On 14 May 2013 01:23, Jack Adrian Zappa wrote: > On Mon, May 13, 2013 at 5:45 PM, Andy Koppe wrote: >> >> On 13 May 2013 22:35, Adrian H wrote: >> > I inadvertently dumped a binary stream to the terminal and it froze >> > mintty. When I tried to kill the process dumping the data, it >> > wouldn't die, even when using -9 switch. When I used Process Explorer >> > to kill it, it died and mintty resumed working. >> >> http://cygwin.com/ml/cygwin/2013-04/msg00362.html > > > Sounds like its related, however, if I only send 369 of the 370 lines, > it is fine and the characters sent to the terminal end up on my > command line. I.e. I get this on my command line after the cat of 369 > lines. > > $ 1;2c1;2c1;2c > > Which is part of the return string sans the leading "\x5b\x3f" a.k.a. > "\e[".
The "\e[" is swallowed by bash/readline. > According to that link, you provided, it sounds like only one > key is necessary to cause this problem. It also requires a large amount of output from the application, so that the pty's output buffer overflows and hence the application's write() to it blocks. > In any case, how does Linux deal with this issue? I don't know. A buffer on the terminal's side of the pty? Automatic deadlock detection in the kernel, causing one of the blocking writes to fail? Andy -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
