Replace hidecursorinactive with hook commands

[joelim-work]

Related issues

• Fixes Feature request: highlighting of cursors when inactive #1757
• Fixes r32 prints ^[[I or I in libvte-based terminals #1673

Description

This PR replaces the hidecursorinactive option introduced in #965 with the more generic on-focus-gained and on-focus-lost hook commands, allowing for greater customization for when the terminal gains or loses focus. This also properly solves #1673 by providing an option focusreport to explictly control terminal focus reporting only enabling terminal focus reporting if either of the hook commands are defined.

Examples

Hide the cursors completely when the terminal loses focus (same effect as set hidecursorinactive true):

# hook command which is invoked when the terminal gains focus
cmd on-focus-gained :{{
    set cursorparentfmt "\033[7m"
    set cursoractivefmt "\033[7m"
    set cursorpreviewfmt "\033[4m"
}}

# hook command which is invoked when the terminal loses focus
cmd on-focus-lost :{{
    set cursorparentfmt ""
    set cursoractivefmt ""
    set cursorpreviewfmt ""
}}

# set initial values upon startup
on-focus-gained

Dim the cursors when the terminal loses focus:

# hook command which is invoked when the terminal gains focus
cmd on-focus-gained :{{
    set cursorparentfmt "\033[7m"
    set cursoractivefmt "\033[7m"
    set cursorpreviewfmt "\033[4m"
}}

# hook command which is invoked when the terminal loses focus
cmd on-focus-lost :{{
    set cursorparentfmt "\033[7;2m"
    set cursoractivefmt "\033[7;2m"
    set cursorpreviewfmt "\033[4;2m"
}}

# set initial values upon startup
on-focus-gained

[DusanLesan]

Very nice solution for cursorparentfmtinactive. Would it be possible to avoid explicitly setting focusreport by checking if the user has either on-focus-gained or on-focus-lost set?

[joelim-work]

Would it be possible to avoid explicitly setting focusreport by checking if the user has either on-focus-gained or on-focus-lost set?

I think this is debatable. On one hand it's nice to avoid adding yet another option (lf has too many), but on the other hand it also makes the code rather 'clever' (one can argue that cmd should only define commands, not do anything else like enabling/disabling focus reporting).

For now I think I will go with your suggestion as it isn't too complex to implement.

[DusanLesan]

I think this is a better way than needing to enable focusreport. Similar to on-redraw, the user can extend app functionality just by adding the appropriate function. In the case of focus hooks, the app can simply skip observing these events if the user has not set callbacks. This way, the user does not need to worry about the app's inner behavior, as the app can determine if it is needed or not by itself.

[joelim-work]

No worries, I have already implemented this in 6fa8aee.