| Summary: | Ctrl-C cygwin/mingw32 | ||
|---|---|---|---|
| Product: | gdb | Reporter: | Jan Smets <jan.smets> |
| Component: | gdb | Assignee: | Not yet assigned to anyone <unassigned> |
| Status: | NEW --- | ||
| Severity: | enhancement | CC: | jojelino, pedro |
| Priority: | P3 | ||
| Version: | 7.5 | ||
| Target Milestone: | --- | ||
| Host: | Target: | ||
| Build: | Last reconfirmed: | ||
| Attachments: | strace | ||
Seems related to 14685 Cygwin strace is unreadable to me. In any case, "set debug remote 1" should show GDB sending a vCont;t packet in response to ctrl-c. That starts out by SIGINT being handled in GDB in handle_remote_sigint, and that setting things up for the event loop to break and call async_remote_interrupt -> target_stop -> ... -> remote_stop_ns. If that isn't working on Cygwin hosts, and vCont;t isn't being sent, someone motivated will need to debug GDB and figure out why not. It was, last I tried (years ago). |
Created attachment 6689 [details] strace (gdb) set target-async on (gdb) set pagination off (gdb) set non-stop on (gdb) set may-insert-fast-tracepoints off (gdb) set may-insert-tracepoints off (gdb) set displaced-stepping off (gdb) set gnutarget a.out-i386 (gdb) target extended-remote 1.2.3.4:2159 Remote debugging using 1.2.3.4:2159 (gdb) (gdb) (gdb) attach 559480756 Attaching to process 559480756 [New Thread 559480756.559480756] 0x07c4e2a6 in ?? () (gdb) c Continuing. <ctrl-c> => nothing happens <window-resize> => thread is stopped. [Thread 559480756.559480756] #1 stopped. A Ctrl-C or (ctrl-break) at this point should suspend the thread again. The ctrl-c is not handled correctly by GDB. In cygwin 1.7 the Ctrl-C is handled correctly after a *window*resize*. See attached strace done on cygwin 1.7. (output taken after the 'continue' has been issued. There are two parts: 1 part is output after ctrl-c , part two is what happened after the window resize.)