On 11/10/2023 17.48, Akihiko Odaki wrote:
On 2023/10/11 17:51, Daniel P. Berrangé wrote:
On Wed, Oct 11, 2023 at 04:03:09PM +0900, Akihiko Odaki wrote:
Make qemu-plugin.h consumable for C++ platform.
Signed-off-by: Akihiko Odaki <akihiko.od...@daynix.com>
---
docs/devel/tcg-plugins.rst | 4 ++++
meson.build | 2 +-
include/qemu/qemu-plugin.h | 4 ++++
tests/plugin/cc.cc | 16 ++++++++++++++++
tests/plugin/meson.build | 5 +++++
tests/tcg/Makefile.target | 3 +--
6 files changed, 31 insertions(+), 3 deletions(-)
create mode 100644 tests/plugin/cc.cc
diff --git a/docs/devel/tcg-plugins.rst b/docs/devel/tcg-plugins.rst
index c9f8b27590..984d8012e9 100644
--- a/docs/devel/tcg-plugins.rst
+++ b/docs/devel/tcg-plugins.rst
@@ -283,6 +283,10 @@ run::
160 1 0
135 1 0
+- contrib/plugins/cc.cc
+
+A pure test plugin to ensure that the plugin API is compatible with C++.
+
IMHO we don't need to be adding a test just to prove the
existance of the G_BEGIN_DECLS/G_END_DECLS macros in the
plugin header.
Strictly speaking, if you include this header file from C++, the code will
be interpreted as C++ instead of C but with C linkage. That worries me that
the header file may get something not allowed in C++ in the future.
I think it should be enough if you add the G_BEGIN_DECLS/G_END_DECLS macros
here. QEMU is a C project, and it was quite difficult to get rid of the C++
code again, so please don't soften the check in meson.build and don't
introduce new .cc files.
If you have some code somewhere that uses C++ for plugins, I think it would
be better to add a regression test there instead.
Thomas
