makefile | 38 38 + 0 - 0 !
1 file changed, 38 insertions(+)

 do not pass -m64 to alpha, ia64, loong64 or riscv64 builds.

makefile | 6 3 + 3 - 0 !
1 file changed, 3 insertions(+), 3 deletions(-)

 do not assume python points to python3.

makefile | 12 12 + 0 - 0 !
scripts/genie.lua | 1 1 + 0 - 0 !
scripts/src/osd/sdl.lua | 3 3 + 0 - 0 !
src/osd/modules/file/posixfile.cpp | 6 6 + 0 - 0 !
src/osd/modules/file/posixptty.cpp | 8 8 + 0 - 0 !
src/osd/sdl/sdlprefix.h | 2 1 + 1 - 0 !
6 files changed, 31 insertions(+), 1 deletion(-)

 several ftbfs fixes for hurd systems.

docs/source/conf.py | 2 1 + 1 - 0 !
1 file changed, 1 insertion(+), 1 deletion(-)

 use the packaged version of mathjax.
 Avoid a privacy breach by using the packaged version of
 libjs-mathjax in the generated sphinx docs.

src/emu/xtal.h | 4 2 + 2 - 0 !
1 file changed, 2 insertions(+), 2 deletions(-)

 fix ftbfs on power "not a constant expression"
There are some explanation upstream :
https://github.com/mamedev/mame/issues/3157
and this probably due to the fact that IBM 128bit long double format
is not constant folded.
I slighlty rewrote ""_kHz_XTAL(long double clock) and ""_MHz_XTAL(long double clock)
the way ""_kHz_XTAL(unsigned long long clock) and ""_MHz_XTAL(unsigned long long clock)
which makes the compiler happy.
Also including an upstream change about rounding to get same results as on x86
: https://github.com/mamedev/mame/pull/5164

3rdparty/lzma/C/Aes.c | 2 2 + 0 - 0 !
3rdparty/lzma/C/AesOpt.c | 2 2 + 0 - 0 !
3rdparty/lzma/C/Sha256.c | 2 2 + 0 - 0 !
3rdparty/lzma/C/Sha256Opt.c | 2 2 + 0 - 0 !
3rdparty/lzma/CPP/7zip/Crypto/MyAes.cpp | 2 2 + 0 - 0 !
5 files changed, 10 insertions(+)

 disable hardware acceleration support on armel

3rdparty/lzma/C/7zCrc.c | 5 3 + 2 - 0 !
1 file changed, 3 insertions(+), 2 deletions(-)

 guard arm v8 feature from old architecture