Mark Johnson
2008-May-29 15:59 UTC
Fwd: fix for pygrub not accepting cursor keys, after CS#15928
-------- Original Message -------- Subject: fix for pygrub not accepting cursor keys, after CS#15928 Date: Thu, 29 May 2008 08:56:23 -0700 (PDT) Author: jk@tools.de Repository: /hg/xen-gate/devel-unstable-patches Latest revision: b830a8d1e8a2c0fd060bf43f91e18ff619755575 Total changesets: 1 Log message: fix for pygrub not accepting cursor keys, after CS#15928 CS#15928 added some NetBSD fixes for xenconsole, xenconsole is now using init_term() on the slave side of the xenconsole<->xend pseudo terminal. init_term() is using tcsetattr with TCSAFLUSH, and that did flush part of pygrub''s vt100 terminal initialization from the pty''s buffer. The vt100 terminal emulation is supposed to run in "vt100 application key mode" mode while the pygrub curses program is running, but was actually running with vt100 application key mode disable because of the missing initialization sequence. For that reason the cursor keys didn''t work as expected. We''re now using TCSANOW instead of TCSAFLUSH, so that the pty''s buffer isn''t flushed when switching to/from raw mode. And now that the NetBSD fix from CS#15928 has added raw mode setup for the slave side of the pty, avoid doing it a second time in the Solaris specific code. Files: update: pty-fixes