diff options
author | Joey Hess <joeyh@joeyh.name> | 2017-04-13 15:45:42 -0400 |
---|---|---|
committer | Joey Hess <joeyh@joeyh.name> | 2017-04-13 16:02:59 -0400 |
commit | cf9eb1f6d1dd9a5a52a9604bff60dcbb766825fe (patch) | |
tree | fc15b5203e6eecedc40865569fa0be00a504ef14 /TODO | |
parent | a4a9a6b39a56db45b75da6727d5864a1af4e83ae (diff) | |
download | debug-me-cf9eb1f6d1dd9a5a52a9604bff60dcbb766825fe.tar.gz |
use posix-pty, and forward SIGWINCH
I discovered the posix-pty library while looking for a way to forward
SIGWINCH to the slave pty. It's taken care of a lot of nasty pty
handling details, so let's use it!
Unfortunately, this broke control-d exiting debug-me, which used to work
great. I think that the problem is that readPty never returns B.empty,
even on control-d. Or perhaps, posix-pty is doing something to the pty
that prevents control-d getting through.
I kept all the withoutMode code; at least withoutMode EnableEcho
is still needed.
This commit was sponsored by Jochen Bartl on Patreon.
Diffstat (limited to 'TODO')
-rw-r--r-- | TODO | 2 |
1 files changed, 1 insertions, 1 deletions
@@ -1,6 +1,6 @@ +* control-d does not exit debug-me * Uncommenting the networkDelay in sendPtyOutput exposes a bug in the Activity chain construction. -* set pty size, and forward resizes to the pty (setting new size and SIGWINCH) * potential DOS where developer sends Activity Entered that all refer back to the first Activity Seen. This requires the user side to keep a Backlog containing all later Activity Seen, so uses |