So no need to hide the Xt type.

Change-Id: Ib7d96745294ab747b3c8ed355d4562a68e3cb0cf

Change-Id: Ib3f43db131cf5562ad011538873c2ee51839665c

Change-Id: If56233b5226cec9516d5e2f8992e1b0beae733bf

Change-Id: Ic496662d623c4732db41476109b3731cf1fd71a0

Change-Id: I2e12d923a07385fc6a036863070a10666542d39d

Change-Id: I853b6b1cfcd4847603d9920a47298d1b9105b46f

See 656ab07b (Sync oox -> writerfilter
alias for wordprocessingml, 2014-08-10).

Change-Id: I5179961b683473a502693923422171bb274fc845

Change-Id: Ifbcb7716640cd05d4b98eef22f9ddf49f3bd79aa

Change-Id: I512713e9b9aa1ceb3d98af7a1a6abd144e370689

Change-Id: Id82d6d54b160b5b78a53c526e6ccab0514c21312

We call the Win32 GetObject() in just one source file and there it is
the A version we mean. (As such, as we call it to get BITMAP objects,
which do not have different A and W variants, we could as well call
GetObjectW().)

Note that we have several own member functions called GetObject(), so
until those are renamed we still need to #undef the GetObject
definition (from <windows.h>).

(No, I am not working on the EasyHack fdo#82580 as a whole. It is
intended for some less mature LO contributor. I just did a few small
things because I couldn't resist.)

Change-Id: Idab79503b3d899ef19f0608677752a62b20302ff

Change-Id: I1edfa97d9874c70fcc61d1b521c905fb499a9bd1

Change-Id: I54830ba65961f8f7ab6fa3555a1e6021b14cdb1c

Change-Id: Icc2c34f51601304e2846d8bd9fce2f384327a4e2

Most of the scripts parsing model.xml just load it as a DOM and extract
info from it needed for their task, which means if you have a typo in an
element / attribute name, it's just silently ignored.

This is annoying, let's prevent that by validating the model before
doing anything with it.

Change-Id: I4229e57c12840fead62cbd150dd1408f2f0d29eb

Change-Id: If9b32b09fd8cbf8b194b75357494d99aaa3fe06e

Change-Id: I1521f174ac53c2e8e96dcb7b5c1264f0bf460fd0

Change-Id: I69f47cf975a9ed91d4a1408351b5ad1751aab19e

Change-Id: I0d2678dd86e42f882e8ed9926917f1e82746997e

Change-Id: I3b0f66ad96813b064d61ea8b2ed6f2c383a11390

Change-Id: I1d2f24def771651d36f06432c20e80429379c9fe

res ctor fallout continues at pace

Change-Id: Ic431c678919a6521ce94719cbf06b87ebf922a0c

Change-Id: Iaabf05075bb51ed17abff5f4f4bb18c101e552a2

Change-Id: I292921ef5b98ffbae9a0d5b631498a2d6efefbfd
Reviewed-on: https://gerrit.libreoffice.org/10941Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>

ODF specifies that extension elements in a document are removed before
the validation. With our old implementation of the number format export
this could result in two number:text elements following each other.

According to the spec a number:text element may not follow another
number:text element.

The number:text element before loext:fill is actually not that
interesting so it makes more sense to make this one an extension
element.

Possible documents showing issues:

fdo45268-1.xlsx
fdo49150-1.xlsx
and many more

Change-Id: I3ad92b8a61a246b3dbc0132d9a2c367488a766d0

Change-Id: I1a63802dfeb702f6f6e9506339fd299b98fd8604

Change-Id: I9de4fdd28c6d75133289b1c8a655381ecc5675c7

This drops the calls to ScMarkData::GetMarkedRanges by a factor of 1k.

Change-Id: I12b1ab3877120a1fbccaccaa7996cbb7efe948b6

Change-Id: I715f8b2c09bd6af2cf1f45029f0561d79817deed

Change-Id: I22f239fbbbfa3ac562b860efb990dff10d0e0e24

Change-Id: If2f35d4ce68f604db1789e1535c341908628838d

Change-Id: If3cfc7d1974036fe6110726da574b07975ab9aef

Change-Id: I770b941c065ca41541f27d01f78727d9d0e32a7e

This reverts commit 90aecb31.

Change-Id: Ic814a2301aaf6e57aa8ca04b837acafb6e4f4566

Change-Id: I3dd0005f00c4a861103ffd1a0422b9fc9397f6fb

Change-Id: I5b9cb4a76739a0c91597c0470c737439d2b07e4f

As it's like RID_SVX_MSGBOX_LANGUAGE_RESTART

Change-Id: I49c68537c22b348dd165fd1cfb78b67d8f3f7c88

Change-Id: Ideeb3ae8aa6493fc2d2a761d46f95baf0158c6e8

that was a bit too much removal

partial revert of fd1e9bbc1ebd5040b092aad1d4a68041286fabda

Change-Id: Iabced82fa6aedd25aacf242cafa384e83fd539cb