14:39 <st3r4g> weird question: can a terminal emulator ever run on a kernel with CONFIG_TTY=n?

14:39 <st3r4g> I guess the answer is no

14:45 <alice> i believe you need that for ptty's too

15:03 <dnkl> st3r4g: it's been a long time since I built my own kernels, but I'm with alice here...

15:04 <dnkl> st3r4g: https://www.linusakesson.net/programming/tty/

15:47 <lanodan> I'm quite wondering why is there even this option if ptty aren't possible with CONFIG_TTY=n

15:48 <lanodan> That said CONFIG_TTY description says it's for support of text terminals and serial ports.

15:51 <lanodan> Ah CONFIG_UNIX98_PTYS (provider for /dev/ptmx and /dev/pts/ ) depends on CONFIG_TTY, so yeah, basically a broken system without CONFIG_TTY.

15:59 <alice> embedded things don't need tty i don't think

15:59 <alice> that's why it would be there at all

16:00 <st3r4g> yeah for desktop TTY=n is too extreme. but VT=n should be doable

16:02 <alice> seems like a pointless thing to disable

16:02 <st3r4g> basically you replace the linux console with something in userspace

16:47 <st3r4g> hm, with VT=n posix_openpt() fails, even though UNIX98_PTYS=y

16:51 <dnkl> st3r4g: is this something you could test?

16:51 <dnkl> int

16:51 <dnkl> posix_openpt(int flags)

16:51 <dnkl> {

16:51 <dnkl> return open("/dev/ptmx", flags);

16:51 <dnkl> }

16:51 <dnkl> (from posix_openpt(3))

17:05 <st3r4g> nevermind I didn't have devpts mounted, now it works

17:05 <dnkl> nice!