Re: 1.7.5 cannot bg mintty
On 30 May 2010 07:00, Christopher Faylor wrote: > On Sat, May 29, 2010 at 11:32:23PM +0100, Andy Koppe wrote: >>There's also another issue with the ^Z not taking effect immediately >>due to signal handling being disabled while mintty waits for Windows >>events, but that would require some fairly major restructuring to fix, >>so leaving that for later. Fixed in mintty 0.8-beta1, now available from mintty.googlecode.com. >> Looking at rxvt, the right thing to do is a >>select() on the pty device and /dev/windows instead of mintty's >>current approach using multiple threads and >>MsgWaitForMultipleObjects.) > > ...assuming that still works. I wonder if anyone has used that > interface in the last few years. > > But, then, I dunno. Maybe rxvt uses it. Yep, rxvt does use /dev/windows; I double-checked via /proc//fd. XWin uses it too. It allowed getting rid of quite a bit of complexity in mintty's I/O handling code. 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
Re: 1.7.5 cannot bg mintty
On Sat, May 29, 2010 at 11:32:23PM +0100, Andy Koppe wrote: >(There's also another issue with the ^Z not taking effect immediately >due to signal handling being disabled while mintty waits for Windows >events, but that would require some fairly major restructuring to fix, >so leaving that for later. Looking at rxvt, the right thing to do is a >select() on the pty device and /dev/windows instead of mintty's >current approach using multiple threads and >MsgWaitForMultipleObjects.) ...assuming that still works. I wonder if anyone has used that interface in the last few years. But, then, I dunno. Maybe rxvt uses it. cgf -- 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
Re: 1.7.5 cannot bg mintty
Paul wrote: > I sometimes forget to append a & when opening a new mintty from an existing > mintty. With rxvt this was never a problem, just ^Z and bg in the parent > terminal (my shell is bash 3.2.49-23). This doesn't seem to work for mintty > (0.6.2-1) though - although it seems that the job has has been put into the > background it never becomes possible to type anything into the newly spawned > terminal window after ^Z has been used. > > If I spawn an rxvt from a mintty (or indeed an rxvt from an rxvt) using ^Z > and bg works without any problems. I can't see anything obviously different > in the output of stty -a between my rxvt and mintty. > > If I start a mintty with a &, fg it and then bg it again, the same effect > occurs - it becomes impossible to type anything into the child mintty. I fixed this in svn, to appear in mintty-0.7. The thread for reading output from the child process wrongly exited when the read() call got interrupted by a signal such as SIGSTOP. (There's also another issue with the ^Z not taking effect immediately due to signal handling being disabled while mintty waits for Windows events, but that would require some fairly major restructuring to fix, so leaving that for later. Looking at rxvt, the right thing to do is a select() on the pty device and /dev/windows instead of mintty's current approach using multiple threads and MsgWaitForMultipleObjects.) 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
Re: 1.7.5 cannot bg mintty
On 18/05/2010, cy.20.superconductor wrote: > I sometimes forget to append a & when opening a new mintty from an > existing mintty. With rxvt this was never a problem, just ^Z and bg in > the parent terminal (my shell is bash 3.2.49-23). This doesn't seem to > work for mintty (0.6.2-1) though - although it seems that the job has > has been put into the background it never becomes possible to type > anything into the newly spawned terminal window after ^Z has been used. > > If I spawn an rxvt from a mintty (or indeed an rxvt from an rxvt) using > ^Z and bg works without any problems. I can't see anything obviously > different in the output of stty -a between my rxvt and mintty. > > If I start a mintty with a &, fg it and then bg it again, the same > effect occurs - it becomes impossible to type anything into the child > mintty. > > I can an alias to avoid tripping myself up on this all the time, but it > seemed worth querying. Is there something I'm missing or a setting > needed somewhere to make mintty background-able (or indeed does anyone > else experience this behaviour?) Thanks for the report, I'll look into this when I get back from holiday. Btw, you can also start a new mintty using Alt+F2. 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
1.7.5 cannot bg mintty
I sometimes forget to append a & when opening a new mintty from an existing mintty. With rxvt this was never a problem, just ^Z and bg in the parent terminal (my shell is bash 3.2.49-23). This doesn't seem to work for mintty (0.6.2-1) though - although it seems that the job has has been put into the background it never becomes possible to type anything into the newly spawned terminal window after ^Z has been used. If I spawn an rxvt from a mintty (or indeed an rxvt from an rxvt) using ^Z and bg works without any problems. I can't see anything obviously different in the output of stty -a between my rxvt and mintty. If I start a mintty with a &, fg it and then bg it again, the same effect occurs - it becomes impossible to type anything into the child mintty. I can an alias to avoid tripping myself up on this all the time, but it seemed worth querying. Is there something I'm missing or a setting needed somewhere to make mintty background-able (or indeed does anyone else experience this behaviour?) Thanks, Paul -- 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