...but rather dynamically created (via the SDK's setsdkenv_unix.sh.in) in the
SDK's per-user output tree.  So better reflect that in instdir by putting the
symlinks into workdir/LinkTarget/Library/ instead on Mac.  (Which has the
additional benefit that the location of the symlinks doesn't change with the
versioned SDKDIRNAME.)

Change-Id: I77237f6a5d64771c3cb0866e125925ed7b0ecdaf

Change-Id: I88dabd5612069a7de3cad188344cc5c33ba22a7e
Reviewed-on: https://gerrit.libreoffice.org/13127Reviewed-by: Samuel Mehrbrodt <s.mehrbrodt@gmail.com>
Tested-by: Samuel Mehrbrodt <s.mehrbrodt@gmail.com>

Change-Id: I1d660da4cdd3f612ac1478e37e613446d4253fae

Change-Id: Iccebda56ad7a6e7f7867d3da8033c633211f9e15

884bb65d "INTEGRATION: CWS native141: #i87732#
version number in package names" removed gid_Profile_Setup_Ini and
b857e461 "INTEGRATION: CWS native141: #i87732#
version number in package names" removed its nine correpsonding
gid_Profileitems, but 00de796b "INTEGRATION: CWS
odfversion241_DEV300" brought back seven of the gid_Profileitems (but not the
gid _Profile_Setup_Ini itself; probably due to a merge conflict; then
fd1db0b8 "Undo basis/brand split: move
program/classes/ from basis to brand" removed one of the gid_Profileitems
again).  Now remove the remaining six.

Change-Id: I464ac1258217b2ecf8e94b2c7e59b795306d5b06

This header is also included all over sc.

Change-Id: I4b8092c83a40bdeefd0781046125746ca8d35c9a

types.hxx is included by pretty much everyone inside sc.  Let's not
stick a boost header in there...

Change-Id: Iaf2aa9e13d0e02437bcd9e71a0143432abfa0a7c

Change-Id: I09ec8ed8f05621b72984a34f47db50d20ebd6c06

This header is also a very high impact header.

Change-Id: Iab63f2ec2edebc14b47820c6377a7f83131cfd06

Change-Id: I0aed7ec1bd8ac30ef18147b385c8649a413e4e3a

(regression from 0a1b1511
 and 26259e0c)

Change-Id: Idf9a6f5620c9efe62d2824df1688b30d13a4782e

...cf. previous 2ad716f4 "Revert 'pyuno: set up
fake command line in getComponentContext().'"

Change-Id: Ia77132465ada17a5839f9d4f9ccaf20f8b113075

This reverts commit 2386a92c and its follow-ups
1acaa577 "sal: add special handling of argc==0
to osl_setCommandArgs()" and 01f27b5e "sal: fix
osl_setCommandArgs() on WNT."

The situation that osl_getCommandArgCount is called without a prior call to
osl_setCommandArgs should be considered as harmless and can legitimately not
only happen in the pyuno case discussed in the reverted commits, but also in
case binary UNO is bootstrapped from within a Java process, as happens in
test-javanative in ure/source/uretest/Makefile.

Change-Id: I2829db390e37dfe5daeda33a4c0659b7d35e565a

of group area listeners.

Change-Id: I3a8d2ceb57462c2c2c4cddb6e34966d2465e149b

Change-Id: I75412790e0afc1dbe3255254ff13fd41c5e9934b

"static" keyword explains why first time the option is well taken into account and not the other times.
So let's remove this and the option will be taken into account all the times.

Change-Id: I9db84373e425dfe21c307b965b1796c781cd7863

Change-Id: I210760099eb630c1306630e49bf84a2f3bf7a3b6

Change-Id: I32a9d6cc625537ab6e0940d93bf9a7544df370c8

(regression from a2feb420)

Change-Id: I211910ba12fa1f2ce578bf467169a239ddf83e1f

Change-Id: I4c8aa46a02799dd8e8cedda0dd552b51b248a0e8

The change was broken. Luckily MSVC noticed: sw\inc\doc.hxx(1707) :
warning C4717: 'SwDoc::GetNumberFormatter' : recursive on all control
paths, function will cause runtime stack overflow.

This reverts commit 7ca3303a.

Change-Id: I6ccd990ba71867f0ff96ad06cf5a9e7c98c93657

Change-Id: Ic7ecfe85826d9148d14fce14e9de3f80a11eb775

Change-Id: Id134b05de3a81dd246baf8743c32fa8eb1e37935

Change-Id: I01f66eb78e357780f35cfd1df6722301cc5753b2

Change-Id: Ifb7e86b078bd549506a9cc1b9ce9fc22fffc5eec

Regression of 10fc1383 that exchanged
the PostPaint with SetDirty, but the paint is also needed.

Change-Id: Ic3e81fa106f405568effaed71ab33b575b4c5c5c

Change-Id: I86d1c3e0c177c671b280c05b47312c79389884ed

Add a work-around for Writer's usage of LONG_MAX to SvxColumnDescription:
on 64-bit platforms the LONG_MAX added in SwView::StateTabWin() for
SID_RULER_ROWS will overflow in vcl's LogicToPixel mapping and that
causes wrong positioning of the row highlight lines.

Probably Writer should use something other than LONG_MAX (no reason why
these types need to be bigger than 32-bit anyway) but that needs a
bigger cleanup.

(regression from 4c60f722)

Change-Id: I08147462356368d48959a85a85ef7dd8dcae0943

Change-Id: I009e3a506d9b3418215341cb7f3e232e7fcfa04b

Change-Id: I208bbe47d1134b6a362e50d54093e6d35e85f51b

Change-Id: I0b7ab0ed27ba23f9ad924819193dd30c70e48fed

Also tweak some of the assertions so they don't fire for character
borders.

Change-Id: Id63631864ab3e5f0366a514adc1ff73f01f57a84

Change-Id: I7f3ed82652e98da767cf7021bc5fc42e31e80b73

usage (5 test runs): vcldemo --test 5

Change-Id: Ib8c3db1e74f4a05f16a5359fe8645ffc61e891d5

Change-Id: I36ef1f24567b9e561f2bd5e981cbd60dba1343d1
Reviewed-on: https://gerrit.libreoffice.org/13103Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

the tests are done headless with dbgutil enabled and
this dialog contents dumped in dbgutil mode tries to
spit out the contents of the various "file format
not known", "file busted" dialogs with some bad casts
and falls over and dies

Change-Id: I353619f0209902d4c099b42064fd48c3c547f846

Change-Id: I89b3b175d75f27704dfcd1d11f953f0bfca5b60c

...and not only for instdir/, as was done in
ab67b79c "set OOo reference version to 4.1 for
incompatible MessageBox changes."

Change-Id: I58931982153e619c678adde6cf01ec461dc0d2f5

Because under Linux use text
"'"
and font "Nimbus Sans L"
and export as pdf and the output is garbled

If we do want to stick WinAnsiEncoding here then we'll presumably have have to
test that pEncoding actually matches WinAnsiEncoding and/or generate an
additional Encoding/Differences from WinAnsiEncoding for that font.

This reverts commit eea16cb3.