Re: multi-tty is broken in CVS HEAD

[Justin Bogner]

Dan Nicolaescu wrote:
> emacs -Q &
> M-x server-start RET
>
> emacsclient -t
>
> after that:
>
> emacsclient -t
>
> from a different terminal does not work anymore. 
>
> I have a build from August 19th where everything was fine.
>
> Given the feature freeze, shouldn't people be more careful and test
> things before checking in?

A change on the 23rd [1] seems to have caused emacsclient -t to break. It was 
just missing some parentheses. A patch like [2] seems to fix it.

[1]
commit 36ba3738efa1eb06a07ea216bca43a9591954819
Author: Eli Zaretskii <address@hidden>
Date:   Sat Aug 23 16:55:52 2008 +0000

    [MSDOS]: Include msdos.h.
    (init_tty) [MSDOS]: Reuse most of WINDOWSNT branch.  Change cpp conditional
    to DOS_NT.  Allow only one call to this function in a session.  Don't
    allocate a new struct tty_display_info; instead, reuse 
the_only_display_info.
    Call get_tty_size to get screen dimensions.  Call init_baud_rate to set
    bad_rate.
    (dissociate_if_controlling_tty) [MSDOS]: Ifdef away function body.
    (Fsuspend_tty) [MSDOS]: Don't close input and output.
    (Fresume_tty) [MSDOS]: Don't reopen the TTY; instead, use stdin/stdout.
    (get_tty_terminal, get_named_tty, Ftty_type)
    (Fcontrolling_tty_p): Handle output_msdos_raw in addition to output_termcap.
    (Fresume_tty, Fsuspend_tty, init_tty, delete_tty): Call
    add_keyboard_wait_descriptor and delete_keyboard_wait_descriptor only when
    subprocesses are supported

[2]
diff --git a/src/term.c b/src/term.c
index 287133a..6303ceb 100644
--- a/src/term.c
+++ b/src/term.c
@@ -2232,7 +2232,7 @@ get_named_tty (name)

   for (t = terminal_list; t; t = t->next_terminal)
     {
-      if (t->type == output_termcap || t->type == output_msdos_raw
+      if ((t->type == output_termcap || t->type == output_msdos_raw)
           && !strcmp (t->display_info.tty->name, name)
           && TERMINAL_ACTIVE_P (t))
         return t;