It could not be used to save whole objects, so get rid of it and use
onRetainNonConfigurationInstance() exclusively.
Change-Id: I7cc3951b60a054b5b2488b49791b58707dc8a68b
It expands to what lo_get_app_data_dir() returns, i.e. the activity's
getApplicationInfo().dataDir.
Use it in the LibreOffice4Android app's rc files instead of a
hardcoded (possibly device- and/or OS version dependent) path.
Change-Id: I8145f2c86eeded39232fb251a79fa64f31f77f55
This means they will have to be uncompressed and copied into actual
files when the app runs for the first time, but that is a small price
to pay for reducing the .apk size with over ten megabytes.
Change-Id: I5bcf3c228197519edd1ca8cbd804b5f027d7699d
We gzip them separately in the Makefile and the gzipped result will be
stored without (further) compression in the .apk.
Use this to store the ttf font files. Shaves off a bit .apk size.
This might seem a bit odd way to do it, why not store these files in
the normal Zip compressed fashion in the .apk? It seems hard to tell
Ant (based on path, not extension) what files to compress and what
not, so we have to keep telling it to not (further) compress any files
at all.
Change-Id: I0d40d8811e6c9df6b28c285845b1db225507f5d4
It seems that previously onCreate served 3 different purposes:
1) bootstrap UNO machinery
2) load the document
3) create the view
Ideally it should do the first two only if not done already -- as a
first step, this commit extracts 1) and does only once.
Change-Id: I087a6b934c05295263080f571c95ef152e4602cb
Use a version script ("version map") that exports only the Java_* and
JNI_OnLoad symbols that the JNI machinery needs. No non-dynamic
symbols are needed (in the .so that goes into the .apk; the one kept
locally for debugging is not stripped).
Change-Id: Ie874e59c593ec9e5d08ba369612cef1a3ea85fe4
Note that with having a targetSdk >= 11 and minSdk <= 10, we can enable
hw acceleration without loosing compatibility with 2.x devices.
Also, remove android:configChanges="orientation", as it seems to improve
rotation experience a bit.
Change-Id: Ia3be9bb3b998d604a032eb5c60f9bb0ec71dfc68
No need to load any native libraries here any more.
For now, using the same native-code.cxx as over in ../DocumentLoader
works. In other words, we end up with the same code in
liblo-native-code.so. This will obviously change if/when the UNO APIs
that are called start to differ.
Change-Id: Ia47ce833c5a290a93986005f92e981c89f63bf71
It doesn't seem to be any slower. (Note that we are *not* using
-Wl,--whole-archive around these libs so it won't link in them in
total, but only what's needed.)
Change-Id: I2cc090562fc910d3375c6293f1f4358c68a6f9f3
Now the DocumentLoader test app seems to work like it does with all
the separate DSOs, knock on wood. The only "attempting to load unknown
library" message I see in logcat is for libspelllo.a, and I
deliberately don't have a mapping for that. A viewer application
presumably has no use for spell checking. (I do link in the
hyphenation stuff, that affects rendering, and we want the rendering
to match LO on the desktop.)
Change-Id: I1ad9eedf352eb0af1187bdfdc71f8f1754912600
IN this branch these changes are not conditional. Unclear yet whether
this is what we finally will want to use or not. Maybe should make
these changes conditional and do this stuff in master instead?
Change-Id: I379d570a0e00648d295c675fd90eba6594ba3182
This patch remove some '@author' for Java souce files, and removes some
commented code founded when removing the '@author'.
Change-Id: Ic4fcd028a9cdbd85c693d93bcd48e41f36386d22
