fix control-character parsing in windows

[WindSoilder]

Sorry for introducing a new bug in #619

The bug is: when user enters ctrl-g, the result will be G with modifier CONTROL. And we need to convert from upper case G to lowercase g.

This pr is trying to fix it.

[archseer]

But does this fully fix the issue, what it ctrl shift g is pressed (ctrl-G)? Could some unit tests be added for the various cases?

[superlou]

With the PR:

Event: Key(KeyEvent { code: Char('g'), modifiers: NONE })            <-- g
Event: Key(KeyEvent { code: Char('G'), modifiers: SHIFT })           <-- shift+g
Event: Key(KeyEvent { code: Char('g'), modifiers: CONTROL })         <-- ctrl+g
Event: Key(KeyEvent { code: Char('g'), modifiers: SHIFT | CONTROL }) <-- ctrl+shift+g

Since shift+G returns Char('G'), is ctrl+shift+G also supposed to return Char('G')?

[WindSoilder]

Could some unit tests be added for the various cases?

I'll try to find where to add this, but it may takes some time.

Since shift+G returns Char('G'), is ctrl+shift+G also supposed to return Char('G')?

For this case, I've compared between this version and 0.22, and tried it on Unix, it seems that they all interpreted as Char('G'), when cap lock is on, shift+G returns Char('g'), I can't find an issue talking about this, so I supposed it's the expected behavior?

[TimonPost]

Not sure what has desired behavior there. Seems like that if SHIFT is pressed with other control keys the character should be just capitalized right?

[superlou]

On Linux Mint, I get:

Event: Key(KeyEvent { code: Char('g'), modifiers: NONE })            <-- g
Event: Key(KeyEvent { code: Char('G'), modifiers: SHIFT })           <-- shift+g
Event: Key(KeyEvent { code: Char('g'), modifiers: CONTROL })         <-- ctrl+g
                                                                     <-- ctrl+shift+g
Event: Key(KeyEvent { code: Char('G'), modifiers: SHIFT | ALT })     <-- alt+shift+g

Unfortunately, I don't get ctrl+shift+G, though something in the desktop environment might be catching it. However, alt+shift+g capitalizes the char code. However, the alt+shift+g case makes me exepct ctrl+shift+g should behave similarly. Is there a document or wikipage somewhere defining how crossterm's intent?

[sophiajt]

Maybe it's just me, but I'd expect Char('g') for all cases, and the modifier SHIFT to tell you if it should be capital or not. This treats the keyboard more like the "raw mode" keyboard interface where we want to know what keys were pressed and then interpret those events later on.

[TimonPost]

When pressing SHIFT an uppercase is printed in all applications, we decided in the past that crossterm should also do this. Similarly, with capslock. When capslock is enabled the user will receive SHIFT + capitalized char. However, I do get your point, more or less a preference thing.

[TimonPost]

On ubuntu, it also don't get SHIFT + CONTROL keys. Tho Unix terminals are more limited in their event support. Windows has a easier API for reading input.

[TimonPost]

Thanks.