Noticed while investigating the message loop. If I had level files
from an interrupted game and was asked "Destroy old game?" when
starting a new one, answering 'n' left the terminal in an unusable
state. Executing 'stty sane' (invisibly since input echo was off)
repaired things but the user shouldn't have to do that.
Change unixtty.c's error() to shut down windowing if that has been
initialized. This might need some tweaking for tty, which will now
clear the screen before showing the startup error message. Other
systems besides unix use unixtty.c so are affected, but I think the
change doesn't introduce anything that should cause trouble (aside
from the potential screen erasure).
I had this in place at one point but must have accidentally undone it
before deciding that yesterday's patch was finished. Defer fetching
'pw' until it's needed.
Fixes#26
Report stated that the attempt to look up the player's username
(on Unix) failed (reason unknown) and nethack refused to allow the
player to execute the #explore command even though sysconf was set
to use character names (CHECK_PLNAME=1) instead of user names.
Setting EXPLORERS to "*" overcomes this glitch, but the fix moves
a bit of code around to honor CHECK_PLNAME before fetching username
so that that isn't necessary.
I ended up doing some formattng clean up (replace tabs with spaces,
whitespace cleanup in 'port_insert_pastebuf()'). The actual change
to fix#26 is only a few lines.
Another part of github issue 227. Casting a function pointer when
passing it to another function is iffy when lying about the return
type. tputs() expects a routine which returns int, so give it one.
Other xputc() usage is equivalent to putchar(), so define xputc()
with the same function signature as that has.
The tputs() declarations in system.h should probably be changed
(third argument is a function which takes an int rather than
unspecified parameters) but I've left them alone. I made that change
to tputs() in sys/share/tclib.c though.
NT and MSDOS changes are untested. tclib.c compiles ok with clang-
as-gcc on OSX but hasn't been tested with the port that uses it (VMS).
Change all the POSTINSTALL 'mkfontdir' to 'mkfontdir -x .lev' although
the mkfontdir version on my system didn't think $(HACKDIR)/*.lev were
font files when I built without dlb.
Also change the PREINSTALL 'cp -n win/X11/nethack.rc ~/.nethackrc' to
keep going if it fails. The linux hints use 'cp -n' for sysconf but
since it is doing so for the playground directory and 'make install'
starts out by clearing away everything in that directory, it shouldn't
fail. But some extra bullet proofing there may be warranted. Only
the initial cp is protected against clobbering an existing file; the
ownership+permission fixups that are applied to the copy of original
file still get applied to an existing one.
travis recently changed linux default dist from trusty to xenial, and bionic is next
include an additional travis linux build under bionic to eliminate surprises there
add a parameter to mkfontdir under linux to prevent it form going after .lev files
phase_of_moon and friday_13th determined using rn2() instead of local
time if fuzzing. Don't reseed using init_random() if fuzzing. Allow
set_random to be called outside of hacklib. rn2_on_display_rng uses
rn2 if fuzzing so that we have a single source of random that we can
ensure is reproducible. Implement rul() that returns a random unsigned
long. Fix bug in fuzzer handling of ntposkey which would cause us to use
unitialized values for x and y. Added command line arguments to allow
auto starting and stopping of fuzzer. Add a logging facility for the
fuzzer to use to record activity. Added some scripts used to automate
fuzzer testing on windows.
When stopping in the debugger after having called impossible, the windowing
state will have been modified since the assertion was hit. This made
examining state that caused the nhassert to fire no longer possible.
To avoid this issue, we now detect the debugger and stop in the debugger
prior to impossible.
