- 22 Haz, 2012 40 kayıt (commit)
-
-
Michael Stahl yazdı
Change-Id: I399d086181a9f513cd95157e815551f0be9b9e95
-
Michael Stahl yazdı
Change-Id: I2bd2782db41b58634627706b96238fcf8297547a
-
Michael Stahl yazdı
Change-Id: I0b370ac227bbd833078804d8a276b48666df734c
-
Michael Stahl yazdı
Change-Id: I9c9087bcd66d1383d08f49099f8d050cff5adb5b
-
Michael Stahl yazdı
Change-Id: Ie8667b1119cb60936a1a522086b1c99d0f4066b7
-
Michael Stahl yazdı
Change-Id: If17439ae83eb063a7ab054c3701e23dd48f1edd1
-
Michael Stahl yazdı
Change-Id: Iecee21acd140ce12ee34ac077b1450205cc28618
-
Michael Stahl yazdı
Change-Id: I8868e19a75376d5abf2d62eda1774c9bd2defa4c
-
Michael Stahl yazdı
Change-Id: I18868b542524f0e94ec7b06e753ee2385ede9a95
-
Michael Stahl yazdı
Change-Id: I5dfc43bdd4d8490a47c718dc49acba0ca5f7b526
-
Michael Stahl yazdı
Change-Id: Ib89d8f72fc7e874968c2d9481df80b53dc982f62
-
Michael Stahl yazdı
Change-Id: Ic27d7728025b3d666c0ebaf9ec4cd2006d0c89bf
-
Michael Stahl yazdı
Change-Id: I2d3120a173097208bd29eb52ab3dea3af870f6ae
-
Michael Stahl yazdı
Change-Id: I3ee442ab6dac31eb7daac32e7a9ed5c6526f07ba
-
Michael Stahl yazdı
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
-
Iain Billett yazdı
-
Takeshi Abe yazdı
except moving ScPostIt into source/filter/inc/xeescher.hxx Change-Id: I85fbfe88e30edce5a48a65c494ced7b2129964ff
-
Takeshi Abe yazdı
Change-Id: Ibb3918b95c2518fd83f566bb726273dbf90cf8c8
-
Michael Meeks yazdı
-
Iain Billett yazdı
-
Iain Billett yazdı
-
Michael Meeks yazdı
-
Miklos Vajna yazdı
The docx spec doesn't say what is the default value, the rtf spec says it's 720, not 1440. Change-Id: Icb331591d4f2f96a7786f808d99af5974e645f8e
-
Michael Meeks yazdı
-
Iain Billett yazdı
-
Luboš Luňák yazdı
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
-
Michael Meeks yazdı
-
Michael Meeks yazdı
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
-
Michael Meeks yazdı
-
Michael Meeks yazdı
-
Luboš Luňák yazdı
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
-
Luboš Luňák yazdı
If the metafile contains just one single bitmap, the drawing code optimizes by just using that one bitmap. It (=the resulting pixmap after scaling etc.) however has not been cached so far, which means smoothscaling (to be done) would be quite slow with every paint. Change-Id: I30950c55fbadfddedc7df31283c66ed064b1a1a6
-
Luboš Luňák yazdı
if( !number ) { ... ++number; } would never get beyond 1. Change-Id: Iac33df3a21280c76fcdd130b699b4ab6466b1f46
-
Luboš Luňák yazdı
The document from bnc#765998 contains a rather huge WMF with poor quality. It could be smoothscaled to get a much better result, but doing that would be too slow with an image of this size, and it would be done every single time the image would need to be painted, because the resulting image would not be cached. One reason for this is that the WMF appears to be kinda broken (or let's say imprecise [which is no wonder after trying to read things like http://msdn.microsoft.com/en-us/library/dd162607(v=vs.85).aspx and trying to understand what exactly rlcFrames is supposed to be]). Change-Id: I017db36ec96f5405ff5965943003d5aa93018a37
-
Luboš Luňák yazdı
Change-Id: Ie3c017eb9c6ceadd4b7982eb2a28c74bf6767631
-
Luboš Luňák yazdı
According to docs http://msdn.microsoft.com/en-us/library/dd162589(v=vs.85), cxDest/cyDest are "Logical width/height of the destination rectangle.", so there should be no +1 as there would be if they were the top-bottom corner. Change-Id: Iefa6db8e6131abe785b7878d97df1c891b73011c
-
Luboš Luňák yazdı
It's rather lame that Rectangle from tools primarily specifies the rectangle as topleft/bottomright, rather than topleft/size, as that easily leads to confusion such as here. Change-Id: Ice86fae90d9159b98e0896b6c875b99c3e1a3686
-
Luboš Luňák yazdı
Change-Id: I8702f8e4b76731ab167533478ba754c94114f419
-
Luboš Luňák yazdı
Change-Id: I284bc45abab57127ea972ed4d4b8164b54bfa18b
-
Miklos Vajna yazdı
Change-Id: I5f0e7aefdea80bbb9cf61b991c5b706bd2023dfa
-