Home • About Us • A-Z Index • Search * Contact Us • Register • Login • Press • Shop

The Open Brand -- Problem Reporting and Interpretations System

Problem Report 0574 Details

Help Show help | Quick Search | Submit a Test Suite Support Request | Click here to view your privileges

This page provides all information on Problem Report 0574.

Report 0574 Actions

    Problem Report Number 0574
    Submitter's Classification Test Suite problem
    State Resolved
    Resolution Rejected (REJ)
    Problem Resolution ID REJ.X.0178
    Raised 1970-01-01 08:00
    Updated 2003-03-13 08:00
    Published null
    Product Standard Commands and Utilities V2 (UNIX 95)
    Certification Program The Open Brand certification program
    Test Suite VSC version 4.1.6
    Test Identification POSIX.cmd/ed/ed_04.ex 189
    Problem Summary PG4C.00099 This request correctly claims that the test may hang intermittently on implementations with an ed command that does not exit when it reads an EOF.
    Problem Text

    ed assertion 189 intermittently hangs on our system. Some of the testcase code
    looks like this:


    send "1,\$t\$ 0a\n" line 11841
    expect {
    \\? { # ok }
    timeout { exit 1 }
    }
    exec sleep 5
    close
    wait line 11848


    It appears that the testcase is missing the command to close the
    file being editted. Adding "send "Q\n"(Capital Q <newline>) before the
    "exec sleep 5" statement clears the hangs. This appears to be consistent
    with other tests that use expect.

    Editted Consultant response


    The submitter is correct in stating that the expect script does
    not exit ed (which is what the Q command does) before attempting
    to terminate the expect script. But we do not believe this is a
    valid reason for the test to fail.

    The expect script starts its termination process by issuing the
    expect command, "close". This command closes standard input to
    the ed process. Once this happens any future attempt by ed to
    read from stdin will return EOF. Although it is not mandated
    this will cause the ed process to exit on most implementations.

    If this not a specification requirement, then you cannot expect it to
    work on all implementations. Therefore this is a not a portable testcase.

    The close command also sends the ed process a SIGHUP signal.
    According to the XCU ed spec

    The ed utility will take the standard action for all signals
    (see ASYNCHRONOUS EVENTS in the XBD specification, Section 2.1,
    Utility Description Defaults) with the following exceptions:

    SIGHUP
    If the buffer is not empty and has changed since the last write,
    the ed utility will attempt to write a copy of the buffer in a
    file. First, the file named ed.hup in the current directory will
    be used; if that fails, the file named ed.hup in the directory
    named by the HOME environment variable will be used. In any
    case, the ed utility will exit without returning to command mode.

    So even if the submitters implementation of ed ignores EOF it
    should exit when it receives the SIGHUP signal.

    Our implementation does exit when it receives a SIGHUP signal. The process
    list when tp189 hangs shows the following:

    vsc0 10770 10000 11 09:14:11 pts/0 0:00 ed ed_out_189_1

    Issueing a kill -1 10770 will kill the ed process and cause the testcase
    to end normally, issueing a PASS test result code. So our ed does respond
    to the SIGHUP signal. It appears the the close does not issue a SIGHUP
    to the ed command as the consultants claim. The Man page for close
    indicates it does send the EOF discussed above, but makes no mention
    of sending a SIGHUP. In addition grepping for kill(SIGHUP) doesn't
    get any hits in the entire $TEST_ROOT/vsc directory.
    Test Output
    No output is produced since the test hangs.

    Review Information

    Review Type TSMA Review
    Start Date null
    Completed null
    Status Complete
    Review Recommendation No Resolution Given
    Review Response
    We recommend this request be refused.

    The submitter is correct. Although the documentation states that
    the close command sends a SIGHUP, the Expect version used in the
    VSC release does not do this.

    However, patch 4.1.6A provides a workaround for this failure.
    We reccommend that the implementation be required to apply this
    patch prior to making a single branding run.

    Review Type SA Review
    Start Date null
    Completed null
    Status Complete
    Review Resolution Rejected (REJ)
    Review Conclusion
    This request is refused.

    Problem Reporting System Options:

     

    Back   

Contact the Certification Authority