Change-Id: Id96026910ffbdda72bfb7a4bf8653c63759e7c96

Change-Id: I90069f562c61359f9b226b147fe99807ecb3ad8a

Change-Id: Ie122a6befd94c836b6f08a7fe3ec8e4f2d67c690

Change-Id: I604244b032855e0e6b94e3ed4217ee87d92c8419

Change-Id: I3229ae20624c753d5e802d56251a628399a85817

Change-Id: I948d7c241f8024edb3bdced21d9d819571f02387

Change-Id: I6ba77dad167e2989e7151a8fa3c085c9133ed623

Change-Id: I5fae2ec81b50203e9d6100f8c938fb3643ff651f

Change-Id: I1625af99f9b13eaae7b74c65b6306bae6d063627

Change-Id: I399d086181a9f513cd95157e815551f0be9b9e95

Change-Id: I2bd2782db41b58634627706b96238fcf8297547a

Change-Id: I0b370ac227bbd833078804d8a276b48666df734c

Change-Id: I9c9087bcd66d1383d08f49099f8d050cff5adb5b

Change-Id: Ie8667b1119cb60936a1a522086b1c99d0f4066b7

Change-Id: If17439ae83eb063a7ab054c3701e23dd48f1edd1

Change-Id: Iecee21acd140ce12ee34ac077b1450205cc28618

Change-Id: I8868e19a75376d5abf2d62eda1774c9bd2defa4c

Change-Id: I18868b542524f0e94ec7b06e753ee2385ede9a95

Change-Id: I5dfc43bdd4d8490a47c718dc49acba0ca5f7b526

Change-Id: Ib89d8f72fc7e874968c2d9481df80b53dc982f62

Change-Id: Ic27d7728025b3d666c0ebaf9ec4cd2006d0c89bf

Change-Id: I2d3120a173097208bd29eb52ab3dea3af870f6ae

Change-Id: I3ee442ab6dac31eb7daac32e7a9ed5c6526f07ba

The existing situtation of not having any explicit rules for header
files does not work because it requires a make restart after
the headers are generated in order for the headers to be delivered.

Because requiring running make twice to get a complete rebuild is bad,
add some rules to force the headers to be delivered immediately.

Change-Id: I5b4d5c8f8e9c9d7d0874fc797e62972eaa1dd904

except moving ScPostIt into source/filter/inc/xeescher.hxx

Change-Id: I85fbfe88e30edce5a48a65c494ced7b2129964ff

Change-Id: Ibb3918b95c2518fd83f566bb726273dbf90cf8c8

The docx spec doesn't say what is the default value, the rtf spec says
it's 720, not 1440.

Change-Id: Icb331591d4f2f96a7786f808d99af5974e645f8e

The Lanczos scaling is of very good quality, but it's rather slow,
which can be very noticeable with large images, so it's not a very
good default for everything. And in general, it's not good to refer
to a specific algorithm when all one usually wants is fast/default/best.

Some of these changes are a bit of a guess between default/best,
but the general logic is that best should be used only for images
that won't be large or where the possible waiting does not matter.

Change-Id: I53765507ecb7ed167890f6dd05e73fe53ffd0231

    Use ksc5601.h header from XFree86 Project Inc. Patch contributed by
    Oliver-Rainer Wittmann with minor changes from Pedro Giffuni
    http://svn.apache.org/viewvc?view=revision&revision=1179296

The used bilinear algorithm is apparently of a rather low quality,
use the new generic box algorithm instead. That makes it somewhat
slower, but the result is cached, and hopefully the speed difference
is not that significant.

Change-Id: I5a4dbe4851d467babc0d0fdcc3375b35441daf93