referring to AHK_L but AHK 1.1. ;)
User avatar
Drugwash
Posts: 560
Joined: 29 May 2014, 21:07
Location: Ploieşti, Romania

Re: Bugs with Input Command and AHK_H v1.1.28.00 H008

09 Apr 2018, 10:50

Haha, AHK 1.1 AKA AHK_L. I guess the latter name is better known amongst the "veterans". Still, if you go to official GitHub page and grab the sources or the docs it will be AutoHotkey_L packages. ;)
Anyway, I just pointed that out in case someone saw the "classic" reference and thought "oh no, yet another 9x […] like Drugwash - let's ignore him". :D
I've deleted my CloudMe account because of GDPR - the now legal base for privacy invasion and data theft.
robodesign
Posts: 349
Joined: 30 Sep 2017, 03:59
Facebook: marius.sucan
GitHub: mariussucan
Location: Romania
Contact:

Re: Bugs with Input Command and AHK_H v1.1.28.00 H008

09 Apr 2018, 16:16

Thank you for the correction. Yes, you are right.

The report applies to any layout with dead keys. Windows metro apps.
-------------------------
KeyPress OSD v4: GitHub or forum. (presentation video)
My home page.
User avatar
Drugwash
Posts: 560
Joined: 29 May 2014, 21:07
Location: Ploieşti, Romania

Re: Bugs with Input Command and AHK_H v1.1.28.00 H008

10 Apr 2018, 03:45

So, to sum it up: you're basically saying that, running the same simple test script, and selecting any keyboard layout that has dead keys, under current AHK_L it works correctly showing the right dead keys and combinations in both regular applications and UWP apps, while under current AHK_H it fails only in UWP apps just as it did in AHK_L before the 1.1.27.05 and 1.1.27.06 fixes were applied? Not distrusting you or anything, just trying to make it clear for myself and any developers reading this.
I've deleted my CloudMe account because of GDPR - the now legal base for privacy invasion and data theft.
robodesign
Posts: 349
Joined: 30 Sep 2017, 03:59
Facebook: marius.sucan
GitHub: mariussucan
Location: Romania
Contact:

Re: Bugs with Input Command and AHK_H v1.1.28.00 H008

10 Apr 2018, 06:57

Yes, exactly:

This code:

Code: [Select all] [Download] GeSHi © Codebox Plus

loop {
Input, InputChar, L1 B V E I, {tab}
tooltip % InputChar
}


In AHK_H v1.1.28.00 H008 ... does not show the keys resulted from dead key combinations in UWP apps [Win10/Metro apps]. This bug was fixed in recent AHK_L versions.

Best regards, Marius.
-------------------------
KeyPress OSD v4: GitHub or forum. (presentation video)
My home page.
robodesign
Posts: 349
Joined: 30 Sep 2017, 03:59
Facebook: marius.sucan
GitHub: mariussucan
Location: Romania
Contact:

Re: Bugs with Input Command and AHK_H v1.1.28.00 H008

10 May 2018, 16:25

Where's HotKeyIt? I see no new activity from him here, on the forum, for the past month.

Is ahkH abandoned? I'm asking because.... It hasn't been updated to reflect the new version of ahkL.... I hope this is not the case...

I hope he's doing well.

Best regards, Marius.
-------------------------
KeyPress OSD v4: GitHub or forum. (presentation video)
My home page.

Return to “AutoHotkey_H”

Who is online

Users browsing this forum: No registered users and 1 guest