Re: [PATCH] Cygwin: Fix putting inferior in foreground (fix input)

[Pedro Alves <pedro@palves.net>]

On 2024-02-27 14:13, Tom Tromey wrote:
>>>>>> "Pedro" == Pedro Alves <pedro@palves.net> writes:
> 
> Pedro> The problem is that inflow.c:child_terminal_inferior is failing to put
> Pedro> the inferior in the foreground, because we're passing down the
> Pedro> inferior's Windows PID instead of the Cygwin PID to Cygwin tcsetpgrp.
> 
> Thanks.  My first thought was that this should really be
> target-specific, but I dug into the code and I see that this is only
> called from inf_child_target -- i.e., it is.  I wonder if a bunch of
> that code could be moved out of inflow.c and into inf-child.c.

I'm not sure.  I think I prefer having all the terminal handling code in
the same place, as the data structures, #ifdefs, etc. are all the same.
Also, there was a suggestion to use terminal shuffling when debugging
with "local gdbserver", and remote isn't an inf_child_target target.

What I think we should do (and I've avoided it thus far because I have a lot of pending
changes to this code), is to rename the file. It is severely misnamed, I think due to
how it ended up being what it is.  At the top you see:

 /* Low level interface to ptrace, for GDB when running under Unix.

That is obviously incorrect.  We have inf-ptrace.c nowadays, and inflow.c has no
ptrace code.  I think what happened is that everything other than terminal handling moved
out of the file over the years.   So I think we could rename it.  Several of the exported
functions from inflow.c are declared in terminal.h, so I guess we could just rename it to
terminal.c.

> 
> Approved-By: Tom Tromey <tom@tromey.com>
> 

Thanks, I've merged it.

Pedro Alves