zyedidia / tcell Goto Github PK

It was brought to my attention that certain key combinations aren't recognized when running micro in foot.

Instead of triggering a key binding, the escape is printed in the terminal:

1 hello world[1;6D]

Examples of non-working key combinations:

• ctrl+shift+arrow
• alt+arrow
• ctrl+alt+arrow
• ctrl+shift+home.

shift+arrow works, as does ctrl+arrow and ctrl+home.

All of these work if I set TERM=xterm-256color, or even TERM=alacritty.

Foot's terminfo is very similar to xterm's. It does expose all the modifier combinations for e.g. right:

 infocmp -x1 foot|grep kRIT
        kRIT=\E[1;2C,
        kRIT3=\E[1;3C,
        kRIT4=\E[1;4C,
        kRIT5=\E[1;5C,
        kRIT6=\E[1;6C,
        kRIT7=\E[1;7C,

We can also look at the diff between xterm-256color and foot:

% infocmp -x1 xterm-256color foot
comparing xterm-256color to foot.
    comparing booleans.
        OTbs: T:F.
        mc5i: T:F.
    comparing numbers.
    comparing strings.
        flash: '\E[?5h$<100/>\E[?5l', '\E]555\E\\'.
        kbs: '^H', '^?'.
        kmous: '\E[<', '\E[M'.
        mc0: '\E[i', NULL.
        mc4: '\E[4i', NULL.
        mc5: '\E[5i', NULL.
        meml: '\El', NULL.
        memu: '\Em', NULL.
        mgc: '\E[?69l', NULL.
        oc: '\E]104\007', '\E]104\E\\'.
        rs1: '\Ec\E]104\007', '\Ec'.
        setab: '\E[%?%p1%{8}%<%t4%p1%d%e%p1%{16}%<%t10%p1%{8}%-%d%e48;5;%p1%d%;m', '\E[%?%p1%{8}%<%t4%p1%d%e%p1%{16}%<%t10%p1%{8}%-%d%e48\:2\:5\:%p1%d%;m'.
        setaf: '\E[%?%p1%{8}%<%t3%p1%d%e%p1%{16}%<%t9%p1%{8}%-%d%e38;5;%p1%d%;m', '\E[%?%p1%{8}%<%t3%p1%d%e%p1%{16}%<%t9%p1%{8}%-%d%e38\:2\:5\:%p1%d%;m'.
        smglr: '\E[?69h\E[%i%p1%d;%p2%ds', NULL.
        Cr: '\E]112\007', '\E]112\E\\'.
        Cs: '\E]12;%p1%s\007', '\E]12;%p1%s\E\\'.
        Ms: '\E]52;%p1%s;%p2%s\007', '\E]52;%p1%s;%p2%s\E\\'.
        Se: '\E[2 q', '\E[ q'.
        Sync: NULL, '\EP=%p1%ds\E\\'.
        XM: '\E[?1006;1000%?%p1%{1}%=%th%el%;', NULL.
        xm: '\E[<%i%p3%d;%p1%d;%p2%d;%?%p4%tM%em%;', NULL.

At least I cannot see anything in there that would/should/could make a difference here. But perhaps there is something in tcell's handling of "builtin" terminfos vs. external ones?

I don't know if this is a tcell or micro bug, anyways, I added a messenger.Message(button) line to all the cases in switch button. Scrolling up and down get interpreted as 256 and 512 respectively, however, when scrolling down and getting a 256, and then moving the cursor it will actually send a 2. When scrolling up, and then moving the cursor will send a 1, which is really annoying, because it will activate a selection. This also happense when scrolling -> moving the cursor -> scrolling again (which can be seen in the webm I linked below)

Here is a webm of the the bug in action, look at the message bar for the button codes. (keep in mind I didn't press the left or middle button (aka button1 and button2) once in this video, the part where it quickly switches between 1 and 2 is actually me just scrolling up and down): https://sr.ht/m_Rw.webm

in my opinion https://github.com/zyedidia/tcell/blob/master/tscreen.go#L870-L881 is most likely to somehow cause this issue, I have no idea though.

I just removed these lines https://github.com/zyedidia/tcell/blob/master/tscreen.go#L870-L881 entirely, and it still seems to happen, this makes me think it has something to do with motion.

#1 fixes the spawning of button1/2 events after scrolling, however, it also disables the button1/2 when clicking, so yeah...

I've been changing 100s of values in tscreen.go, but not luck fixing this issue as of yet. I think something in parseXtermMouse() is causing this. Changing:

        case 3:
            btn = int(b[i])
            state++

to

        case 3:
            btn = int(b[i]) - 32
            state++

Makes it spawn additional scroll events on motion after scrolling, instead of button1/2 events, so I think changing something in case 3 might fix this somehow...

Does this fork implement the text selection feature?
like issue: gdamore#273

I got some error when I tried to build termbox/compat.go:

# github.com/zyedidia/tcell/termbox
./compat.go:305:2: impossible type switch case: tev (type tcell.Event) cannot have dynamic type *tcell.EventInterrupt (missing EscSeq method)
./compat.go:351:42: cannot use tcell.NewEventInterrupt(nil) (type *tcell.EventInterrupt) as type tcell.Event in argument to screen.PostEvent:
	*tcell.EventInterrupt does not implement tcell.Event (missing EscSeq method)

this is related to zyedidia/micro#1854 (I was testing the solution and it wouldn't launch, I got to this point via delve)

tl;dr if t.TrueColor will segfault if t is nil
the check for t == nil should be moved up further

however, the secondary issue here is that it fails to detect the foot terminal, despite infocmp being present/installed (admittedly in ~/.terminfo, which is standard (see terminfo(5)'s "Fetching Compiled Descriptions")
I couldn't find (in a reasonable timeframe) exactly where infocmp (or anything similar) is invoked, but this is clearly differing behavior (esp. since infocmp is expected to be installed)

Hi, I found this through micro's go.mod and I'm investigating the library right now -- is there any reason for this fork? It looks like from the diff that you're adding new features/fixing bugs -- but there's no plan to send those upstream? Should this fork be used for production over the main repo (despite the main repo having had more recent releases/commits, and is also still in development)?

I have micro running inside tmux with set clipboard terminal
When I press Ctrl-v micro crashes with the following output:

panic: runtime error: slice bounds out of range [7:6]

goroutine 19 [running]:
github.com/zyedidia/tcell/v2.(*tScreen).parseOSC52Paste(0xc0002b4d80, 0xc00052ef00, 0xc00052edf8, 0xc00025a140)
 github.com/zyedidia/tcell/[email protected]/tscreen.go:1402 +0x67d
github.com/zyedidia/tcell/v2.(*tScreen).collectEventsFromInput(0xc0002b4d80, 0xc00052ef00, 0x7ff907197300, 0x0, 0x0, 0x0)
 github.com/zyedidia/tcell/[email protected]/tscreen.go:1486 +0x174
github.com/zyedidia/tcell/v2.(*tScreen).scanInput(0xc0002b4d80, 0xc00052ef00, 0xf55300)
 github.com/zyedidia/tcell/[email protected]/tscreen.go:1452 +0x45
github.com/zyedidia/tcell/v2.(*tScreen).mainLoop(0xc0002b4d80)
 github.com/zyedidia/tcell/[email protected]/tscreen.go:1616 +0x32f
created by github.com/zyedidia/tcell/v2.(*tScreen).Init
 github.com/zyedidia/tcell/[email protected]/tscreen.go:207 +0x6ad

And if I don't have anything in paste buffer micro crashes with:

panic: runtime error: index out of range [0] with length 0

goroutine 19 [running]:
github.com/zyedidia/tcell/v2.(*tScreen).parseRune(0xc000234d80, 0xc00005df00, 0xc00005ddf8, 0xc000140001)
 github.com/zyedidia/tcell/[email protected]/tscreen.go:1317 +0x7ac
github.com/zyedidia/tcell/v2.(*tScreen).collectEventsFromInput(0xc000234d80, 0xc00005df00, 0x7f366abadb00, 0x0, 0x0, 0x0)
 github.com/zyedidia/tcell/[email protected]/tscreen.go:1498 +0x21e
github.com/zyedidia/tcell/v2.(*tScreen).scanInput(0xc000234d80, 0xc00005df00, 0xf55300)
 github.com/zyedidia/tcell/[email protected]/tscreen.go:1452 +0x45
github.com/zyedidia/tcell/v2.(*tScreen).mainLoop(0xc000234d80)
 github.com/zyedidia/tcell/[email protected]/tscreen.go:1616 +0x32f
created by github.com/zyedidia/tcell/v2.(*tScreen).Init
 github.com/zyedidia/tcell/[email protected]/tscreen.go:207 +0x6ad

Also set -s set-clipboard on is needed in .tmux.conf to enable full clipboard support for applications inside tmux.

However they can be made to work by setting the ENABLE_LVB_GRID_WORLDWIDE 0x0010 flag on initialization. I want to run micro from the Windows console so a fix for this would be nice. See also gdamore#627 .