[jira] [Created] (XERCESC-2263) NaN bypasses minInclusive/maxInclusive bounds check via unhandled compareValues INDETERMINATE(2) result

Guichard Desrosiers created XERCESC-2263:
--------------------------------------------

             Summary: NaN bypasses minInclusive/maxInclusive bounds check v=
ia unhandled compareValues INDETERMINATE(2) result
                 Key: XERCESC-2263
                 URL: https://issues.apache.org/jira/browse/XERCESC-2263
             Project: Xerces-C++
          Issue Type: Bug
          Components: Validating Parser (XML Schema)
            Reporter: Guichard Desrosiers


SUMMARY
-------
NaN silently passes minInclusive and maxInclusive facet validation on xs:do=
uble and xs:float types.=C2=A0
This is a regression caused by a mismatch between compareValues return doma=
in and boundsCheck=C2=A0
handling logic.

ROOT CAUSE
----------
compareValues was updated at some point to return 2 (INDETERMINATE) for NaN=
 comparisons instead=C2=A0
of 1. The boundsCheck logic in minInclusive and maxInclusive was never upda=
ted to account for=C2=A0
this 4th return value, breaking the assumed trichotomy of \{-1, 0, 1}.

compareValues return values:
=C2=A0 -1 =C2=A0=E2=86=92 less than
=C2=A0 =C2=A00 =C2=A0=E2=86=92 equal
=C2=A0 =C2=A01 =C2=A0=E2=86=92 greater than
=C2=A0 =C2=A02 =C2=A0=E2=86=92 INDETERMINATE (NaN)

Buggy checks:
=C2=A0 minInclusive: if (result =3D=3D -1) =E2=86=92 FAIL =C2=A0 misses IND=
ETERMINATE(2) =E2=86=92 NaN passes =E2=9C=97
=C2=A0 maxInclusive: if (result =3D=3D 1) =C2=A0=E2=86=92 FAIL =C2=A0 misse=
s INDETERMINATE(2) =E2=86=92 NaN passes =E2=9C=97

UNAFFECTED FACETS
-----------------
minExclusive and maxExclusive are not affected as their inverted logic inci=
dentally catches=C2=A0
INDETERMINATE(2):
=C2=A0 minExclusive: if (result !=3D 1) =C2=A0=E2=86=92 FAIL =C2=A0 catches=
 -1, 0, and 2 =E2=9C=93
=C2=A0 maxExclusive: if (result !=3D -1) =E2=86=92 FAIL =C2=A0 catches =C2=
=A00, 1, and 2 =E2=9C=93

AFFECTED FACETS
---------------
=C2=A0 xs:minInclusive on xs:double and xs:float
=C2=A0 xs:maxInclusive on xs:double and xs:float

FIX
---
Tighten acceptance condition to only pass on known-valid orderings:
=C2=A0 minInclusive: reject if (result !=3D 1 && result !=3D 0)
=C2=A0 maxInclusive: reject if (result !=3D -1 && result !=3D 0)

=C2=A0

REFERENCES
----------
[fix: NaN bypasses minInclusive/maxInclusive facets by gdesrosiers1805 =C2=
=B7 Pull Request #64 =C2=B7 apache/xerces-c|https://github.com/apache/xerce=
s-c/pull/64]



--
This message was sent by Atlassian Jira
(v8.20.10#820010)