RFR(S): 8171248: Minor HarfBuzz fixes to pacify Coverity code scan

[Volker Simonis]

Hi,

can I please have a review for the following small changes which fix
some Coverity code scan issues in the Harfbuzz library:

https://bugs.openjdk.java.net/browse/JDK-8171248
http://cr.openjdk.java.net/~simonis/webrevs/2016/8171248/

These changes only make sense if they are also accepted in the
upstream HarfBuzz repository. I've therefore send out a pull request
with the same changes and kindly requested Behdad to accept them
upstream:

https://github.com/behdad/harfbuzz/pull/377

Following are the details of the fix:

we regularly run Coverity code scans on the OpenJDK sources and
recently discovered two issues with HarfBuzz. While the discovered
issues are not real errors, we think that fixing them my be
nevertheless worthwile in order to increase the readability of the
source code.

We just wanted to ask, if you are willing to accept these changes in
the upstream HarfBuzz repository because only then it would make sense
to also fix them in the OpenJDK copy of HarfBuzz.

The first issue found by Coverity is the last of the following four
lines from src/hb-ot-font.cc:

    if (!subtable) subtable = cmap->find_subtable (0, 2);
    if (!subtable) subtable = cmap->find_subtable (0, 1);
    if (!subtable) subtable = cmap->find_subtable (0, 0);
    if (!subtable)(subtable = cmap->find_subtable (3, 0)) && (symbol = true);

>From the whole context it really took me some time to understand that
'symbol' should only be set to true if 'subtable' is set from
'cmap->find_subtable (3, 0)'. Coverity reports an "assignment instead
of compare" which is a false positive, but we think the could would be
much more readable if changed to look as follows:

    if (!subtable)
    {
      subtable = cmap->find_subtable (3, 0);
      if (subtable) symbol = true;
    }

The second issue is related to the following definition in
src/hb-ot-layout-gpos-table.hh:

  ValueFormat  valueFormat1;           /* Defines the types of data in
                                        * ValueRecord1--for the first glyph
                                        * in the pair--may be zero (0) */
  ValueFormat  valueFormat2;           /* Defines the types of data in
                                        * ValueRecord2--for the second glyph
                                        * in the pair--may be zero (0) */

Throughout hb-ot-layout-gpos-table.hh, '&valueFormat1' is used as if
it were an array of two ValueFormat objects. While extremely unlikely,
a compiler could theoretically insert padding between 'valueFormat1'
and 'valueFormat2' which would make the code incorrect. We would
therefore propose to simply change the previous definiton into a real
array:

  ValueFormat  valueFormat[2];         /* [0] Defines the types of data in
                                        * ValueRecord1--for the first glyph
                                        * in the pair--may be zero (0) */
                                       /* [1] Defines the types of data in
                                        * ValueRecord2--for the second glyph
                                        * in the pair--may be zero (0) */

and change the code which uses 'valueFormat' accordingly.

Thank youand best regards,
Volker