Microsoft and other non-GNU compilers don't recognize gcc tricks
like /*NOTREACHED*/ to suppress individual warnings. clang recognizes most
of them because it tries to be gcc-compatible. Because of that, a lot of
potentially useful warnings have had to be completely suppressed in the
past in all source files when using the non-gcc compatible compilers.
Now that the code is C99, take advantage of a way to suppress warnings for
individual functions, a big step up from suppressing the warnings
altogether.
Unfortunately, it does require a bit of ugliness caused by the
insertion of some macros in a few spots, but I'm not aware of
a cleaner alternative that still allows warnings to be enabled
in general, while suppressing a warning for known white-listed
instances.
Prior to the warning-tiggering function, place whichever one of
the following is needed to suppress the warning being encountered:
DISABLE_WARNING_UNREACHABLE_CODE
DISABLE_WARNING_CONDEXPR_IS_CONSTANT
After the warning-triggering function, place this:
RESTORE_WARNINGS
Under the hood, the compiler-appropriate warning-disabling
mechanics involve the use of C99 _Pragma, which can be used
in macros.
For unrecognized or inappropriate compilers, or if
DISABLE_WARNING_PRAGMAS is defined, the macros expand
to nothing.
After the most recent round of moving old stuff to 'outdated',
src/windows.c contained two references to non-existent files.
That broke 'make depend'. Updating it to turn those two into
comments seems risky because someone might add an include for
some new interface later in the file. So comment them out in
the source instead. Also, redo previous 'make depend' update
from about three weeks ago to do the same thing.
get the warning flags on linux and macOS into better alignment
move the compiler flags into hints/include/compiler.2020 so that
maintenance changes can be made there and take effect for both
linux and macOS
this will likely trigger some initial new code warnings to be
investigated on either or both platforms
The inclusion shouldn't really come before *conf.h because the
things it sets up could depend on those. This is better, but
having Amiga 90% dead and 10% comatose seems to be more trouble
that it's worth.
The gone but still referenced state of amiconf.h broke 'make depend'.
Fixing that turns it into a comment in the generated dependencies
but that ran into a problem with it being followed by fnamesiz.h
instead of being last in the list containing it. So in addition to
the depend fix, move #include "fnamesiz.h" from config.h to global.h
in order to have it come before amiconf.h.
Provide a reasonably straightforward way to leave lua alone when
changing git branches (which I precede with 'make spotless').
make clean-keep-lib
and
make spotless-keep-lib
are new alternate forms of top level 'make clean' and 'make spotless'
that won't touch lib/lua/.
Remove all references to the unused vision tables in the main source
and unix build. Leave makedefs able to generate the vision tables.
makdefs will be cleaned up in a different commit, once all ports
are clear of dependencies.
Add a missing update to sys/unix/Makefile.top. Makefile.dat
only requires that tiles2bmp exist, without knowing anything
about whether it needs to be rebuilt. So force Makefile.top to
make sure that it's up to date, similar to how tiles2x11 gets
handled.
I've been ignoring submodules so far. For the old method of
dealing with lua, the instructions
You might need to do
make spotless
make fetch-lua
aren't adequate. They should be
When lua version has changed in Makefile.top, before running
setup.sh to put that new Makefile in place, do
make spotless
then
sys/unix/setup.sh [hints/...]
make fetch-lua
otherwise it will try to clean up the not yet fetched new lua
version instead of the old one.
I was looking into adding a confirmation prompt for '!' and it
isn't very promising due to sequencing issues. (The check for
whether '!' is allowed should happen before the prompt about
running it but the latter should take place in the core rather
than in the port code.) In the mean time, I noticed that VMS was
ignoring the SHELLERS value from SYSCF.
Untested implementation of a SHELLERS check on VMS. Even if it
works, it should not be using $USER as the user name to verify.
Tweaks the Unix implementation of check_user_string() but doesn't
switch the testing loop to the simpler version used by VMS which
is derived from the generic users test used by Qt.
Because some Makefile.top dependencies triggered the build
of the host native lua library, the cross-compiles were
building it needlessly.
Make it a make variable so that it can be overridden by
cross-compile recipes in sys/unix/hints/include/cross-*.2020
Replace use of $(LINK) with $(CLINK) or $(CXXLINK) as warranted.
When the Qt interface is enabled, the utility programs were all
(except dlb) being linked with C++ support. That didn't cause
any problems, just looked wrong. Link them as C instead of C++.
Two actually do need C++ support (and still have it) but both
are dead: 'tile2beos' because the source file doesn't exist (not
even in 'outdated'), 'tileedit' because it won't build with Qt5.
I didn't bother with QUIETCC support for them.
There were still a couple of references to dgn_comp (for the lint
target; just in the name of a macro, not its value); remove those.
