why do we set prms_id/bug_id in GDB testcases?

[Joel Brobecker <brobecker at adacore dot com>]

Hello,

It's a question that has been nagging at me problably since I decided
that I was going to understand how to write testing with only stuff
that is actually useful, and following the recommended style...

We often start testcases with the following statements:

    set prms_id 0
    set bug_id 0

On rare occcasions, we'll even set an actual PR number, which I assume
is the Mozilla PR number. Does this actually do anything?  The dejagnu
documentation only says that we can use these variables with
conventional meaning:

  prms_id is the PR number in the "official" tracking system;
  bug_id is a bug identification from another organization

If it is useful to set the prms_id (and bug_id), I think we can
automatically default them to 0/0 by doing it in gdb_init.  Since
this function is called at the start of every testcase.  Then a testcase
that wants to set it to something specific can override it.

But for now, I don't seem to be able to find them being used anywhere:

    % find . -type f -print | grep -v ChangeLog | xargs grep prms_id | grep -v 'set prms_id 0$'
    ./gdb.base/hashline3.exp:set prms_id 2360
    ./gdb.base/hashline1.exp:set prms_id 2360
    ./gdb.base/hashline2.exp:set prms_id 2360
    ./gdb.base/reread.exp:set prms_id 13484
    ./gdb.base/structs2.exp:set prms_id 13536
    ./gdb.base/echo.exp:#   ex: set prms_id 643
    ./gdb.cp/classes.exp:    global prms_id
    ./gdb.cp/inherit.exp:    global prms_id
    ./gdb.cp/demangle.exp:    global prms_id
    ./gdb.cp/virtfunc.exp:    global prms_id
    ./gdb.cp/misc.exp:    global prms_id
    ./gdb.cp/cplusfuncs.exp:    global prms_id
    ./gdb.cp/templates.exp:    global prms_id
    ./gdb.cp/gdb2384.exp:set prms_id 2384
    ./gdb.cp/gdb2495.exp:set prms_id 2495
    ./gdb.objc/nondebug.exp:    global prms_id
    ./gdb.objc/objcdecode.exp:    global prms_id
    ./gdb.objc/basicclass.exp:    global prms_id
    ./gdb.threads/step.exp:    global prms_id

Maybe it's used internally by dejagnu? (I couldn't find evidence of that
in the dejagnu manual)

Thoughts?

-- 
Joel