| 22994af5 | 24-Oct-2026 |
Pedro Giffuni <10236985+pgiffuni@users.noreply.github.com> |
Merge pull request #512 from pkubaj/ppc64-fix
FreeBSD PowerPC64: fix the C++-UNO bridge, build on little-endian
I am happy with this change as it completes the series of support for different FreeB
Merge pull request #512 from pkubaj/ppc64-fix
FreeBSD PowerPC64: fix the C++-UNO bridge, build on little-endian
I am happy with this change as it completes the series of support for different FreeBSD platforms, and AOO is likely to get shipped and well tested there.
We should check why BridgeTest was disabled for PowerPC: sadly we don't have resources to look at all PPC platforms.
show more ...
|
| 90aa4121 | 22-Oct-2026 |
Piotr Kubaj <pkubaj@FreeBSD.org> |
FreeBSD PowerPC64 bridge: return small structs in registers
The bridge used a hidden return pointer for every struct, but under the
ELFv2 ABI structs of up to 16 bytes come back in r3/r4, and struct
FreeBSD PowerPC64 bridge: return small structs in registers
The bridge used a hidden return pointer for every struct, but under the
ELFv2 ABI structs of up to 16 bytes come back in r3/r4, and structs made
of up to eight floats or doubles come back in f1-f8. Calling a C++
method that returns such a struct from Basic, Python or Java, e.g.
getCellAddress() or getPosition(), passed the return buffer where the
callee expected this, and the testtools bridgetest crashed.
Use return_in_hidden_param() for both call directions and for the vtable
snippets, pass f1-f8 to MapReturn(), and load f1-f8 in
privateSnippetExecutor when a C++ caller expects such a float struct.
With this the bridgetest passes on FreeBSD 15.1 powerpc64 and
powerpc64le, in-process and over URP, against both the C++ and the Java
test objects.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
show more ...
|
| 78c44356 | 22-Oct-2026 |
Pedro Giffuni <pfg@apache.org> |
Fix missing variable reference in FreeBSD ARM bridge makefile
The make condition had a missing $ before OS, causing the build
configuration block to never match. This would result in an empty
bridge
Fix missing variable reference in FreeBSD ARM bridge makefile
The make condition had a missing $ before OS, causing the build
configuration block to never match. This would result in an empty
bridge library being built.
show more ...
|
| 32049e8c | 22-Oct-2026 |
Pedro Giffuni <pfg@FreeBSD.org> |
Fix FreeBSD aarch64 C++-UNO bridge stack handling for AAPCS64
Ensure each stack argument occupies a full 8-byte slot per AAPCS64
requirements. Small types (BYTE, SHORT, LONG, BOOLEAN, etc.) are now
Fix FreeBSD aarch64 C++-UNO bridge stack handling for AAPCS64
Ensure each stack argument occupies a full 8-byte slot per AAPCS64
requirements. Small types (BYTE, SHORT, LONG, BOOLEAN, etc.) are now
sign/zero-extended into a sal_uInt64 scratch slot before being copied
to the stack, preventing garbage values from uninitialized upper bytes.
This fixes the testPackedStack BridgeTest failure where nStack was
24 instead of the expected 48, and resolves the cascading transportAny
crash caused by corrupt stack values.
show more ...
|
| a95bc489 | 21-Oct-2026 |
Pedro Giffuni <pfg@FreeBSD.org> |
freebsd_aarch64: Fix stack alignment and return handling
Fix two issues in the FreeBSD aarch64 C++-UNO bridge:
1. abi.cxx: align_stack_offset and stack_size now use the type's
natural alignment
freebsd_aarch64: Fix stack alignment and return handling
Fix two issues in the FreeBSD aarch64 C++-UNO bridge:
1. abi.cxx: align_stack_offset and stack_size now use the type's
natural alignment (nAlignment) and raw size (nSize) instead of
hardcoded 8/alignment and 8-byte slot rounding. This matches the
AAPCS64/System V ABI requirement that stack arguments use their
natural alignment and size, matching the macOS reference.
2. uno2cpp.cxx: Restructure the return-value switch to use 'default'
with a bSimpleReturn check (matching the macOS reference), so
fill_struct is only called for register-returned aggregates, not
for types that use the x8 hidden-result buffer.
show more ...
|
| 9ff3e225 | 03-Aug-2026 |
Jim Jagielski <jimjag@gmail.com> |
Synthesise UNO exception RTTI on macOS/arm64
RTTI::getRTTI() in the arm64 C++-UNO bridge resolved exception
type_info via dlsym(RTLD_DEFAULT, "_ZTIN...E"). On arm64 Darwin this
can never succeed: cl
Synthesise UNO exception RTTI on macOS/arm64
RTTI::getRTTI() in the arm64 C++-UNO bridge resolved exception
type_info via dlsym(RTLD_DEFAULT, "_ZTIN...E"). On arm64 Darwin this
can never succeed: clang emits the typeinfo of any keyless class -
which is every UNO exception - as hidden regardless of visibility
attributes, version scripts, or export lists, whereas the same class
gets exported typeinfo on x86_64 Darwin.
show more ...
|
| 079530e0 | 01-Aug-2026 |
Peter <leginee@users.noreply.github.com> |
Win64 bridge backport (#492)
* bridges/msvc_win64: pass simple argument values, not the temp's address
uno2cpp.cxx converts each simple by-value argument into an 8-byte alloca
temp and stores the p
Win64 bridge backport (#492)
* bridges/msvc_win64: pass simple argument values, not the temp's address
uno2cpp.cxx converts each simple by-value argument into an 8-byte alloca
temp and stores the pointer in pCppArgs[nPos]:
uno_copyAndConvertData( pCppArgs[nPos] = alloca( 8 ), ... );
Copying the value into the outgoing slot therefore has to dereference that
pointer. The marshalling switch used &pCppArgs[nPos] instead, which reads
the pCppArgs array slot itself, i.e. the temp's address -- so every HYPER,
LONG, ENUM, SHORT, CHAR, BOOLEAN, BYTE, FLOAT and DOUBLE parameter reached
the callee as a stack address (or its low 16/32 bits) rather than a value.
The same expression is correct in the complex/ref branch, where the pointer
IS the argument; written there as (sal_uInt64)pCppArgs[nPos] to make the
difference between the two branches obvious.
This only affects calls that cross between the uno and cpp environments, so
a pure-C++ session never trips it; it surfaces via pyuno and the invocation
adapters.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* bridges/msvc_win64: return queryInterface's Any in the hidden return buffer
The Win64 call stack is "ret addr, this, [ret *], params", so a method with
a hidden return pointer finds it at pCallStack[2] -- cpp2uno_call() reads
pCppReturn from [2], and the queryInterface shortcut in cpp_mediate() reads
its Type argument from [3] accordingly.
That shortcut nevertheless built the returned Any at pCallStack[1] and
returned [1] in the register slot. [1] is `this`, so it constructed a
24-byte Any over the proxy object and never wrote the caller's own return
buffer. The caller then read and destructed an uninitialised Any, faulting
in uno_any_destruct() on a stale stack value used as a
typelib_TypeDescription*.
The x86 bridge has always used pCallStack[2] here; the Win64 port carried
over the x86 index, where the layout puts the hidden pointer before `this`.
Reproducer: any queryInterface on a raw uno_Interface proxy -- e.g. pyuno's
Adapter::getOutIndexes() inspecting an InvocationAdapterFactory adapter,
which crashed on every use of the Python macro organiser.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* pre-commit: fix whitespace in the msvc_win64 bridge sources
The x64 bridge backport in ebdb555ed2 brought these files over from
origin/windows-amd64, a branch predating the tree-wide whitespace
cleanups in #383 and #462, so it silently reintroduced trailing
whitespace in the license headers of call.asm, cpp2uno.cxx and
except.cxx plus a stray blank line at the end of except.cxx.
Since the pre-commit workflow runs with --all-files, this fails the
check on every pull request against trunk, not just ones touching
these files.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
show more ...
|
| 9d1a529d | 08-Jun-2026 |
Jim Jagielski <jimjag@gmail.com> |
Fix crash, uninitialized-return, and buffer bugs in update check and macOS x86 bridge
updateprotocol.cxx: guard XNodeList dereferences so a swallowed
XPathException no longer leaves a null reference
Fix crash, uninitialized-return, and buffer bugs in update check and macOS x86 bridge
updateprotocol.cxx: guard XNodeList dereferences so a swallowed
XPathException no longer leaves a null reference to be dereferenced
(crash on malformed/hostile update descriptors); clear the list before
the relnote query so a thrown exception can't leave us iterating the
previous (sources) node list.
cpp2uno.cxx: the queryInterface fast-path guard used 'break', which
exited the switch and skipped the default case, returning an
uninitialized typelib_TypeClass. Guard the block with
'if (bHasHiddenReturn)' instead so control falls through to default and
eRet is always assigned.
updatecheck.cxx (getImageFromFileName): use bitwise '|' instead of
logical '||' when combining osl_Process flags; rewrite the trailing
CR/LF trim so it never reads/writes before the start of the buffer
(e.g. when osl_readFile returns 0 bytes).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
show more ...
|