I observed a stack overflow in news command on my x86_64 system
running Fedora 18.
Empire 2 settled on this formula for the stack size:
stacksize = 100000
/* budget */ + MAX(WORLD_SZ() * sizeof(int) * 7,
/* power */ MAXNOC * sizeof(struct powstr));
Obviously attempts to provide space for known configuration-dependent
stack hogs.
The first hog is allegedly budget. Bogus since day one: its large
arrays were static in Empire 2, and became dynamically allocated in
Empire 3.
The second one makes some sense: powe() has a struct powstr[MAXNOC].
It also has an int[MAXNOC], which isn't accounted for.
Except for ridiculously small worlds, the second term is smaller, and
only the (bogus) first term matters.
Two hogs are missing: head() has a struct histstr[MAXNOC][MAXNOC], and
news() has a short[MAXNOC][MAXNOC]. It also calls head().
I looked for more hogs with "gcc -fstack-usage", and found none.
On my x86_64 system, a news command needs almost 107KiB of stack.
Only slightly less when compiled for 32 bit. Stack overrun for worlds
with fewer than some 320 sectors, thus unlikely to bite in real games.
Increase player stack size to 1 MiB. Using MAXNOC to size the stack
isn't worth the trouble.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Normally, git's pre-commit hook protects us from them. However, when
expected test output contains trailing white space, we have to bypass
commit hooks. Unwanted space can then slip in if you don't pay
attention. I obviously didn't; clean up.
The previous commit should reduce the need for such hook suppression.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Units owned by POGO are not in use. Giving a unit to POGO destroys
it. The opposite doesn't work, however: the unit prewrite hooks give
it right back to POGO, because efficiency is below the minimum. Make
it work by also increasing efficiency to minimum.
Note that you can't use this to create a unit that doesn't already
exist in the respective file. That's because edit's second argument
selects from existing objects only.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
SO_KEEPALIVE isn't useless with long idle timeouts. Before v4.3.30's
idle timeout changes, such long timeouts had serious drawbacks (easy
denial of service), but they're just fine since then.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Bug bites when a ship supplies the sector it's in. First the sector
is charged zero mob for moving the stuff, and is written back. Next,
the sector receives the stuff, and is written back, clobbering the
first write (no effect), and triggering a seqno oops.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
print_plane() ends its output with '\t' instead of '\n'. Next is a
prompt, which supplies the missing newline (see pr_id()).
Ugly, clean up.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Breaking sanctuary turns 100% sanctuaries into capitals, all others
into highways. Before Empire 3, sanctuaries always became capitals.
Give deities more control over initial state of countries than
choosing between 100% capital and <100% highway sectors: turn
sanctuaries into their new designation on break. Except make a
capital when the new designation is sanctuary (which it normally is).
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Accept general <SECTS|SHIPS|PLANES|LANDS|NATS> argument instead of
just <SECT|SHIP|PLANE|LAND|NAT>.
edit with <KEY> <VALUE>... arguments applies the arguments to all
selected objects. Without such arguments, edit lets you edit the
selected objects interactively one after the other.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Unlike other edits, editing a country modifies the object in-place.
It has to, because editing may send bulletins, which updates nat_tgms,
and writing back a copy would clobber these updates, triggering a
seqno mismatch oops.
However, the in-place editing neglects to detect concurrent updates.
Fix that.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Interactive edit prints the edited object, then reads player input.
If it gets key and value, it applies them to the object, and repeats.
If it gets nothing, it prints the edited object again, and stops.
Remove this last print, because it's not really useful. The object is
commonly the same after reading input as before. Except when a nation
gets updated while "edit c" is waiting for input: then the second
print actually reflects the updates. Has always been that way.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Print a message describing the actual change for keys 't' (nat_tgms)
and 's' (nat_status). The message is necessary to give the deity a
chance to catch unexpected changes, e.g. a player reading telegrams
just before the deity edits nat_tgms.
Send a bulletin to the country for key 's'.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Print "FOO of CNAME (#CNUM)" instead of just "FOO". Print "unchanged"
instead of "changed from X to X" on no-op.
Send a bulletin to the country and report news when appropriate.
Affected keys:
key struct member before after notes
-------------------------------------------
n nat_cnam -- BN 1
r nat_pnam -- B- 2
b nat_btu -- BN 3
m nat_reserve BN BN
c nat_xcap,ycap -- B-
o nat_xorg,yorg -- B-
u nat_timeused -- BN
M nat_money B- BN
Notes:
1. Reports N_NAME_CHNG rather than N_AIDS/N_HURTS.
2. Message improved to accurately reflect string truncation.
3. Greengrocers' apostrophe in message fixed.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Print "Technology" instead of "Tech". Print "FOO of CNAME (#CNUM)"
instead of just "FOO". Print "unchanged" instead of "changed from X
to X" on no-op.
Send a bulletin to the unit owner and report news on change.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Print a message (always), send a bulletin to the unit owner and report
news (sometimes).
The message is necessary to give the deity a chance to catch
unexpected changes, e.g. a player spending mobility right before the
deity edits it. Watching out for such changes is especially important
with non-interactive edit.
Affected keys:
cmd key struct member action
----------------------------------
edit s L shp_x,y B-
E shp_effic BN
M shp_mobil BN
F shp_fleet B-
T shp_tech BN
a shp_pstage --
b shp_ptime --
R shp_rpath B-
edit p l pln_x,y B-
e pln_effic BN
m pln_mobil BN
w pln_wing B-
t pln_tech BN
r pln_range B-
edit u L lnd_x,y B-
e lnd_effic BN
M lnd_mobil BN
a lnd_army B-
t lnd_tech BN
F lnd_harden BN
Z lnd_retreat B-
R lnd_rpath B-
The two characters in column action show whether the command sends a
bulletin (B) or not (-), and whether it reports news (N) or not (-).
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Print a message, send bulletin to owner. Affects ship key 'W', land
unit key 'W', and plane key 'f'. The message is necessary to give the
deity a chance to catch unexpected changes, e.g. a player modifying
retreat conditions right before the deity edits them. Watching out
for such changes is especially important with non-interactive edit.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Edit lets deities load units onto remote carriers, resulting in a
carriers having cargo in another sector. Not good. Cargo gets
teleported to its carrier belatedly when the carrier moves.
Better let edit take care of the teleport.
Also tell the deity that he just caused a teleport. Necessary to give
the deity a chance to catch unexpected changes, e.g. a player moving a
plane right before the deity edits it. Watching out for such changes
is especially important with non-interactive edit.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
They really need to stay with their carrier. Else, cargo gets
silently teleported to its carrier when the carrier moves.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Print a message, send bulletin to owner. Affects plane keys 's', 'y',
and land unit keys 'S', 'Y'. The message is necessary to give the
deity a chance to catch unexpected changes, e.g. a player loading a
plane right before the deity edits it. Watching out for such changes
is especially important with non-interactive edit.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Send bulletin to owner and report news exactly like for key 'O' or
sector key 'L'.
Print a "duplicated" message, just for consistency with other keys.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Two failure modes:
* When the argument is negative, edit() attempts an invalid write,
triggering an oops. edit() detects the write failure, and returns
RET_FAIL.
Detect the negative argument instead, and return RET_SYN, like we do
for other invalid edit arguments.
* When the argument requires the table to be extended, but the
extension fails, edit() returns RET_FAIL.
Additionally report the failure to the player.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
