Re: vt: Improve CSI parsing

On Sat, Dec 15, 2018 at 03:34:19PM +0100, Martin Hostettler wrote:

This patch series improves parsing of csi sequences to be more compliant
with current practice.

ECMA-64 defines the format of CSI sequences which allow more characters
than what the vt parser currently accepts. More importantly many of
these characters are used in sequences that more capable terminal
terminal implementations use.

Adjust the parsing of CSI sequences to match xterm* by ignoring all
unknown sequences of the form
   (ESC [)|CSI [\x20-\x3f]*[\x40-\x7e]

This avoids printing unwanted characters when application send valid
sequences not supported by linux either while querying the terminal for
it's identity or when applications print sequences without knowing what
terminal implementation they are connected to (e.g. when connected over
serial lines, android's adb, simple tcp connects, etc)

* and other common terminals

Thanks for these, now queued up.

greg k-h