The value of diffx() had the wrong sign when the arguments differed by
WORLD_X / 2. Same for diffy() and WORLD_Y / 2. satmap() used them to
find the vector from map center to ship or land unit to put on the
map, and got incorrect values for ships and land units directly
opposite to the center in x or y. The bug made satmap() read a
pointer out bounds of its malloced radbuf[], and then write through
that with unpredictable consequences.
Broken in 4.2.12. The original bug was in Empire 1.1: it
miscalculated where to put ships on the map (no crash). An incomplete
fix for radmap() and satmap() appeared in Chainsaw 2 (still no crash).
radmap() got fixed correctly in Chainsaw 3, but satmap() was
forgotten. That one got "fixed" in 4.2.7, and again in 4.2.12, but
both "fixes" were flawed and could crash.
Fix by backing out the flawed fixes and adopting the fix from radmap()
instead.
Commit da8a1dae (v4.3.12) introduced virtual selectors, but neglected
to update xundump. Xundump can't work for them, because they don't
provide a setter method.
This didn't actually break anything, because all virtual selectors
have flag NSC_EXTRA set, or are in table EF_VERSION, which xundump
refuses to touch.
Make deffld() oops on virtual selector, just to be safe.
When fixing planes stuck in the air, we fixed them only in memory, so
when a fixed plane wasn't written to disk for other reasons before the
next game start, it had to be fixed again.
Change pln_zap_transient_flags() to write them out.
Some losing implementations of strptime() such as FreeBSD's happily
succeed when they fully consumed the first argument, regardless of
whether they matched the full second argument or not. This causes
lines without directives to be interpreted as "next Sunday".
Work around the lossage in parse_time().
This is because we want to define them in src/lib/global/, and code
there can't use getnatp(), because that requires
src/lib/common/file.c. Which renders a cnum parameter pretty useless.
Virtual selectors requiring code from common/ could well come up again
in the future, but let's not worry about that now.
Planes normally sit in their base (sector or carrier), where they can
be spied, damaged, captured, loaded, unloaded, upgraded and so forth.
All this must not be possible while they fly. There are two kinds of
flying planes: satellites in orbit, and planes flying a sortie.
Satellites in orbit have always been marked with flag PLN_LAUNCHED.
Works. What didn't work was tracking planes flying a sortie.
If you look at one sortie in isolation, up to three groups of planes
can be flying at any point of time: the primary group, which carries
out the sortie's mission (bomb, transport, ...), their escorts, and a
group of hostile planes flying interception or air defense.
The old code attempted to track these planes by passing those groups
to the places that need to know whether a plane is flying. This was
complex and incomplete, and broke down completely for the pin-bombing
command.
It was complex, because the plane code needs to keep track of all the
call chains that can lead to a place that needs to know whether a
plane flies, and pass the groups down the call chains. This leads to
a rather ugly passing of plane groups all over the place.
It was incomplete, because it generally failed to pass the escorts.
And the whole scheme broke down for the pin-bombing command. That's
because pin-bombing asks the player for targets while his planes are
loitering above the target sector. This yields the processor and lets
other code run. Which does not get the flying planes passed.
The new code marks planes and SAMs (but not other missiles) flying a
sortie with flag PLN_LAUNCHED (the previous commit laid the groundwork
for that), and does away with passing around groups of flying planes.
This fixes the following bugs:
* Many commands could interact with foreign planes flying for a
pin-bombing command as if they were sitting in their base. This
includes spying, damaging, capturing, loading, or upgrading them,
and even getting intercepted by them. Any changes to those planes
were wiped out when they landed. Abusable.
* The bomb command could bomb its own escorts, directly (pin-bomb
planes) or through collateral damage, strategic sector damage,
collapsing bridges or nuke damage. The damage to the escorts was
wiped out when they landed.
* If you asked for a plane to fly both in the primary group and the
escort group, you got charged fuel for two sorties instead of one.
* pln_put1() and pln_put() now recognize planes that didn't take off,
and refrain from making them land. Intercept (since commit
c64e2149) and air defense can do that. Making them land had no
ill-effects, but it was still wrong.
There's one new problem: if PLN_LAUNCHED doesn't get reset properly,
due to game crash during flight or some other bug, the plane gets
stuck in the air. Catch and fix that on game start in ef_verify().
Make deffld() reject fields whose selector has flag NSC_EXTRA set.
Since xundump() doesn't provides space for these, the bug could lead
to buffer overruns.
This is for the forthcoming empdump utility program. The xdump
command still does not support the human-readable dialect.
xundump has supported the human-readable dialect for a long time
(commit 4871a10a).
Where ordinary selectors specify a value stored in some object,
virtual selectors specify a function to call to compute a value
associated with some object.
Use them to replace the special case xdump ver by new table
EF_VERSION.
Move configkeys[] to lib/common because nsc_init() needs it to
initialize empfile[EF_VERSION].cadef.
New struct emptypedstr to avoid depending on empobj.h there.
Remove now superfluous manual initializations elsewhere.
This doesn't fix any missing initializations.
Use ef_ensure_space() in getobj(). This also makes sure objects are
properly initialized before undumping writes to them.
Clean up how sentinels are appended: instead of keeping its slot
reserved while undumping, keep it in the table, and strip it off when
done.
ef_extend() extended the file bypassing the cache, which screws up the
cache if it's EFF_MEM. It fixed that by closing and reopening the
table. Cheesy, and worked only for file-backed tables.
Rewrite ef_extend() to remap the cache properly for EFF_MEM. While
there, simplify the !EFF_MEM case: steal a cache slot instead of
allocating a buffer.
Factor cache mapping out of ef_open() and ef_extend() into new
ef_realloc_cache().
Read-only was a bit of a misnomer: you could write to the table by
obtaining a pointer into it from ef_ptr(), you just couldn't write to
the backing file.
Semantic changes:
* ef_flush() is now allowed when the table is file-backed or privately
mapped. Before, it had to be file-backed. Flushing a privately
mapped table does nothing, just like flushing a read-only table did.
* ef_write() is now allowed when the table is file-backed or fully
cached. Before, it had to be file-backed and not read-only.
Writing to a privately mapped file-backed table doesn't write to the
file.
* ef_extend() is not implemented for privately mapped tables, just
like it wasn't implemented for read-only tables.
Split ef_init() into two functions: empfile_init() for initialization,
and empfile_fixup() to fix it up for configuration. Put them next to
empfile[]. Move the call to empfile_init() from behind emp_config()
to before it.
Windows code leaked memory (result of _fullpath()).
POSIX code passed a null buffer to getcwd(), which is not portable,
and failed to check for errors.
read_builtin_tables() wanted to run in builtindir, and
read_custom_tables() wanted to run in configdir. Bothersome. Use new
fopenat() to relax those requirements.
The chdir() satisfying them are now superflous, remove them.
File names in econfig need to be interpreted relative to configdir.
This wasn't the case everywhere for keys data and info.
Fix this by changing variables gamedir and infodir to hold absolute
names. Change builtindir likewise, for consistency. Store the values
from econfig in gamedir_conf, infodir_conf and builtindir_conf.
Uses new fnameat() to derive absolute names from possibly relative
ones.
Move stuff to untangle the ugly cyclic dependencies between the
archives built for selected subdirectories of src/lib/:
* Move common/io.c to empthread/ because it requires empthread stuff
* Move parts of subs/nstr.c to common/nstreval.c to satisfy
common/ef_verify.o
* Move getstarg.c getstring.c onearg.c from gen/ to subs/ because they
require stuff from there
* Move bridgefall.c check.c damage.c empobj.c journal.c maps.c
sectdamage.c from common/ to subs/ because they require stuff from
there
* Move cnumb.c from subs/ to common/ to satisfy common/type.o
* Move log.c fsize.c from common/ to gen/ because they really belong
there
* Move emp_config.c mapdist.c from gen/ to common/ because they really
belong there, and require stuff from libglobal.a
Also package as/ as libas.a to satisfy common/path.o.
Remaining dependencies:
lib needs
--------------------------------------------
libas.a libglobal.a
libcommon.a libas.a libglobal.a libgen.a
libgen.a
libglobal.a
liblwp.a libgen.a
libw32.a[*] libgen.a
[*] Except for service.o, which can only be linked into the server
Link order now: liblwp.a libcommon.a libas.a libgen.a libglobal.a
libw32.a. The position of libw32.a is not quite right, but works
anyway.
There are several files with land unit subroutines. This one is in an
awkward place: it depends on stuff from ../subs, which contributes to
libcommon.a's ugly dependencies. Move its contents to logical places
(use internal linkage where possible), and remove it.
