GAS REGRESSION: .sleb128 tests

[Nick Clifton <nickc at redhat dot com>]

Hi Julian,

  I am seeing a lot of new regressions in the GAS testsuites for various
  targets:

  Checking Binutils in: mmix-mmixware ...
   GAS REGRESSION: .sleb128 tests (2)  
   GAS REGRESSION: .sleb128 tests (4)  
   GAS REGRESSION: .sleb128 tests (5)  
   GAS REGRESSION: .sleb128 tests (7)  

  Checking Binutils in: rx-elf ... 
   GAS REGRESSION: .sleb128 tests (2)  
   GAS REGRESSION: .sleb128 tests (3)  
   GAS REGRESSION: .sleb128 tests (4)  
   GAS REGRESSION: .sleb128 tests (5)  
   GAS REGRESSION: .sleb128 tests (7)
   
  Checking Binutils in: sh-elf ... 
   GAS REGRESSION: .sleb128 tests (2)  
   GAS REGRESSION: .sleb128 tests (5)  
   GAS REGRESSION: .sleb128 tests (7)
  
  Checking Binutils in: tic4x-coff 
   GAS REGRESSION: .sleb128 tests (2)  
   GAS REGRESSION: .sleb128 tests (3)  
   GAS REGRESSION: .sleb128 tests (4)  
   GAS REGRESSION: .sleb128 tests (5)  
   GAS REGRESSION: .sleb128 tests (7)
   
  Checking Binutils in: tic54x-coff ... 
   GAS REGRESSION: .sleb128 tests (2)  
   GAS REGRESSION: .sleb128 tests (3)  
   GAS REGRESSION: .sleb128 tests (4)  
   GAS REGRESSION: .sleb128 tests (5)  
   GAS REGRESSION: .sleb128 tests (7)
   
  Checking Binutils in: xtensa-elf ... 
   GAS REGRESSION: .sleb128 tests (4)  


There seems to be a variety of different, target specific reasons for
the failures.  The MMIX failures for example come from:

  gas/testsuite/gas/all/sleb128-2.s: Assembler messages:
  gas/testsuite/gas/all/sleb128-2.s:4: Error: unknown pseudo-op: `.l1:'
  gas/testsuite/gas/all/sleb128-2.s:8: Error: unknown pseudo-op: `.l2:'

Whilst the RX failures are due to:

  binutils/objdump: section '$DATA$' mentioned in a -j option, but not found in any input file
  binutils/objdump: section 'D_1' mentioned in a -j option, but not found in any input file

The SH failures are due to results like this:

  regexp_diff match failure
  regexp "^ .* 7d2a.*$"
  line   " 0000 fdffffff ffffffff ff012a             ..........*     "
  FAIL: .sleb128 tests (2)

The TIC4X failures are due to:

  gas/testsuite/gas/all/sleb128-2.s:2: Error: Unknown opcode `foo'.
  gas/testsuite/gas/all/sleb128-2.s:5: Error: junk at end of line, first unrecognized character is `0'
  gas/testsuite/gas/all/sleb128-2.s:6: Error: junk at end of line, first unrecognized character is `0'
  gas/testsuite/gas/all/sleb128-2.s:7: Error: junk at end of line, first unrecognized character is `0'
  gas/testsuite/gas/all/sleb128-2.s:12: Error: unknown pseudo-op: `.l1'
  gas/testsuite/gas/all/sleb128-2.s:13: Error: junk at end of line, first unrecognized character is `4'

Similarly the TIC5X failures are due to:

  gas/testsuite/gas/all/sleb128-2.s: Assembler messages:
  gas/testsuite/gas/all/sleb128-2.s:1: Error: Invalid label '.text'
  gas/testsuite/gas/all/sleb128-2.s:2: Error: Invalid label '.globl'
  gas/testsuite/gas/all/sleb128-2.s:2: Error: Unrecognized instruction "foo"
  gas/testsuite/gas/all/sleb128-2.s:5: Error: Invalid label '.byte'
  gas/testsuite/gas/all/sleb128-2.s:5: Error: junk at end of line, first unrecognized character is `0'
  gas/testsuite/gas/all/sleb128-2.s:6: Error: Invalid label '.byte'
  gas/testsuite/gas/all/sleb128-2.s:6: Error: junk at end of line, first unrecognized character is `0'
  gas/testsuite/gas/all/sleb128-2.s:7: Error: Invalid label '.byte'
  gas/testsuite/gas/all/sleb128-2.s:7: Error: junk at end of line, first unrecognized character is `0'
  gas/testsuite/gas/all/sleb128-2.s:10: Error: Invalid label '.data'
  gas/testsuite/gas/all/sleb128-2.s:12: Error: Invalid label '.sleb128'
  gas/testsuite/gas/all/sleb128-2.s:12: Error: unknown pseudo-op: `.l1'
  gas/testsuite/gas/all/sleb128-2.s:13: Error: Invalid label '.byte'
  gas/testsuite/gas/all/sleb128-2.s:13: Error: junk at end of line, first unrecognized character is `4'

Lastly the XTENSA failure is from this result:

  regexp_diff match failure
  regexp "^ .* 83808080 082a.*$"
  line   " 0000 83808080 782a                        ....x*          "
  FAIL: .sleb128 tests (4)


Please could you have a look at these failures and either fix the tests
so that they will work correctly or else arrange to skip the tests for
the particular targets that are affected.  If you do skip a target,
please could you add a comment explaining why the test does not work on
that particular target.

Cheers
  Nick