gdb/testsuite/TODO

   1 The highest priority item is not on this list: Fix bugs in the
   2 existing testsuite, fix the GDB/compiler/shell/etc bugs which it
   3 detects (particularly when they are hard to XFAIL), make it run
   4 reliably without unexpected failures on the "standard" machines, etc.
   5 This list exists largely as "tests we can add when we are ready to
   6 risk destabilizing it again".
   7
   8 return.exp--reenable test (there is nothing known to be wrong with the
   9 test, but it hasn't been tried on a wide range of architectures).
  10
  11 nodebug.exp--test printing variables.
  12
  13 Get crossload tests to use --with-targets and reenable them.
  14
  15 corefile.exp:
  16 1.  Print variables from the core file (data and stack), and text
  17 (from the exec file).  This tests whether the corefile sections are
  18 mapped to the right addresses.
  19 2.  Test what happens when we get a new exec file without explicitly
  20 getting rid of the core file (we at least must avoid core dumps and such).
  21 3.  Test backtrace in corefile.exp.
  22 4.  Test ability to run program when there is a core target, then go
  23 back to the core file when the program exits.
  24
  25 Test handling of floating point variables
  26 1.  float, double, or long double
  27 2.  in register or saved register or memory.  Also the case where a
  28 double is in two float registers and only one of them is saved.
  29 3.  print them or set them
  30 4.  (Alpha) integer (32 or 64 bit) in floating point register.
  31
  32 Print registers--"p $r5", "p sizeof ($r5)".  Test that they print
  33 appropriately (integer registers in decimal, registers which always
  34 contain addresses (pc, probably sp and fp, maybe others) in hex,
  35 floating point).
  36
  37 Test "info line" with all kinds of linespecs.  Test that the last line
  38 of the file works right.
  39
  40 weird.exp--test that unrecognized cross-reference types or
  41 unrecognized visibility or virtual characters get skipped properly
  42 (see stabs.texinfo).
  43
  44 Test "jump" including jump to a breakpoint (the latter will need an
  45 xfail for UDI and probably VxWorks (PR 1786 for vxworks; PR 2416
  46 contains some info for 29k).
  47
  48 Test C++ nested types (especially if PR 1954 is fixed; even if not
  49 *some* things already should work even in the presence of nested
  50 types).  Test classes nested more than 9 levels deep (g++ mangles
  51 these differently) (both a demangle test and some tests which also
  52 test the compiler).  Test calling a method of a class nested more than
  53 9 levels (for gdb_mangle_name and demangling).
  54
  55 Test that printing const-qualified versions of various types works.
  56 In particular, on the sparc and probably other machines, "double" is
  57 handled differently from most types because it requires more alignment
  58 and thus goes in a different section (there is a gcc 2.4.5 bug with
  59 "const double" on sparc).
  60
  61 Test that GDB's "source" command works and that things work if stdin
  62 is redirected (to a file or a pipe).  Test user defined command.  Run
  63 an inferior each of these ways (to test that inflow.c works).  Test
  64 that GDB works if the last line of stdin or a source'd file lacks a
  65 newline.
  66
  67 Test that module__2do (for example) in a C program does not get
  68 demangled.
  69
  70 Test that unmatched single quotes produce error messages, both in
  71 expressions and linespecs.
  72
  73 Test "cd".  "foo/bar/.." should get simplified to "foo".  "/../.."
  74 should not get simplified (for Mach).  "/.." should not get simplified
  75 (for other networked OSes; POSIX.1 section B.2.3.7).  All these
  76 examples should continue to work with trailing slashes.
  77
  78 Test scoping; here is a start
  79  1      int i=2;
  80  2      int j=3;
  81  3      main()
  82  4      {
  83  5          int i;
  84  6          for (i=600; i>0; i--)
  85  7              print_line(i);
  86  8      }
  87  9
  88 10      print_line(i)
  89 11          int i;
  90 12      {
  91 13          h();
  92 14          printf("%d\n",i);
  93 15      }
  94 16
  95 17      h()
  96 18      {
  97 19          printf("In h...");
  98 20      }
  99 Set a breakpoint in h, and print i, print_line::i, and main::i.  Set a
 100 breakpoint in main (or don't run the program), and test that
 101 print_line::i is an error.  But if i were static, "p main::i" should
 102 work even if the program is not being run.
 103
 104 Write a test for the reentracy bug with rs6000_struct_return_address
 105 in rs6000-tdep.c.
 106
 107 Test "return" from dummy frames.
 108
 109 FORTRAN common blocks (a.out and xcoff--weird.exp has the start of
 110 one but it is not quite right as of 19 Nov 1993).
 111
 112 Test that "x" command sets $_ and $__.  Test $_ in general.
 113
 114 Test ability to process NMAGIC a.out files.
 115
 116 If there are two breakpoints in the same place, and exactly one of
 117 them has its condition true, test that the correct breakpoint gets
 118 printed.
 119
 120 Set a watchpoint on a local variable (to be interesting, make a few
 121 calls, to be more interesting, make a recursive call).  Test that it
 122 gets disabled when leaving that scope.
 123
 124 Test calling a function, hitting a breakpoint in the called function,
 125 calling another function, and hitting a breakpoint.  Test backtrace
 126 works in the presence of multiple dummy frames.  Test that "continue"
 127 will get you out of the inner called function, and "continue" again
 128 will get you back to where you were when you called the first one.
 129
 130 Test special longjmp handling in wait_for_inferior (need to figure out
 131 in detail what the proper behavior in each case is).  Test longjmp to
 132 a place where there is a breakpoint (such that
 133 BPSTAT_WHAT_CLEAR_LONGJMP_RESUME_SINGLE happens).  In general, test
 134 interactions between longjmp and watchpoints, breakpoints, stepping,
 135 call function, etc.
 136
 137 Test jumping right past a breakpoint (the case where wait_for_inferior
 138 passes not_a_breakpoint to bpstat_stop_status).  Might already be
 139 tested by some of the sun3 tests.  Probably want a .s test to avoid
 140 compiler dependencies.
 141
 142 Test more obscure wait_for_inferior cases, expanding on the tests in
 143 watchpoint.exp, signals.exp, etc.
 144
 145 Test that the copyright year in the startup message matches the
 146 current year (would produce a single spurious FAIL on old GDB's, but
 147 probably still a good idea).
 148 \f
 149 (this is for editing this file with GNU emacs)
 150 Local Variables:
 151 mode: text
 152 End: