Change-Id: Ic038a1c95721d4e578face5deea0b8d8a4ab7304

Change-Id: Ib763efb268bc9165704185c2b588ed282121cfa0

With just the Oracle JRE installed, it is not possible to load the
"jvm.dll" directly, because the required MSVC runtime DLL is not
found (unless some other software happens to install it system-wide),
as described in http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7184978

There was a hack "do_msvcr71_magic" to manually load the runtime that
is bundled with Oracle JRE 6 and this generalizes that to work with the
MSVCR100.DLL that is bundled with Oracle JRE 7 as well.

An additional adaption of the virtual addresses in the file is done,
and it's a mystery whether it even worked before without that.

This issue was not user-visible before because LO releases 3.5 - 4.2
bundle the MSVCR100.DLL themselves.

Change-Id: If61565df80ff8a68472a76000ab5b10d6c78e11c

Change-Id: Ib4d7c2da71a6fa083fc1ed56f058c8dfbc6e58fb

Change-Id: I629b5f6e57c5d8368bafc544c1c23785ee209a1e

Change-Id: Iab50c52a132c90389992ef68c2d31df95a193ab9

Change-Id: Ia00d7e52de5edfce09c3a0a8aa4390e3e1582a01

"warning: suggest a space before ';' or explicit braces around empty body"

Change-Id: I71a55cc888d3480217621491ecace7d0af3dca6e
Reviewed-on: https://gerrit.libreoffice.org/9441Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
Tested-by: Tomaž Vajngerl <quikee@gmail.com>

Change-Id: I6c2cc90668a2158a0b5bf638ce520c8af7731f4e

The motivation is that this test takes time to execute (and we're
testing a limit here, so it's not easy to speed it up), though moving it
out of CppunitTest_sw_ooxmlexport is still possible (which is the
current critical path during a toplevel 'make').

Change-Id: Ibe6580f8aea1439f4feb281996f2f4bbd46ff4ce

The \vern output was never parsed, and Word uses the generator field, so
let us do the same. And with the standard generator string, if we ever
need to parse it, we already have a parser for it in the ODF filter.

Change-Id: I57f25a6414b2b753a23cf82b691e364d84e4bb44

Copy them to the same folder where the conversion output is written.

Change-Id: I38b93bf3fc427974cff7d6975151451a253b1cce

... to allow continuous typing.

Change-Id: I4717e72e422037a441e582004bc87689ea7bc455

Change-Id: Ibf0cb4c0fa951adcd0a7f185d7dd45b06913a40d

Change-Id: Icc2a4829b8a859756ed194f35df214f8bb55c7ae

Change-Id: I6acc7538fcbf8aa7a5e50ba72b27409252cffa15

Change-Id: Id57eb51b69e50be78f85d19a7b3623c1acdf6509

Change-Id: I102566a95b5b34daf60730ffef290913caf28eb2

Change-Id: I095f45b537690fe98bf385d99194db6319d7ac93

...as flagged by -fsanitize=undefined.  But is it really undefined?
[conv.double] "If the source value is between two adjacent destination values,
the result of the conversion is an implementation-defined choice of either of
those values."  So if the double is between std::numeric_limits<float>::max()
and std::numeric_limits<float>::infinity()...

Change-Id: I6389c8ac4a922991e240638d231dd2a39e173882

Change-Id: Iffe85763097829cb04b766314077b5f2a9b5b8d5

Change-Id: Ic1bae7fcebfaf35b62aecd86ed8b3a70fae3e4d3

As an InfoBar, so hopefully not annoying...

Change-Id: I5d4e79273edc03829fdab2d6d5a291576c954e4b

Change-Id: I0dbb2546f1e776ebf95b61c16ebedbce0ead4c7f

E.g. when pasting from terminal you will get "\r\n".

Writer inserts '\r' character into the document as an
invisible character. If you save that to .odt and reload, you will
get ' ' instead. Or when saving to .doc and reload, you will get '\n'.

Change-Id: Id6780bf6d71a8f0a17c1e6ee7a7dcb4d297446dc

Change-Id: Ife28845af92883d8090bb80283c59a0a9a350b32

collada2gltf does not handle Windows pathes, the used COLLADABU::URI
class can't parse native Windows pathes so call COLLADA2GLTFWriter
with URL only.
Other problems:
collada2gltf uses the URL path for creating/opening files,
instead of native ones.
collada2gltf initializes _outputFilePath member in different way,
sometimes as an URL path, sometimes as a native one, but handle it
on the same way(as an URL).

Change-Id: I104182653c07e7dcf6fc5b6e32c7031d59246b84

Change-Id: I0b60fa5b0af8224b4c517367130aa9738f071531

GeneratedSaxParser::Utils:toURI() method has a bool output
parameter called failed, which is assumed to be set inside
the method before return. At some place the caller code does
not initialize the bool variable passed to this failed parameter
and so when the caller checkes the returned value it is
undefined.
e.g. it can be false when the method called successfully so
the error handling throws away the returned URI.

Change-Id: I3f90fab657a86b42bba0f492518e36c343e69d21

(and breaks the build on windows/cygwin for some reason, although
running the find/xargs command by itself works just fine)
So just use awk instead and have a nicer solution without breaking the
build

Change-Id: I24b9f52c0b84c2e8019f3f6719a307fe508fb755

File is getting crashed while saving in LO, as it tries
to access the value of vector after erasing that value.
Code changed to access the vector value first and
then erase that value from vector.

Change-Id: I29369b7582fdbd7363d0e6fdae091a30c79a747d
Reviewed-on: https://gerrit.libreoffice.org/9433Tested-by: Michael Stahl <mstahl@redhat.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

Change-Id: I26cf27c7edc51d0a28e228c56465c3ec2ad7d35e

Change-Id: I41ec24a02867ba3c5bf4f39b5d79bf6a3254ad0d

...which mentions extern "C", so CppUnit's TestPlugInSignature must reflect
that.  It is a bit odd how CPPUNIT_PLUGIN_EXPORT needs to be passed into
CppUnit, so probably best to keep this as a local patch for now.  (Clang's
-fsanitize=undefined complained about the mismatch.)

Change-Id: Ied179a1afe82ceb04de4739c14cf8fadff31b80f

Change-Id: Ia71c9abedfd325b9f5bb2a5e28b2386db5c2fd6b

...though many more remain, as the design appears to be let heap grow
until exit.

Change-Id: I3e5cc95a0896c8add33a767209ab0ff60d6a4b52

Change-Id: I1a651fb059e6cbaece53c8dbac7436b493f66cd5

Change-Id: I213ad0f0ab853aec5cb523474218dfad2c44d11b

Change-Id: I66ccb5332e6c5eceb2f199911ebd4296c93cac4b

Change-Id: Ic809ddad81608f107749498c1432606403ee2314