Terminal mode didn't restore back to canonical behaviour
GoogleCodeExporter opened this issue · comments
What steps will reproduce the problem?
1. run the clish executable, in my case: bin/clish -x
xml-examples/klish/etc/clish
2. issue "enable" command following the Prompt, in my case: > enable
3. press ctrl+c to interrupt processing after authentication is required, in my
case: 密码:(ctrl+c goes HERE)>
4. exit clish shell: > exit
What is the expected output? What do you see instead?
In the same terminal, any key pressed should be echo at the terminal, however
nothing happened now whatever press.
What version of the product are you using? On what operating system?
klish-1.6.4 and Fedora 16
Please provide any additional information below.
when I inspect the attribute of the terminal in another terminal, I observed
the following result:
[yrming@localhost klish-1.6.4]$ w
09:39:49 up 34 min, 3 users, load average: 0.01, 0.05, 0.12
USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT
yrming tty2 :0 09:06 34:10 48.80s 0.03s gdm-session-wor
yrming pts/0 :0 09:06 3:34 0.29s 0.29s bash
yrming pts/1 :0 09:20 0.00s 0.11s 0.00s w
[yrming@localhost klish-1.6.4]$ stty -F /dev/pts/0
speed 38400 baud; line = 0;
eol = M-^?; eol2 = M-^?; swtch = M-^?; lnext = <undef>; min = 1; time = 0;
-icrnl ixany iutf8
-icanon -echo
[yrming@localhost klish-1.6.4]$ stty -F /dev/pts/1
speed 38400 baud; line = 0;
eol = M-^?; eol2 = M-^?; swtch = M-^?;
ixany iutf8
Obviously attributes of virtual terminal pts/0 are strange since clish has
exited.
I don't think the issue is critical, however it's a problem, right? Sometimes
people forget the password of root may decide abort the enable command, they
will encounter the problem at last.
FYI and hope it can be fixed. :)
Original issue reported on code.google.com by yrm...@gmail.com
on 12 Mar 2013 at 2:21
I can't reproduce the problem. My su (while 'enable' command) blocks ctrl^c and
do nothing. But i have found a possible cause of broken tty. So try to use the
attached patch and write back if it works. Then i'll commit the patch to
repository.
Original comment by serj.kalichev@gmail.com
on 12 Mar 2013 at 9:40
- Changed state: Started
Attachments:
Hi, Serj:
Thanks for quick response.
After applying the patch, according to my tests, the problem is gone.
BTW, when describing the procedures to reproduce it, I forgot to change my
locale, so when anthentication requires, it printed "����" instead of
"password". Sorry for that.
Yurui
Original comment by yrm...@gmail.com
on 13 Mar 2013 at 2:15
Hi. It was fixed in git repository. Thanks for bug report.
Original comment by serj.kalichev@gmail.com
on 13 Mar 2013 at 7:40
- Changed state: Fixed