Centralize thread ID printing

Centralize thread ID printing

Add a new function to print a thread ID, in the style of paddress,
plongest, etc. and adjust all CLI-reachable paths to use it.

This gives us a single place to tweak to print inferior-qualified
thread IDs later:

 - [Switching to thread 1 (Thread 0x7ffff7fc2740 (LWP 8155))]
 + [Switching to thread 1.1 (Thread 0x7ffff7fc2740 (LWP 8155))]

etc., though for now, this has no user-visible change.

No regressions on x86_64 Fedora 20.

gdb/ChangeLog:
2016-01-13  Pedro Alves  <palves@redhat.com>

	* breakpoint.c (remove_threaded_breakpoints)
	(print_one_breakpoint_location): Use print_thread_id.
	* btrace.c (btrace_enable, btrace_disable, btrace_teardown)
	(btrace_fetch, btrace_clear): Use print_thread_id.
	* common/print-utils.c (CELLSIZE): Delete.
	(get_cell): Rename to ...
	(get_print_cell): ... this and made extern.  Adjust call callers.
	Adjust to use PRINT_CELL_SIZE.
	* common/print-utils.h (get_print_cell): Declare.
	(PRINT_CELL_SIZE): New.
	* gdbthread.h (print_thread_id): Declare.
	* infcmd.c (signal_command): Use print_thread_id.
	* inferior.c (print_inferior): Use print_thread_id.
	* infrun.c (handle_signal_stop)
	(insert_exception_resume_breakpoint)
	(insert_exception_resume_from_probe)
	(print_signal_received_reason): Use print_thread_id.
	* record-btrace.c (record_btrace_info)
	(record_btrace_resume_thread, record_btrace_cancel_resume)
	(record_btrace_step_thread, record_btrace_wait): Use
	print_thread_id.
	* thread.c (thread_apply_all_command): Use print_thread_id.
	(print_thread_id): New function.
	(thread_apply_command): Use print_thread_id.
	(thread_command, thread_find_command, do_captured_thread_select):
	Use print_thread_id.