Win32 pseudoconsoles --- ideal interface?

Andy Koppe
Sat Dec 18 04:51:00 GMT 2010

On 18 December 2010 01:14, Daniel Colascione wrote:
> What would an ideal Windows kernel-level pseudoterminal master interface
> look like for Cygwin? (The slave, of course, would be identical to a
> console handle today.)

As long as it faithfully relayed all slave requests rather than
premunging them into a screen buffer I wouldn't particularly care what
it looks like in detail.

One big difference from Unix ptys that would need to be addressed is
that clients get to allocate/free consoles, whereas on Unix it's the
server (e.g. terminal emulator or sshd) that creates ptys and then
passes them along to clients. Therefore there'd need to be the concept
of a console provider or something like that. That would default to
the console we all know and loath, but a process would need to be able
to change the console provider and pass that change on to child

Idle speculation in any case, unless you have a particular reason to ask this?


More information about the Cygwin-developers mailing list