Using 0 as index offset to set user data in window handle,
is failing on x64 platform with:

ERROR_INVALID_INDEX 1413 (0x585)

To find it out GetLastError() can be used:

SetLastError(0);
LONG_PTR salFrameBack = SetWindowLongPtrW(hWnd, 0, (LONG_PTR)pFrame);
DWORD status = GetLastError();

Switching to documented GWLP_USERDATA instead, solved the problem.

Change-Id: I496ed2915e44309eaaaa4b576dca30b0051bdbfd
Reviewed-on: https://gerrit.libreoffice.org/13807Reviewed-by: David Ostrovsky <david@ostrovsky.org>
Tested-by: David Ostrovsky <david@ostrovsky.org>

Change-Id: I7e483efd3aa96c34fb02affa73ee5ffb9840209c

Change-Id: I68bacfc41dc7b104a07c44135c0124e3f6464d67

Change-Id: Iea8552570146bd516ca2ee241caa4fcc3047b9f7

Change-Id: I584045f8b626ff3a51e868c586581219bf4a758c

Change-Id: Icb6d782a4a3aa91e05c6466edb3d84c5dfc08f08

Change-Id: I43463b56cabfea4c9ee2b98445f7fb522221197d

Since we have a SwContentTree::Notify() now, might as well clear our
pointers to recently deceased SwWrtShells when the obituary arrives.

(regression from b5c1f352)

Change-Id: I44282658468f1b758038fbf1c7fc13dfe1b4b8ed

Probably 03cca068 was not sufficient,
because the ~ScCellRangesBase may access a deleted ScDocument after the
early return.

READ of size 8 at 0x61b000191cb0 thread T7 (cppu_threadpool)
    #0 0x2b3c8fd7c8b3 in ScDocument::RemoveUnoObject(SfxListener&) sc/source/core/data/documen3.cxx:863:9
    #1 0x2b3c94517824 in ScCellRangesBase::~ScCellRangesBase() sc/source/ui/unoobj/cellsuno.cxx:1467:9
    #2 0x2b3c9459cf98 in ScCellRangeObj::~ScCellRangeObj() sc/source/ui/unoobj/cellsuno.cxx:4786:1
    #3 0x2b3c945eca28 in ScCellObj::~ScCellObj() sc/source/ui/unoobj/cellsuno.cxx:6088:1
    #4 0x2b3c945ecef1 in ScCellObj::~ScCellObj() sc/source/ui/unoobj/cellsuno.cxx:6087:1
    #5 0x2b3c945ed093 in non-virtual thunk to ScCellObj::~ScCellObj() sc/source/ui/unoobj/cellsuno.cxx:6088:1
    #6 0x2b3c56cb35ec in cppu::OWeakObject::release() cppuhelper/source/weak.cxx:205:9
    #7 0x2b3c94523d63 in ScCellRangesBase::release() sc/source/ui/unoobj/cellsuno.cxx:1772:5
    #8 0x2b3c945a77ec in ScCellRangeObj::release() sc/source/ui/unoobj/cellsuno.cxx:4835:5
    #9 0x2b3c945f842c in ScCellObj::release() sc/source/ui/unoobj/cellsuno.cxx:6128:5
    #10 0x2b3c945fb953 in non-virtual thunk to ScCellObj::release() sc/source/ui/unoobj/cellsuno.cxx:6129:1
    #11 0x2b3c7f305faa in bridges::cpp_uno::shared::freeUnoInterfaceProxy(_uno_ExtEnvironment*, void*) bridges/source/cpp_uno/shared/unointerfaceproxy.cxx:42:5

freed by thread T14 (cppu_threadpool) here:
    #0 0x43436b in operator delete(void*) (/data/lo/build_clang/instdir/program/soffice.bin+0x43436b)
    #1 0x2b3c9382462e in ScDocShell::~ScDocShell() sc/source/ui/docshell/docsh.cxx:2722:1
    #2 0x2b3c93824d46 in virtual thunk to ScDocShell::~ScDocShell() sc/source/ui/docshell/docsh.cxx:2755:1
    #3 0x2b3c58e2058a in SvRefBase::ReleaseRef() include/tools/ref.hxx:194:29
    #4 0x2b3c58e12726 in tools::SvRef<SfxObjectShell>::~SvRef() include/tools/ref.hxx:52:24
    #5 0x2b3c5a8a8f99 in IMPL_SfxBaseModel_DataContainer::~IMPL_SfxBaseModel_DataContainer() sfx2/source/doc/sfxbasemodel.cxx:247:5
    #6 0x2b3c5a8a9141 in IMPL_SfxBaseModel_DataContainer::~IMPL_SfxBaseModel_DataContainer() sfx2/source/doc/sfxbasemodel.cxx:246:5
    #7 0x2b3c5a7b1905 in SfxBaseModel::dispose() sfx2/source/doc/sfxbasemodel.cxx:795:5
    #8 0x2b3c5a7c8146 in SfxBaseModel::close(unsigned char) sfx2/source/doc/sfxbasemodel.cxx:1418:5

Change-Id: Ia041dc6596d1b0b6b979a20fa93e1490c657e15b

Change-Id: Ia53dc393e4284989f6e967c4fa42e7ba6fc58183

Change-Id: Ia8471c9745b86c4137297fd9cd0a3fea712a14fa

Change-Id: Id8524328f1bdacbc37dd6846f053e5bcfbcd782c

Would be nicer to fix it though.

Change-Id: Ia5f49a69374e9e79a9a16542bc44f279b9e5a6f7

At least on some 64-bit Cygwin installations, the input gbuild-to-ide gets
include \r\n EOL. Python3 regex excludes \n from .* but leaves \r in the matching
group. This training cr ends up in the file paths in VS project files, breaking
it as they are invalid characters.

Change-Id: I481c1adbf6f6db7ab0ee0d3a768b7a397dc78ee5
Reviewed-on: https://gerrit.libreoffice.org/13803Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>

Change-Id: I9fd7b873b6e64a6450c1bd946cd9bb08c293290a

Change-Id: I844532d08123ed0c91ccec73cc806277fd821089

Change-Id: Iaa46849742c215798722d03d9ee59bb39d8033f7

Also OOXML calls these API functions, this is the central place to handle it.

If there are no matching tab names for a FileId, preserve at least the known
reference parts. In case of 2D references the sheet name is in the token if
for example read from .xlsx, only for 3D references the second sheet name
would be needed. The underlying makeExternalRefStr() and its subroutines
handle the missing tabname elements gracefully.

Still this situation is worth an assertion.

Change-Id: I3df065af8e4ef44734f468fd455c3b7c93d7fbc6

This reverts commit 808b0bf8.

This is not the right fix, it cures a symptom but probably will break
other cases. See also my comment on https://gerrit.libreoffice.org/13670

Change-Id: I5bb661ba2bcbd387a69aacf08cb13b503315c5b1

Change-Id: Ica5ebb04ec50c6579f71d1152ae7619dd37b879e

Change-Id: I921c631545be6da66639709642e423c44e82c5af

Change-Id: I46225e30f6326e0af5a8ac3bebe9847f4dbe50d0

...at least in CppunitTest_sd_export_tests, the dynamic_cast in
PPTWriterBase::GetStyleSheets (sd/source/filter/eppt/pptx-epptbase.cxx) can
fail, but the resulting reference appears to not be accessed then.  So change
from reference to (potentially null) pointer.

Change-Id: Ide49be7bffefa7906e84349642f7c3d0666d414e

Change-Id: Iad41da9294c9a38e501e5bcd1b3185f2e3b26ff2

The ENABLE_CRASHREP is no longer AC_SUBSTed by configure, so it is
always non-empty in config_host.mk, resulting in -Zi without -FS,
which breaks MSVC 2013 builds.

Leave a placeholder in configure for the new crash reporter vaporware,
and force the CFLAGS to be the same as in the --eanble-symbols case.

(regression from 8ab4d2f2)

Change-Id: I476fb2e4f50de5304632628b25e3997ea6c31c27

and remove the need to worry about keeping indexes correct

Change-Id: I9a5fc00f7e28f305279b41099274c96daebebb95

Change-Id: I22a3aba3266ae8209242cf6e02ca8c66b1be863b

Rename OcKumZinsZ to ocCumIpmt, SC_OPCODE_KUM_ZINS_Z to SC_OPCODE_IPMT,
ScInterpreter::ScKumZinsZ() to ScInterpreter::ScCumIpmt().

Change-Id: I7a062374fa0dad375c9277ca7e1c23f6d6855013
Reviewed-on: https://gerrit.libreoffice.org/13816Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>

Rename OcKapz to OcPpmt, SC_OPCODE_KAPZ to SC_OPCODE_PPMT
ScInterpreter::ScKapz() to ScInterpreter::ScPpmt().

Change-Id: I5bc556e086970ea487e14b4663a9b3d2aaf9bb66
Reviewed-on: https://gerrit.libreoffice.org/13815Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>

Rename OcKombin to OcCombin, SC_OPCODE_KOMBIN to SC_OPCODE_COMBIN,
ScInterpreter::ScKombin() to ScInterpreter::ScCombin(), OpKombin to
OpCombin.

Rename OcKombin2 to OcCombinA, SC_OPCODE_KOMBIN_2 to
SC_OPCODE_COMBIN_A, ScInterpreter::ScKombin2() to
ScInterpreter::ScCombinA(), OpKombin2 to OpCombinA.

Rename OcVariationen to OcPermut, SC_OPCODE_VARIATIONEN to
SC_OPCODE_PERMUT, ScInterpreter::ScVariationen() to
ScInterpreter::ScPermut(), OpVariationen to OpPermut.

Rename OcVariationen2 to OcPermutationA, SC_OPCODE_VARIATIONEN_2 to
SC_OPCODE_PERMUTATION_A, ScInterpreter::ScVariationen2() to
ScInterpreter::ScPermutationA(), OpVariationen2 to OpPermutationA.

Change-Id: I2efa01e9b160e98058879d08e2e2313284a5ec91
Reviewed-on: https://gerrit.libreoffice.org/13814Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>

Rename OcZinsZ to OcIpmt, SC_OPCODE_ZINS_Z to SC_OPCODE_IPMT,
ScInterpreter::ScZinsZ() to ScInterpreter::ScIpmt()

Change-Id: I6ed671aff85b7d2ec06dd66cd7ec9390776008b1
Reviewed-on: https://gerrit.libreoffice.org/13813Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>

Change-Id: I5e29626fe2803d2751bdec9c6919662ea37cf64c

Change-Id: I99d3a715cce203eb2303c76da3b20e6d853a9d23

and adjust the basic test-case, which is surely wrong

The syntax for REPLACE is:

// Replace(expression, find, replace[, start[, count[, compare]]])

surely in the case of a start of 3 the preceeding chars should
be returned unchanged in the result, not stripped off as before
this change.

This reverts commit 869402a5.

Change-Id: Ie710e4de9e7e35c84abe2770142a963532820af4

(mpLayoutCache may be null, but that was harmless as Write was effectively static)

Change-Id: I41b10593c5ae98de92e7b305d9db75d0c4c7b334

Change-Id: Ibffc44da6e9d6fb02d43d8a2b7e143ff994d76a8

so make them final and make their destructors non-virtual

Change-Id: I94453786d2befb885996996040a766bc4ccbb4e3

Known issues: the mimetype returned for .odp files is null and share
does not work on these files.

Change-Id: I61e752763a6934657c30fff3cefc57ac0144ec71