Cole Munz 0b40585124 Set the stringop-overflow flags before add_subdirectory(deps) ...

With the linker pin in place the proxspace job gets past the bootrom and
dies later instead, in the cmake build:

  client/deps/jansson/load.c:193:35: error: writing 1 byte into a region
  of size 0 [-Werror=stringop-overflow=]

client/CMakeLists.txt already turns that false positive off for GCC 10
and newer, but it does it at line 876 while add_subdirectory(deps) runs
at line 295. CMake copies the calling scope into a subdirectory at the
add_subdirectory call, so anything set afterwards never reaches the
bundled deps, and jansson is built without the suppression that the rest
of the client gets.

Move the block above the add_subdirectory call in both client and
experimental_lib. Checked by configuring before and after: 0 of 17 deps
targets carried -Wno-error=stringop-overflow in their flags.make before,
17 of 17 after.

ubuntu-cmake and macos-cmake stay green either way; their compilers do
not emit this particular false positive, which is why the gap went
unnoticed while the bootrom link was failing first.

Signed-off-by: Cole Munz <Munzzyy1@proton.me>

2026-08-19 13:18:46 +02:00

..

example_c

cppcheck unused variable

2025-03-23 23:35:34 +01:00

example_lua

pm3_console() in Python/Lua/C: replace passthru by capture and quiet

2024-10-29 21:13:38 +01:00

example_py

experimental lib & client: force symlinks

2025-03-24 09:18:21 +01:00

.gitignore

SWIG experiments

2020-11-10 23:27:21 +01:00

00make_swig.sh

lib scripts

2020-11-13 22:40:08 +01:00

01make_lib_continue.sh

SWIG experiments

2020-11-10 23:27:21 +01:00

01make_lib.sh

experimental_lib/01make_lib.sh

2021-05-22 23:39:40 +02:00

CMakeLists.txt

Set the stringop-overflow flags before add_subdirectory(deps)

2026-08-19 13:18:46 +02:00