Author: truckman
Date: Fri Aug 5 05:49:46 2016
New Revision: 1755245
URL: http://svn.apache.org/viewvc?rev=1755245&view=rev
Log:
Fix --enable-debug build with clang as the compiler.
Clang warns about this expression in graphite:
if (!m_srAttachTo == 0)
Operator precedence rules require this to be evaluated as:
if ((!m_srAttachTo) == 0)
but clang worries that the author might have intended this to mean:
if (!(m_srAttachTo == 0))
Fortunately these two are equivalent, as is:
if (m_srAttachTo != 0)
and the latter is much easier for humans to understand.
For some reason, --enable-debug adds -Werror to the graphite CFLAGS even
though its makefile.mk specifies:
EXTERNAL_WARNINGS_NOT_ERRORS := TRUE
Modified:
openoffice/trunk/main/graphite/graphite-2.3.1.patch
Modified: openoffice/trunk/main/graphite/graphite-2.3.1.patch
URL:
http://svn.apache.org/viewvc/openoffice/trunk/main/graphite/graphite-2.3.1.patch?rev=1755245&r1=1755244&r2=1755245&view=diff
==============================================================================
--- openoffice/trunk/main/graphite/graphite-2.3.1.patch (original)
+++ openoffice/trunk/main/graphite/graphite-2.3.1.patch Fri Aug 5 05:49:46 2016
@@ -2452,3 +2452,15 @@
bool operator!=(const GlyphSetIterator & rhs) const throw() {
return !(*this == rhs); }
// Random access iterator requirements
+diff -ur misc/silgraphite-2.3.1/engine/src/segment/GrSlotState.h
misc/build/silgraphite-2.3.1/engine/src/segment/GrSlotState.h
+--- misc/silgraphite-2.3.1/engine/src/segment/GrSlotState.h 2009-01-27
18:01:29.000000000 -0800
++++ misc/build/silgraphite-2.3.1/engine/src/segment/GrSlotState.h
2016-08-02 17:22:26.300396000 -0700
+@@ -481,7 +481,7 @@
+ {
+ Assert(false); // Should have already been set in
HandleModifiedPosition,
+ // but just in case.
+- if (!m_srAttachTo == 0)
++ if (m_srAttachTo != 0)
+ return 0;
+ else
+ m_mAttachAtX = 0; // attach on the right
