AutoHotkey Community

Here are the changes for v1.0.42.01:

Fixed the following variables to work correctly when concatenated onto other strings: A_DetectHiddenWindows, A_DetectHiddenText, A_AutoTrim, and A_StringCaseSense. [thanks jballi]

Fixed KeyWait and GetKeyState in the subroutines of hook hotkeys (broken by 1.0.39) [thanks Laszlo & TobStro]. For example:

Here are the changes for v1.0.42.02:

Fixed UpDown controls to snap onto the right control when inside a Tab control that contains a ListView. [thanks Thalon]

Improved the Hotkey command with an "Off" option (typically used to create a hotkey in an initially-disabled state).

Improved A_Cursor not to interfere with double-clicking. [thanks Shimanov]

At last, the A_Cursor limitation is fixed

_________________
GitHub • Scripts • IronAHK • Contact by email not private message.

Here are the changes for v1.0.42.03:

Fixed crash of WinActivate in certain rare circumstances. [thanks twhyman]

Changed hotstrings to require the mouse hook by default (see next item). This can be prevented via #Hotstring NoMouse.

Improved hotstrings to take into account mouse clicks. This allows a hotstring to fire when the user's click focuses a new control or moves the caret.

Improved the Random command with a means to set a new seed. [thanks Laszlo]

Improved #ClipboardTimeout to reattempt data reading when the first attempt fails. Previously, only the opening of the clipboard was reattempted.

Added built-in variable A_LastError for DllCall and Run/RunWait. It contains the result from the OS's GetLastError(). [thanks Shimanov and PhiLho]

Does this mean that GetLastError() will be called after each call? If yes, could this be made optional?

According to benchmarks, calling GetLastError() does not measurably impact performance (since the difference couldn't be measured, it's probably below 0.1%). I benchmarked it with GetTickCount because it's known to be a very fast function:


The result after the change was the same as it was before. Did you have some concern other than performance?

Although performance was the main concern, it seems a bit unnecessary to add an extra call every time.

Calling GetLastError() unconditionally bothered me a bit too because it clearly isn't useful for most functions. However, even a non-API function can call SetLastError() as a means of returning extra info to its caller, so there is some value added there.

Also, it seems likely that GetLastError() is implemented a lot like GetTickCount() in the sense that it instantly returns a value from a fixed memory address. Such functions have the fastest possible execution time, almost as fast as a single line of inline code. In fact, the compiler might translate such functions into inline code (on the other hand, maybe it only does that for C-library functions like strlen).

Eh, sometime I wait for upgrading, skipping some minor releases, but this time, I run to download the latest version!
Thank you Chris.
I guess the first scripts will have to test for emptiness of the variable, for compatibility with old versions of AHK.

_________________
vPhiLho := RegExReplace("Philippe Lhoste", "^(\w{3})\w*\s+\b(\w{3})\w*$", "$1$2")

Here are the changes for v1.0.42.04:

Fixed ClipboardAll when used with apps such as Word and Excel (broken by previous version). [thanks Roussi Nikolov]

Fixed ClipboardAll to prevent appearance of bookmarks in MS Word. [thanks Laszlo & 02clams]

Fixed A_TimeIdlePhysical being disrupted by simulated AltGr keystrokes. [thanks skrommel]

Fixed Send{Blind} causing sticking Win/Ctrl/Alt/Shift when the user released such a key in the middle of a Send.

Improved the Send command not to press and release the shift key after each uppercase letter (e.g. Send ABC).

Improved SoundSet/Get to support ComponentType "Headphones". [thanks Tobias Winkler]


I didn't get much coding done lately because the rehosting of autohotkey.com took a lot of time. Hopefully that's almost all taken care of now.

v1.0.42.05 is functionally identical to the previous release, whose EXE and BIN were about 1% larger than they should have been due to a compiler quirk.

Here are the changes for v1.0.42.06:

Fixed crash of A_ScriptDir (broken by 1.0.42.01).

Fixed Run/RunWait's passing of an extra space character at the end of the parameter list when launching shortcuts. [thanks arbe]

In v1.0.42.07, a crash of functions that recursively pass ByRef parameters was fixed. [thanks PhiLho]