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

The Open Brand -- Problem Reporting and Interpretations System

Problem Report 1218 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 1218.

Report 1218 Actions

    Problem Report Number 1218
    Submitter's Classification Specification problem
    State Resolved
    Resolution Rejected (REJ)
    Problem Resolution ID REJ.X.0420
    Raised 1970-01-01 08:00
    Updated 2003-03-13 08:00
    Published null
    Product Standard Internationalised System Calls and Libraries Extended (UNIX 95)
    Certification Program The Open Brand certification program
    Test Suite VSU version 4.1.0
    Test Identification CAPIbase/ffcvt 1 2 5
    Problem Summary PG4U.00047 The test is failing because the interpretation of ndigit is wrong.
    Problem Text
    The test is failing because the interpretation of ndigit is wrong.

    We are failing VSU test suite test cases for fcvt() and we think
    either the spec is wrong or very ambiguous in it's wording.
    We have generally found that our implementation matches other
    existing Unix systems:

    fcvt(0.01234987, 11) returns "01234598700" and decpt == zero.

    The test suite would prefer that we return "1234598700" and a decpt of -1.

    ecvt()'s requirements seem to be in conflict with the fcvt() spec:
    fcvt()'s ndigit argument specifies the number of digits after the radix point.
    In this case ndigit is 11 and we return 11 digits.

    ecvt(), however, has a rule that the "...high-order digit is non-zero unless
    the value is 0. fcvt(): "Identical to ecvt() except that ndigit specifies
    the number of digits desired after the radix point."

    I read this as saying that fcvt() would differ from ecvt() where any of
    ecvt()'s behaviors would prevent ndigit from specifying the number of
    digits desired after the radix point. Since the non-zero leading digit
    rule of ecvt() sometimes prevents fcvt() from expressing the correct
    number of digits after the decimal point, I would say the two behaviors
    are in conflict and that therefore the non-zero rule of ecvt() does not
    apply to fcvt().

    This may be a test suite error but I'm assuming the test suite
    thinks it is correctly testing the spec. Please note that non-zero
    rule also affects the 'decpt' returned by fcvt().
    Test Output
    70||"SPEC1170TESTSUITE AREA fcvt 11"
    10|778 /tset/CAPIbase/ffcvt/ffcvt1 17:28:49|TC Start, scenario ref 76-42
    15|760 1.10 1|TCM Start
    400|760 1 1 17:28:52|IC Start
    200|760 1 17:28:52|TP Start
    520|760 1 9764896 1 1|SPEC1170TESTSUITE CASE 1
    520|760 1 9764896 1 2|A call to char *fcvt(double value, int
    520|760 1 9764896 1 3|ndigit, int *decpt, int *sign) shall convert value
    520|760 1 9764896 1 4|a NULL-terminated string with ndigit digits to the
    520|760 1 9764896 1 5|right of the radix point and return a pointer to t
    520|760 1 9764896 1 6|string.
    520|760 1 31457304 1 1|TEST: fcvt(1.000000, 6) should return "1000000"
    520|760 1 31457304 1 2|TEST: fcvt(1290000.000000, 5) should return "1290
    520|760 1 31457304 1 3|TEST: fcvt(19.399984, 8) should return "193999837
    520|760 1 31457304 1 4|TEST: fcvt(15.499900, 3) should return "15500"
    520|760 1 31457304 1 5|TEST: fcvt(-2284.347840, 6) should return "228434
    520|760 1 31457304 1 6|TEST: fcvt(-38.293900, 7) should return "38293900
    520|760 1 31457304 1 7|TEST: fcvt(-1000.000000, 4) should return "100000
    520|760 1 31457304 1 8|TEST: fcvt(123.450000, 10) should return "1234500
    520|760 1 31457304 1 9|TEST: fcvt(-123.450000, 3) should return "123450"
    520|760 1 31457304 1 10|TEST: fcvt(1.450000, 9) should return "145000000
    520|760 1 31457304 1 11|TEST: fcvt(12345.987000, 11) should return "1234
    520|760 1 31457304 1 12|TEST: fcvt(0.123460, 11) should return "12345987
    520|760 1 31457304 1 13|TEST: fcvt(0.012346, 11) should return "12345987
    520|760 1 31457304 1 14|ERROR: fcvt() returned "01234598700"
    220|760 1 1 17:28:52|FAIL
    410|760 1 1 17:28:53|IC End
    80|778 0 17:28:53|TC End
    10|779 /tset/CAPIbase/ffcvt/ffcvt2 17:28:53|TC Start, scenario ref 76-42
    15|761 1.10 1|TCM Start
    400|761 2 1 17:28:56|IC Start
    200|761 1 17:28:56|TP Start
    520|761 1 9830432 1 1|SPEC1170TESTSUITE CASE 2
    520|761 1 9830432 1 2|When value is not 0 the high order digit of the st
    520|761 1 9830432 1 3|returned by a call to char *fcvt(double
    520|761 1 9830432 1 4|value, int ndigit, int *decpt, int *sign) shall be
    520|761 1 9830432 1 5|non-zero.
    520|761 1 31522840 1 1|TEST: fcvt(19.399984, 8) should have non-zero ini
    520|761 1 31522840 1 2|TEST: fcvt(15.499900, 3) should have non-zero ini
    520|761 1 31522840 1 3|TEST: fcvt(-2284.347840, 6) should have non-zero
    520|761 1 31522840 1 4|TEST: fcvt(-38.293900, 7) should have non-zero in
    520|761 1 31522840 1 5|TEST: fcvt(-1000.000000, 4) should have non-zero
    520|761 1 31522840 1 6|TEST: fcvt(123.450000, 10) should have non-zero i
    520|761 1 31522840 1 7|TEST: fcvt(-123.450000, 3) should have non-zero i
    520|761 1 31522840 1 8|TEST: fcvt(1.450000, 9) should have non-zero init
    520|761 1 31522840 1 9|TEST: fcvt(12345.987000, 11) should have non-zero
    520|761 1 31522840 1 10|TEST: fcvt(0.123460, 11) should have non-zero in
    520|761 1 31522840 1 11|TEST: fcvt(0.012346, 11) should have non-zero in
    520|761 1 31522840 1 12|ERROR: The high order digit is incorrect, expect
    220|761 1 1 17:28:56|FAIL
    410|761 2 1 17:28:56|IC End
    80|779 0 17:28:56|TC End

    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.

    We believe the ecvt() spec statement that ecvt() returns a pointer
    to a NULL terminated string of digits whose...

    high-order digit is non-zero, unless the value is 0

    clearly applies to fcvt() as well since the spec states that
    fcvt() is identical to ecvt() except for the semantics of ndigit.

    For ecvt(), ndigits specifies the number of digits contained in
    the string of digits.

    For fcvt(), ndigits specifies the number of digits desired after
    the radix point. So there is no relationship between ndigits and
    the number of digits in the string of digits. Since there is no
    relationship between these two the string of digits returned by
    fcvt() must follow the mandate that the high-order digit of the
    string of digits be non-zero.

    For the example, fcvt(0.012345987, 11), 11 digits past the radix
    point are desired, which would be 0.01234598700.

    The mandate that the high-order digit be non-zero still applies
    so the leading zeros and period are removed, resulting in the
    string, "1234598700", and a decpt value of -1. Taken together
    these 2 values properly meet the requirement of 11 digits past
    the radix character. There is no requirement that all 11 of
    those digits appear in the string.

    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