Re: [PATCH 2/3] forkpty, openpty: split functions into new modules

[Bruno Haible]

Hi Eric,

You requested a rapid review:

Eric Blake wrote:
> +  AC_SEARCH_LIBS([forkpty], [util libutil],
> +    [if test "$ac_cv_search_forkpty" != "none required"; then
> +       PTY_LIB="$ac_cv_search_forkpty"
> +     fi])

The second argument [util libutil] can be simplified to [util], no?
Searching for liblibutil sounds useless.

> diff --git a/modules/forkpty b/modules/forkpty
> new file mode 100644
> index 0000000..52dbc49
> --- /dev/null
> +++ b/modules/forkpty
> @@ -0,0 +1,27 @@
> +Description:
> +Provide the forkpty() function.
> +
> +Files:
> +m4/forkpty.m4
> +
> +Depends-on:
> +openpty
> +pty
> ...
> diff --git a/modules/openpty b/modules/openpty
> new file mode 100644
> index 0000000..2a019a5
> --- /dev/null
> +++ b/modules/openpty
> @@ -0,0 +1,26 @@
> +Description:
> +Provide the openpty() function.
> +
> +Files:
> +m4/forkpty.m4
> +
> +Depends-on:
> +pty

I don't understand two things here:
  - Why does 'forkpty' depend on 'openpty'?
  - Since forkpty.m4 is used for both 'forkpty' and 'openpty', it is asymetric
    and a bit misleading to call it 'forkpty.m4'. Why not call it 'pty.m4'?

Bruno