Age | Commit message (Collapse) | Author |
|
Change-Id: I7d875dd74f8eff05f7a291100841779abbf5fd1c
|
|
Change-Id: I75e195be6bcedc972bc7c820c8c8efb648c91526
|
|
Evidently on Windows, the newfangled ucpp handles #include "foo"
differently from #include <foo> and treats it as a relative path, while
the angle brackets always result in absolute paths.
Since relative paths result in infinite rebuilds if make is invoked in a
different directory, don't use #include "foo" in IDL files.
Change-Id: Iedcda3a4be5542389a0be086f14541cda8dc5323
|
|
Change-Id: Ie0b0ec3120f828e02b1ba9c21710c28159e7cd3c
|
|
Change-Id: I36f59184c70b05e45929a01c2052c80cb04ed3de
|
|
|
|
Change-Id: Id4210229b20a1d54b680ff1a2ccbcdd3684ecf5f
|
|
This is a followup of :
commit e748b096633c5ef747f0630d13708d8dd7308a86
Author: Kristian Rietveld <kris@lanedo.com>
Date: Sat Oct 15 01:15:21 2011 +0200
Mark ecx register as clobbered
The inline assembly code executes a function call and functions are
free to use eax, ecx and edx without preservation. We must thus mark
ecx as a clobber register.
Change-Id: I4e9ed38189b23c2bb6d4fbdedc34c318026acfe7
|
|
Some of them were commented out for documentation purpose.
Change-Id: I29a6b2cbe774b527f0c3a0be5675160817a7f3ce
|
|
|
|
Those never get included at any point in the code.
Change-Id: I17736e005635b9e57759cfc95176a99e7c5e8988
|
|
|
|
Change-Id: If58683331c50f2a95204e8e2dea11edbef3ccb63
|
|
More than two lines are removed for readability.
Change-Id: Ibff6cf68d7c512e240a54065b54a225bb23a782b
|
|
Change-Id: If02be864f958de96ae1c54fe575b1828410390b1
|
|
Change-Id: Id2a12ec3a3c394bb40aa233f29807932b1fa84d5
|
|
since 99bf0c318b89c7d6acee364161b34b13d348a3bb
Change-Id: I0fbd88d842f750d101e3eec5a6b18f7e8bc062be
|
|
Change-Id: I50f0887ceb4517d7ef234f970ca2ba679d533382
|
|
this removes dmake completely out of the build for migrated modules
build.pl now assumes modules to be gbuild, unless there is a
prj/dmake file
Change-Id: I674a036b182ee13c5ec093e83cb3d38133112d3b
|
|
|
|
This ports 84dbc4fe2547f8fc341a46d7f000e721c81e63ee "do not let gcc use
registers we are setting ourselves," in the hope that it is also an
improvement for Mac OS X PPC. (But the patch was applied blindly, so
please revert if it makes things worse instead.)
|
|
gcc uses a register for the function call--and it tried r9 here...
|
|
The "mr" stores r11 to indeterminate register (r9 in my case) before
storing it into the output variable, thus overwriting one of our input
arguments... This later leads to nice segfault in
testtools/source/bridgetest ...
I suppose there is a better way to get the variable, but I do not know
it .-)
|
|
|
|
Had an embarrassing thinko in the assembler generation for the iOS
simulator, had forgotten the $ for a literal number. While at it, make
also the simulator read-only code snippet data structures
position-independent. Some minor comment improvements. Bridge seems to
work now on simulator.
|
|
|
|
Otherwise exception throwing seems to be broken.
This is a partial revert of b0515ea5fa6c29faebed616ae3e0213c72d24904.
|
|
|
|
...the same way cpp_uno/gcc3_linux_x86-64 was already adapted.
|
|
|
|
|
|
...the same way cpp_uno/mingw_intel was already adapted.
|
|
...for reasons that are yet unclear to me.
|
|
|
|
|
|
|
|
See comment in callvirtualmethod.cxx for details.
|
|
|
|
...which incompatibly changed the calling convention of the third (dest)
function pointer parameter of __cxa_throw to __thiscall.
|
|
This reverts commit b5ca34fa409bc4cbeb41a84b900612e302f5b0fb.
829a9b9c1041626efb688157658c1e0e2f6e8414 "In Java, css.uno.RuntimeException is
not derived from css.uno.Exception" is a better fix.
|
|
which lead to Bridge::map_to_java (jni_data.cxx) copying the Context member of RuntimeException
twice, once with a fieldID for Exception, which fails on Dalvik.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|