This is the mail archive of the gdb-testers@sources.redhat.com mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Sunday Project test results, gdb, 2002-02-09


. Summary

  . Coverage

     1  platform
     4  versions of gdb
     5  versions of gcc
     2  debugging formats
    40  total test runs

  . Aggregate results

    . libiberty

      all tests passed

    . gdb 5.1

        0  build failures
        1  test run failures
       45  ERROR
       16  WARNING
      596  FAIL

    . gdb 5.1.1

        0  build failures
        0  test run failures
       21  ERROR
       14  WARNING
      453  FAIL

    . gdb gdb_5_1-2001-07-29-branch

        0  build failures
        0  test run failures
       21  ERROR
       14  WARNING
      452  FAIL

    . gdb HEAD

        0  build failures
        0  test run failures
       21  ERROR
       14  WARNING
      361  FAIL

  . Tables

    http://www.shout.net/~mec/sunday/2002-02-09/index.html

  . Old bugs fixed

    I did not have time to itemize these.

  . New bugs detected

    http://sources.redhat.com/cgi-bin/gnatsweb.pl?database=gdb&cmd=view&pr=348
    gdb testsuite programs: "warning: no return statement in function returning non-void"



. Test protocol changes since last report.

  I upgraded my target platform from red hat linux 7.1 to red hat linux 7.2.
  I did before-and-after test runs (not published) and found no regressions.

  I upgraded my baseline compiler from gcc 3.0.2 to gcc 3.0.3.  This is the
  compiler that I build the software with, including gcc and gdb.  It is not
  the compiler that I use to build the test suite (that is a separate list of
  compilers).  I did before-and-after test runs (not published) and found
  no regressions.

  I added gdb 5.1.1 to the gdb versions under test.

  I added gcc 3.0.3 to the gcc versions under test.  I found one improvement
  versus gcc 3.0.2: gdb.c++/inherit.exp runs much better with gdb-HEAD in both
  -gdwarf-2 and -gstabs+.  I found no regressions.

  I added a new table, "Counts".  These are simply the result counts from
  the current test runs.



. Coverage

  . Baseline software

    make 3.79.1
    binutils 2.11.2
    gcc 3.0.3
    tcl 8.3.4
    expect 5.33.0
    dejagnu 1.4.2

  . Test matrix

    targets:   i686-pc-linux-gnulibc2.2-rh7.2
    hosts:     native
    date:      2002-02-09T10:40:17-08:00 to 2002-02-09T23:47:29-08:00
    gdb:       gdb 5.1, gdb 5.1.1, gdb gdb_5_1-2001-07-29-branch, gdb HEAD
    gcc:       gcc 2.95.3, gcc 3.0.2, gcc 3.0.3, gcc gcc-3_0-branch, gcc HEAD
    gopt:      -gdwarf-2, -gstabs+

    "date" shows the beginning time and end time of my CVS checkouts.
    versions which start with a number are official releases.
    versions named HEAD are the cvs HEAD (mainline).
    versions with any other name are cvs branches.

    gcc release versions are built with binutils 2.11.2.
    gcc CVS versions are built with binutils HEAD.



. Analysis

  . libiberty

    . results

      . target i686-pc-linux-gnulibc2.2-rh7.2

        . host native

          binutils HEAD                         All 647 tests passed
          gcc gcc-3_0-branch                    All 646 tests passed
          gcc HEAD                              All 647 tests passed
          gdb gdb_5_1-2001-07-29-branch         All 645 tests passed
          gdb HEAD                              All 647 tests failed

    . commentary

      It appears that gcc gcc-3_0-branch and gdb gdb_5_1-2001-07-29-branch
      have older versions of libiberty.  That is okay.

  . gdb

    The last report was 2002-01-19.

    . 5.1

      . Summary

	I continued testing gdb 5.1 in order to compare gdb 5.1 -> gdb 5.1.1.

	There are no significant regressions or improvements with gcc 2.95.3
	or gcc 3.0.2, as expected.  (If there were any differences, they would
	come from variations in my test bed).

        There are no significant regressions or improvements with
        gcc-3_0-branch.

        There is continuing deterioration with gcc HEAD.  I believe that gdb
        5.1 + gcc HEAD is no longer a supported configuration, so that this is
        okay.

    . 5.1.1

      . Summary

        This is the first report with gdb 5.1.1 so there are no prior results
        to compare with.

        I've published a separate report on the difference from gdb 5.1 to
        gdb 5.1.1.  There was one regression:

          http://sources.redhat.com/cgi-bin/gnatsweb.pl?database=gdb&cmd=view&pr=334

	gdb 5.1.1 with gcc HEAD suffers from PR gdb/348.


    . gdb_5_1-2001-07-29-branch

      . Regressions

	gdb.base/funcargs.exp and gdb.threads/linux-dp.exp with gcc HEAD:
	regressed.  The test programs do not compile.  PR gdb/348.

	gdb.threads/pthreads.exp with gcc HEAD: regressed.  This test went
	from a healthy state to UNSUPPORTED.

      . Improvements

	gdb.base/interrupt.exp
	gdb.base/return.exp
	gdb.base/return2.exp
	gdb.c++/anon-union.exp
	gdb.c++/method.exp

    . HEAD

      . Regressions

	gdb.base/funcargs.exp and gdb.threads/linux-dp.exp with gcc HEAD:
	regressed.  The test programs do not compile.  PR gdb/348.

	gdb.base/ending-run.exp with gcc 2.95.3: regressed.

	gdb.c++/local.exp with gcc 2.95.3 and -gdwarf-2: regressed.

	gdb.mi/mi-var-cmd.exp: regressed.

	gdb.mi/mi0-var-cmd.exp: regressed.

	gdb.threads/gcore-thread.exp: regressed.

	gdb.threads/pthreads.exp with gcc HEAD: regressed.  This test went
	from a healthy state to UNSUPPORTED.

      . Improvements

	gdb.base/break.exp
	gdb.c++/anon-union.exp
	gdb.c++/classes.exp
	gdb.c++/method.exp
	gdb.c++/userdef.exp
	gdb.c++/virtfunc.exp
	gdb.mi/mi-stack.exp
	gdb.mi/mi-var-child.exp
	gdb.mi/mi0-stack.exp
	gdb.mi/mi0-var-child.exp

      . Other

	gdb.base/callfwmall.exp: test script removed.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]