Re: [Ada] Namespace pollution in GNAT run-time library

[Geert Bosch <bosch at gnat dot com>]

On Wed, 7 Nov 2001, Florian Weimer wrote:

  On GNU/Linux, the following symbols are defined by the GNAT run-time
  library without a __gnat prefix:
  
  __get_errno __gl_exception_tracebacks __gl_locking_policy
  __gl_main_priority __gl_queuing_policy __gl_task_dispatching_policy
  __gl_time_slice_val __gl_unreserve_all_interrupts __gl_wc_encoding
  __set_errno _gnat_builtin_longjmp c_fileno c_stderr c_stdin c_stdout
  get_char get_int getc_immediate getc_immediate_common
  getc_immediate_nowait gnat_argc gnat_argv gnat_envp gnat_exit_status
  gnat_time mode_append_binary mode_append_binary_plus mode_append_text
  mode_append_text_plus mode_read_binary mode_read_binary_plus
  mode_read_text mode_read_text_plus mode_write_binary
  mode_write_binary_plus mode_write_text mode_write_text_plus
  null_function put_char put_char_stderr put_int put_int_stderr
  seek_end_function seek_set_function
  
  I guess at least the functions without any 'gnat' string in the name
  should be renamed.
  
  Comments?

Yes, I agree this should be fixed. I am not aware of any patches from
you for which review is pending. Could you give an URL for that patch?

When you submit a patch, I will check it and run all regression tests against
it. If that's OK, the patch can go in both branches. There really is not
much difference between the trees, which we want to keep that way.

  -Geert