Change-Id: Id4e65908cc73978205bc0858a3a809b882864b62

Actually, we check HAVE_FEATURE_READONLY_INSTALLSET and not
HAVE_FEATURE_MACOSX_SANDBOX, but in practice we set the former only
for the latter case.

We could do this test cross-platform but I didn't bother now. Nobody
uses --enable-readonly-installset when building for Linux or Windows
anyway, I think.

(There must be more places where we could and should check
HAVE_FEATURE_READONLY_INSTALLSET.)

Change-Id: Ie54f436616a1db3119b0bc0148149784127b156d

- problem was hang after replaceall or searchall function
- fix slow performance by looping only selected cols.

Change-Id: Ic0178af33bf381e52584bd4366bff9e128891b64
Reviewed-on: https://gerrit.libreoffice.org/11829Tested-by: LibreOffice gerrit bot <gerrit@libreoffice.org>
Reviewed-by: Matthew Francis <mjay.francis@gmail.com>
Tested-by: Matthew Francis <mjay.francis@gmail.com>
Reviewed-by: Kohei Yoshida <libreoffice@kohei.us>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: I4fd649ee20f76063784cea5f46e42cadb350077e

Change-Id: I5b432273ce24d8575ac36f5336717669db733c35
Reviewed-on: https://gerrit.libreoffice.org/11730Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

from SvxMSDffShapeOrders.

Change-Id: Idf551f1fbde907759d6a296141e2837264dbbb10
Reviewed-on: https://gerrit.libreoffice.org/11840Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Use gperf instead of ls to find binpath, so that vs2012/3 can be tested
on non-Windows systems.
Output the vcxproj and vcxproj.filters files with indentation and
formatting.
Make some minor changes to improve pep8 compliance.

Change-Id: Ie8ac11ecf75ad170f0756f803c45356685792e5c
Reviewed-on: https://gerrit.libreoffice.org/11797Tested-by: David Ostrovsky <david@ostrovsky.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

from ItemFormatMap.

Change-Id: I956b5797e677d22eb71fe801b650db7c982d6d51
Reviewed-on: https://gerrit.libreoffice.org/11854Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: I5897a774b8949f336a40096a0727a347bae012ac

Change-Id: Ia2bfaff2a284c2ea40198f1152716860d5963482

tools to correct svg:d imports

(cherry picked from commit f077f99d)

Conflicts:
	basegfx/source/inc/stringconversiontools.hxx
	basegfx/source/tools/stringconversiontools.cxx

Change-Id: I7de1558682990f83e66fdded3d9d30e21e4f97fe

Calling std::list::remove() iterates over all elements...

Change-Id: I46d00217911c1863f70d3ba72dc7099411714471

Originally this used both SolarMutex and an own mutex, then a deadlock
was resolved in 2fe6a4a3 by not locking
SolarMutex.

Since the class will call event listeners without dropping the mutex
e.g. in insertNoCheck(), using the SolarMutex appears better anyway.

With this, installing a BASIC extension no longer triggers SolarMutex
asserts in SfxBroadcaster.

Change-Id: Ib9a2ee491ef53b1a53855af0fc22e863c5e7cb91

The locking strategy in the basic module is totally unclear to me, there
does not appear to be a dedicated mutex for the core stuff callable via
StarBASIC, just a bunch of SolarMutexGuards at random locations;
let's try to use SolarMutex at the UNO entry points...

Change-Id: Ia9c45fdcfb5ffd0a4acc77ef5d2fabfb8743ad38

Looks like they are backwards to me

Change-Id: I20f38050172c1671a81d6ab7de80123ad1703528

Change-Id: Ie00a983cc93736528e16b4244aaf1361536ed9b6

Does nothing useful so far.

Change-Id: I9d288a8b8ab92e01a28c92bb41162af4410e0163

Incorrectly converted in a384b21c
where SFX_IMPL_ONEINSTANCEFACTORY was used.

AFAICS it's the first converted single-instance service which implements
css::lang::XInitialization. That's kind of strange but can do its job.

sbergman@redhat.com:  Three things were necessary in order to not call the
~ShutdownIcon code too late during exit now:

* Move the relevant code from ~ShutdownIcon to ShutdownIcon::disposing.

* Add a dummy <singleton name="com.sun.star.office.theQuickstart"/> so the
  service manager will eventually dispose the (single) instance.

* In
  cppuhelper::ServiceManager::Data::Implementation::createInstanceWithArguments
  do not shortcut updateDisposeSingleton in that odd case of calling
  createInstanceWithArguments on an implementation that (effectively) is a
  singleton (as otherwise the service manager would still not dispose it).  It
  looks to me like that "return inst;" was an inadvertent leftover in
  874c4818 "Simplify service manager's tracking
  of singletons" and wasn't intended to serve some subtle purpose.

Change-Id: Icd4d3168ec0bbb820b17ac321fe897ac9f9ce7fc

causing m_aDataArray to be cleared, so the rItem the result of GetString
is being assigned to is invalid

Change-Id: I1dc30d675233ddc53234bdb5349c5f374b036ca8

Change-Id: I52eb3400769999d7f554c3bdb8746f65b7990388

Apparently fc04f763 used some sed script
to change all ...

Change-Id: Ie609bd02a2c5d70109fc6185cf4440480f29d8f5

Apparently fc04f763 used some sed script
to change all ...

Change-Id: I0a8171aa474070bb1a8584e5c8c714fe34694618

aa33dd16 changed the currency symbol
from ¢ to ₡ but changed only the symbol. Actually changing the number
format codes wasn't necessary at all, but changing only the symbol is
wrong, see also the ATTLIST Currency  legacyOnly description in
i18npool/source/localedata/data/locale.dtd

Hopefully nothing relies on the actual symbol used in the number format
codes ...

Change-Id: I92e79403fbb80f18c44dc3ecf5a3830a52fd2ff8

...so must not be checked for if --without-system-neon

Change-Id: If598108f98d6f94f32650feba7044224a78a4c12

to get the NONE constant out of the global namespace

Change-Id: I68b7ab93fe0d4f8d02beb96798cd79e8aa3a0c71

<sberg> kendy, as you recently changed Timer handling on Windows: debugging why
 quickstarter no longer works on Windows, I run into the phenomenon that
 IdleTerminate::Timeout (sfx2/source/appl/shutdownicon.cxx) executes on the main
 thread, and from within the m_xDesktop->terminate() call Timer::ImplDeInitTimer
 is called which deletes the pTimerData corresponding to our IldeTerminate, so
 that Timer::ImplTimerCallbackProc, after the return from
 IdleTerminate::Timeout, will operate on a stale pTimerData and crash; could
 that be related to those recent changes?
<kendy> sberg: I think mst told that we were previously never deleting the
 timer, and that he did some changes there [...]
<mst__> sberg, IdleTerminate needs a different implementation
<mst__> sberg, does it work to do this via PostUserEvent, as
 "Application::Quit()" does?
<sberg> mst__, do you think calling terminate from a Timer did work in the past?
 [...]
<mst__> sberg, perhaps it did before [a recent] commit but i would guess it
 would be more by accident than by design

Change-Id: I23b14ba59a963cc2209a261a1459d532a88acdc0

after my hasty commit
bc002034  "convert PathKind enum in SD to enum class"

Change-Id: I97263e907bf7363121dc1b437eaf564efdb66ad2

Change-Id: Ia418876b8535cb69b544217fa7976514e77a88b7

to get the NONE constant out of the global namespace

Change-Id: Iecbe39126c1ec08aa852fcb903cee1a6e6f4594b

mostly to get the NONE constant out of the global namespace, so I can
convert other enums which also want to use the NONE constant

Change-Id: I988d5d7e5316dd5e4ca0be0cc8f043f69bbb900c
Reviewed-on: https://gerrit.libreoffice.org/11848Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>

Change-Id: I9676e38029c511678efed1679f9488f1f15f4f8d

Change-Id: If423bd3fde53695be33a4f890f7d363ccce8dbf0

Change-Id: I99f3010e30f81786b938dc11736ea1597cd5530d

Handles the .uno: and .slot: commands.  Cannot load / save the info yet, only
dumps it to the screen at the application exit.

Disabled in the code now, needs a configuration option.

Change-Id: I4e689f534fe5c8a5af84df472e47f276fb7af89b

Change-Id: Ibf4f2d46bfbb977f0ba27110d86d7b7f1173631c

smb was included in the set of supported protocols since those X-KDE-Protocol
lines got introduced with 673be8e7 "add
X-KDE-Protocols," but as discussed in the mail thread starting at
<http://lists.freedesktop.org/archives/libreoffice/2014-September/063621.html>
"X-KDE-Protocols=...,smb,...," actual support for smb URLs in LO under KDE is
very fragile and typically requires further steps from the user (which vary
across the various LO configure options, --enable-gio vs. --enable-gnome-vfs,
and Linux distros) to work at all.

So it is probably best to not unconditionally claim smb support in those
X-KDE_Protocols lines at least in the vanilla LO sources (esp. since the TDF
builds with --enable-gnome-vfs need a really weird OOO_FORCE_DESKTOP=gnome
workaround to make it work at all, cf.
<http://lists.freedesktop.org/archives/libreoffice/2014-October/063876.html>
"Re: X-KDE-Protocols=...,smb,...").  Any Linux distros that are confident that
smb support in LO under KDE reliably works for them can patch those lines
locally.

Change-Id: I1c247d85825c96f0d74efcce3003e327a1522fa3

Change-Id: I307d7833fb5556c5509edd698b4b5ecd7b7a5fb3

Change-Id: Id1c6357c84e03c71136eb92434002e5e0a4b71d7

The core package will obsolete the individual ones (of the same
major.minor version)

Change-Id: I30dea2c08ec7eca052db2706b7f9f30f91e37095

Change-Id: If26d21593f36a39de3bc5b81717e5000795c97ae