merge together a gadzillion hand-crafted
isSimpleChinese/isTraditionalChinese/isKoreanVariants/isCJK
implementations which should fix a goodly amount of them
add a MsLangId::isFamilyNameFirst for locales where
family name appears first while I'm at it.
Change-Id: I65377793be037d16fe7250cd7450b28aec689e83
Another attempt, whoever has written this apparently didn't consider
the possibility of recursion at all, and this still feels a bit hackish.
Writerfilter keeps just one oox::shape::ShapeContextHandler object during
the entire time of parsing the document, because e.g. <v:shapetype> needs
to be reachable even across VML block (see sw testcases for bnc#705956).
This however presents a problem when VML contains <w:txbxContent> which
contains another VML, as this code previously just took whatever has been
read and returned it to writerfilter, and it broke with recursion.
So now try to mark recursion entry and returns the right shape.
Related to 36c12c246d886b2d96d7a2d4d0c250db9d925c74 and the previous
commits it reverted.
Change-Id: I949a6b52ec7540aa59b047c7b6e908b10fb3bdc1
Stupid containers with shared_ptr. Trying to remove the object
from it gets it deleted because it's owned by the shared_ptr
and there's no sensible way to wrestle it out of it. This will
need to be redone somehow.
This reverts commit 04d600d4be.
Pass on to VirtualDevice where used to set the MapMode of the device
appropriately. Adapt DocumentLoader, use to scale the page rendering
to exactly fit the virtual device.
Change-Id: I4b0bc67e12114d3d9d493ff1aca2ef5d2cc78912
operator() gets called in a Java GC thread. A JNIEnv* saved in the
constructor (which is called from the UI thread) is not valid in other
threads (although for now, Dalvik notices, warns, and works around
it). So don't bother keeping the JNIEnv*. Instead fetch one when
needed. Move the method implementations from inline in the header to
libotouch's android.cxx.
Change-Id: I7de6fc54bb8d9a59146576d6e8c325fe917393ee
1) Set FormatsSupplier property only if underlying object has one.
Else, exception is thrown and the format is not set
2) getTyperelatedFieldData uses the format keys, so initialize them
before call, not after.
Change-Id: I68c4c96a9da9a6afdc3ab8964e973588f53ee814
The About dialog now houses these functions.
Change-Id: I7f751027564897a063d2910644c69680d471b6d9
Signed-off-by: Fridrich Štrba <fridrich.strba@bluewin.ch>
there's no way j was intended to be used as a rtl_TextEncoding
here. From context it's clearly meant to create a UniString from
a subset of the input ByteString
Change-Id: Ic0a0773f90686d73ec4eb8b8e03a454585758578
Seems very awkward to create an entire temporary O[U]String just to compare it
to another when we want to know if that other contains just one char of a given
value
Change-Id: Iacd4e0d87dc5c5211fa06c61c55e1ed85b1e91f3
These warning arise because of the additional of varargs parameters
in Java1.6. Casting the parameter eliminates the compiler
confusion.
Change-Id: I4906bcfa2700ef80a67b79c61c6848a18e8a7168
