Steffen Nurpmeso wrote in <20190311163800.-pRmF%steffen_at_sdaoden.eu>:
|Steffen Nurpmeso wrote in <20190311160330.GfYQR%steffen_at_sdaoden.eu>:
||Leonardo Arena wrote in <CAGG_d8CK5FJP8mBX-Lhv9G6NCNX31Z5CaZO6y5RuDQ=qdW\
|||On Fri, Mar 8, 2019 at 5:35 PM Steffen Nurpmeso <steffen_at_sdaoden.eu[/\
|||build fails for me with:
||I have never seen these errors:
|||main.c:(.text.startup+0xbb): undefined reference to `su_state_set'
|I see now, -Os; yes, we now require inline functions.
|This is what happens with automatic heuristics; i for one write
|INLINE when i want to have a function inlined, otherwise i do not.
|I do not want to reorganize sources and move statics to after
|their use cases in order to avoid crappy heuristics taking place
|and inline functions by themselves. And even that will no longer
|work with link time optimizations and whatever they are now doing.
|But isn't that a compiler bug, i can use EXTRA_CFLAGS=-O3 which
|sets -finline-functions according to  and works just fine?
That is to say that -Os also sets this according to :
It also enables -finline-functions, causes the compiler to tune
for code size rather than execution speed, and performs further
optimizations designed to reduce code size.
So it seems maybe some pieces of cake think inlining is done
whereas others do not.
This leads me to that terribline "inline" behaviour of ISO C all
in all. In C++ if a compilation unit needs that thing or an
address thereof it simply creates an instance, damn. And
_Static_assert is also c..p compared to C++ static_assert or so.
S....y d..n ISO C decisions, all over the place.
|  http://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html
|Der Kragenbaer, The moon bear,
|der holt sich munter he cheerfully and one by one
|einen nach dem anderen runter wa.ks himself off
|(By Robert Gernhardt)
Received on Mon Mar 11 2019 - 17:41:37 UTC