This is the mail archive of the gdb-prs@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug server/19855] readline abort when debugging with rr


https://sourceware.org/bugzilla/show_bug.cgi?id=19855

--- Comment #10 from Pedro Alves <palves at redhat dot com> ---
(In reply to Yichao Yu from comment #9)
> OK, finally back to this issue. A few things that I discovered,

Thanks.

> 3. There are calls to `remote_terminal_inferior` however
> `target_async_permitted` is always `0` so it never does anything.

>     The value is set to `0` with bt

>         command=0x148f8b0 "maint set target-async 0") at event-top.c:463

Ah...  Is there a reason RR disables target-async?  Users are not supposed to
need this.  It's an escape hatch for gdb bugs, so I imagine it was done to
paper over some old GDB bug?  Changing this setting affects the target backend
/ event-loop integration, which affects stdin/readline handling -- different
code paths.

The !target-async paths are getting less tested as time progresses.  All 
hosts except DJGPP do target async when remote debugging by default, I
believe...

>     Is this something caused by rr?

So I'd think that stopping RR from using the "maint set target-async 0" command
would make this go away.

Did you try the users/palves/immediate_quit branch?  It may also fix this
since https://sourceware.org/ml/gdb-patches/2016-03/msg00358.html removes the
target_async_permitted checks from
remote_terminal_inferior/remote_terminal_ours.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]