Markus Armbruster [Thu, 24 Feb 2011 18:29:33 +0000 (19:29 +0100)]
Speed up A* neighbor cache hits
struct sctstr members sct_x, sct_y are normalized, no need to
normalize them again.
The neighbor cache now speeds up distribution path assembly by about
10% without the path cache, and by about 5% with the path cache.
Markus Armbruster [Sun, 20 Feb 2011 07:43:44 +0000 (08:43 +0100)]
Permit disabling of A* neighbor cache at compile-time
Mostly to measure its effectiveness. Compile with
AS_NO_NEIGHBOR_CACHE defined to disable it.
The neighbor cache turns out to be useless in my tests: it eats memory
without speeding up the update's distribution path assembly.
Markus Armbruster [Sat, 19 Feb 2011 14:00:03 +0000 (15:00 +0100)]
Permit disabling of A* path cache at compile-time
Mostly to measure its effectiveness. Compile with AS_NO_PATH_CACHE
defined to disable it.
Turns out the path cache is quite effective. For my continental test
case (Hvy Metal 2 updates), it reduces the number of searches by a
factor of 18.5, speeding up distribution path assembly by a factor of
7. The price is memory: it uses 135 times more memory than the A*
library. For my island test case (Hvy Plastic 2 updates), I get 4
times search reduction, 3.5 times faster distribution path assembly,
36 times more memory.
Markus Armbruster [Sat, 19 Feb 2011 12:50:10 +0000 (13:50 +0100)]
Fix when best_path() prints A* performance statistics
Print them when A* actually runs, not when best_path() finds a path.
Statistics for unsuccessful runs were lost, and old statistics were
printed for path cache hits.
Markus Armbruster [Sat, 19 Feb 2011 12:19:27 +0000 (13:19 +0100)]
A* path and neighbor cache performance statistics
as_clear_cachepath() now prints cache hits, misses, number of entries,
and memory use to stderr, when compiled with AS_STATS defined.
Markus Armbruster [Sat, 19 Feb 2011 09:47:32 +0000 (10:47 +0100)]
More precise and complete A* performance statistics
Memory usage didn't include path (adp->path), neighbor cache
(adp->neighbor_coords, adp->neighbor_nodes), and the hash table
(adp->hashtab).
While there, print path length.
To get A* statistics on stderr, compile with AS_STATS defined.
Markus Armbruster [Fri, 18 Feb 2011 19:08:26 +0000 (20:08 +0100)]
Clean up A* sector cache leftovers
The sector cache was disabled in v4.2.2, and dropped in commit
8f40f5ad, v4.2.20. A bit of cache statistics code was left behind.
Remove it.
Markus Armbruster [Sun, 20 Mar 2011 09:06:33 +0000 (10:06 +0100)]
Document buffer overrun for long land paths
BestLandPath(), BestDistPath() and best_path() are unsafe by design:
they take a path[] argument without a size, and blindly assume there's
enough space. When that's wrong, bp_path() overruns the caller's
buffer.
move_ground() and assemble_dist_paths() provide space for 512
characters. best(), dist(), path(), att_reacting_units(), s_commod()
and do_unit_move() provide space for 1024 characters.
A malicious player can arrange paths longer than that, but it takes a
lot of land.
BestAirPath() and BestShipPath() also take a path[] argument without a
size, but they're actually safe: bestownedpath() writes at most 100
(MAXROUTE) characters, perform_mission_bomb() provides space for 512,
sorde(), getpath(), do_unit_move() and nav_ship() for 1024.
Markus Armbruster [Fri, 25 Feb 2011 18:45:46 +0000 (19:45 +0100)]
Remove pointless check for sea from finish_sects()
Checking "sea or unowned" is pointless, because sea is always unowned.
Markus Armbruster [Fri, 25 Feb 2011 18:40:08 +0000 (19:40 +0100)]
Optimize assemble_dist_paths() for foreign distribution center
You can't distribute to a foreign sector. This case is relatively
rare. However, unsuccessful path search is relatively expensive, and
the extra check doesn't really slow down the common case.
Markus Armbruster [Fri, 25 Feb 2011 07:08:05 +0000 (08:08 +0100)]
Optimize dodistribute() for sectors with no distribution center
import_cost is now -1 in that case, so checking that suffices.
Markus Armbruster [Fri, 25 Feb 2011 07:00:59 +0000 (08:00 +0100)]
Fix assemble_dist_paths()'s recovery from invalid dist center
The recovery avoided crashing here, but left the path costs undefined.
If they happend to be non-negative, dodistribute() still crashed. Set
the costs to -1 to avoid that.
While there, oops on invalid distribution center.
Markus Armbruster [Sat, 19 Feb 2011 07:21:33 +0000 (08:21 +0100)]
Speed up export cost calculation in assemble_dist_paths()
Import and export paths enter the same sectors, except for the last
one. Compute export cost from import cost instead of reverting the
import path. Do it in dodistribute(), so that we need to store only
import costs.
Markus Armbruster [Fri, 18 Feb 2011 19:05:32 +0000 (20:05 +0100)]
Supply charged mobility for backward path
It used the path from supply recipient back to supply source. Has
always been broken that way.
Markus Armbruster [Fri, 18 Feb 2011 18:10:28 +0000 (19:10 +0100)]
Document memory leak in as_search()
Markus Armbruster [Fri, 18 Feb 2011 18:29:29 +0000 (19:29 +0100)]
Make bestpath work for deities in foreign land
Before, it only worked in land owned by the deity.
As always, paths can't cross international borders.
Markus Armbruster [Fri, 18 Feb 2011 18:07:31 +0000 (19:07 +0100)]
Merge dodistribute() parameters dist_i_cost, dist_e_cost
Only one of them is used, depending on argument imex. Replace them by
a single parameter path_cost.
Markus Armbruster [Fri, 18 Feb 2011 18:01:25 +0000 (19:01 +0100)]
Remove dodistribute() parameter path
It was only used to see whether a path to the dist center exists. Use
negative cost for that.
Markus Armbruster [Fri, 18 Feb 2011 17:53:37 +0000 (18:53 +0100)]
SAVE_FINISH_PATHS hasn't been used since 4.2.2, remove it
Since 4.2.2, assemble_dist_paths() stores a dummy path instead of the
real path to the dist center. That's possible because distribution
doesn't actually use the path, only whether it exists.
The code to store and free the real path is still around, under #ifdef
SAVE_FINISH_PATHS. Remove it.
Markus Armbruster [Fri, 25 Feb 2011 05:13:23 +0000 (06:13 +0100)]
Print distribution costs when compiled with DISTRIBUTE_DEBUG
Markus Armbruster [Sun, 11 Jul 2010 09:30:02 +0000 (11:30 +0200)]
Collect path-related stuff in path.h
Markus Armbruster [Wed, 16 Mar 2011 17:25:18 +0000 (18:25 +0100)]
Log distribution path assembly's CPU use (user and system time)
Markus Armbruster [Wed, 16 Mar 2011 17:24:58 +0000 (18:24 +0100)]
Log update's CPU use (user and system time)
Ron Koenderink [Sat, 19 Mar 2011 22:05:35 +0000 (16:05 -0600)]
Fleshed out getrusage() stub for Windows
Markus Armbruster [Wed, 16 Mar 2011 06:16:21 +0000 (07:16 +0100)]
Provide a getrusage() stub for Windows
Markus Armbruster [Sun, 13 Mar 2011 13:23:13 +0000 (14:23 +0100)]
Make savecore mind available disk space
Markus Armbruster [Sun, 13 Mar 2011 13:22:30 +0000 (14:22 +0100)]
Make savecore check core file is accessible
Markus Armbruster [Mon, 28 Feb 2011 05:58:51 +0000 (06:58 +0100)]
Fix bitmap overruns when WORLD_X * WORLD_Y not a multiple of 16
World-sized bitmaps were allocated with size WORLD_SZ() / 8, which
expands to (WORLD_X * WORLD_Y / 2) / 8. The divisions truncate unless
WORLD_X * WORLD_Y is a multiple of 16. The bitmaps were one byte too
small then. Bitmap overruns happen when:
* A lookout looks at one of the last sectors of the sector file.
Besides commands look and llook, this affects navigate and march
sub-command 'l'.
* Command spy spies into one of the last sectors of the sector file.
* A map or nmap (but not a bmap) shows one of the last sectors of the
sector file, or a sector that can see one of the last sectors
(visual range is two sectors at 100% efficiency). Besides commands
lmap, map, nmap, pmap, smap, this affects move and transport
sub-command 'm'.
Diagnosed with valgrind.
Already broken in BSD Empire 1.1 (bitmaps were on the stack then).
Markus Armbruster [Fri, 18 Feb 2011 18:36:55 +0000 (19:36 +0100)]
Fix pathrange()'s computation of the range's right limit
Because of the bug, the path command's maps weren't always fitted to
the path correctly. Broken in commit
0f458d2c, v4.3.17
Markus Armbruster [Sun, 6 Feb 2011 15:28:12 +0000 (16:28 +0100)]
Land units no longer hit allied mines
Markus Armbruster [Sun, 6 Feb 2011 09:29:14 +0000 (10:29 +0100)]
Use relations_with() in sendmessage()
We know player != other. Because we can have only one player in state
PS_PLAYING per country, and we know other->state == PS_PLAYING, it
follows that player->cnum != other->cnum. Thus, no functional change.
Adds another call to getnatp() hidden in relations_with(), though.
Keeping that optimized isn't worth it.
Markus Armbruster [Sun, 6 Feb 2011 09:27:45 +0000 (10:27 +0100)]
Use relations_with() for getrel(NP, THEM) where NP isn't THEM
Replacing getrel(NP, THEM), where NP is known to be getnatp(US), by
relations_with(US, THEM) makes a difference only when US equals THEM.
Replace in places where it's obvious that they're not equal.
Adds a few calls to getnatp() hidden in relations_with(). Keeping
that optimized isn't worth it.
Markus Armbruster [Sun, 6 Feb 2011 09:06:38 +0000 (10:06 +0100)]
Use relations_with() for US==THEM || getrel(NP, THEM)
Replace patterns like "US == THEM || getrel(NP, THEM)...", where NP is
known to be getnatp(US), by "relations_with(US, THEM)...". No
functional change.
Adds a few calls to getnatp() hidden in relations_with(), though.
Keeping that optimized isn't worth it.
Markus Armbruster [Tue, 1 Feb 2011 06:49:32 +0000 (07:49 +0100)]
Use relations_with() in unit_interdict()
No functional change, because the value of rel only matters when cn !=
victim, and then it's the same as before.
The new value of rel permits simplifying cn != victim && rel <=
NEUTRAL to just rel <= NEUTRAL
Markus Armbruster [Sun, 30 Jan 2011 17:32:19 +0000 (18:32 +0100)]
Use relations_with() in owned_and_navigable()
This removes a special case for POGO (#0). Before, unoccupied sectors
were treated as "own or allied" for POGO, but not for other deities.
Impact on callers:
* BestAirPath() is not affected, because the change is only reachable
with a non-null bigmap argument.
* sorde() and nav_ship() pass a non-zero ship owner. sorde() ensures
that itself, and prod_ship() does it for nav_ship().
* unit_path() passes the player number when called with a ship
argument, i.e. in the navigate command. Player number is zero for
POGO. Since deities can't navigate foreign ships, this can happen
only when POGO navigates dead ships. Yes, that's possible, needs
fixing.
* getpath() passes the player number (zero for POGO) when called with
argument P_SAILING, i.e. by the sail command.
Thus, the change makes navigate's and sail's path finding work for
POGO exactly like it does for other deities. That's fine.
Markus Armbruster [Sun, 30 Jan 2011 16:37:36 +0000 (17:37 +0100)]
Use relations_with() in getilists()
No functional change, even though this changes rel[intruder] from
NEUTRAL to ALLIED. Uses of rel[]:
* getilists() and ac_encounter() compare rel[cn] to HOSTILE. No
change, because NEUTRAL and ALLIED are both greater than HOSTILE.
* ac_encounter() compares rel[cn] to ALLIED, but only when cn !=
plane_owner. Because it passes plane_owner as argument for
getilists() parameter intruder, rel[cn] can't refer to the changed
element of rel[] here.
The new value of rel[plane_owner] permits simplifying cn ==
plane_owner || rel[cn] == ALLIED to just rel[cn] == ALLIED.
Markus Armbruster [Sun, 30 Jan 2011 16:10:43 +0000 (17:10 +0100)]
Use relations_with() in nav_loadship()
No functional change, because the value of rel only matters when
sectp->sct_own != sp->shp_own, and then it's the same as before.
The new value of rel permits simplifying sectp->sct_own == sp->shp_own
|| rel >= FRIENDLY to just rel >= FRIENDLY.
Markus Armbruster [Sun, 30 Jan 2011 15:39:38 +0000 (16:39 +0100)]
Use relations_with() in shp_fort_interdiction()
No functional change, because the change affects only
notified[victim], which isn't used in the loop around
notify_coastguard(), and gets overwritten before the interdiction fire
loop.
Markus Armbruster [Sun, 30 Jan 2011 15:31:02 +0000 (16:31 +0100)]
Use relations_with() in lnd_mar_one_sector()
No functional change, because the value of rel only matters when
sect.sct_own != actor, and then it's the same as before.
The new value of rel permits simplifying sect.sct_own != actor && rel
!= ALLIED to just rel != ALLIED.
Markus Armbruster [Sun, 30 Jan 2011 15:26:22 +0000 (16:26 +0100)]
Use relations_with() in player_relstr()
No functional change, because argument is never player->cnum.
Markus Armbruster [Sat, 29 Jan 2011 13:34:40 +0000 (14:34 +0100)]
Use relations_with() for getrel(getnatp(US), THEM) where US!=THEM
Replacing getrel(getnatp(US), THEM) by relations_with(US, THEM) makes
a difference only when US equals THEM. Replace in places where it's
obvious that they're not equal.
Note: getsect() sets player->owner to "player is god or owns this
sector". Thus, after getsect(..., §), sect.sct_own ==
player->cnum implies player->owner. Conversely, !player->owner
implies sect.sct_own != player->cnum. Similarly for getship(),
getplane() and nxtitem().
Markus Armbruster [Sun, 30 Jan 2011 10:46:17 +0000 (11:46 +0100)]
Use relations_with() where its different value doesn't matter
Switching from getrel() to relations_with() can change the value from
NEUTRAL to ALLIED. The change doesn't matter when the value's only
compared to HOSTILE, as both old and new value are greater than
HOSTILE. Likewise for >= NEUTRAL.
Markus Armbruster [Wed, 26 Jan 2011 18:41:28 +0000 (19:41 +0100)]
Use relations_with() for US==THEM || getrel(getnatp(US), THEM)
Replacing getrel(getnatp(US), THEM) by relations_with(US, THEM) makes
a difference only when US equals THEM.
Replace patterns like "us == them || getrel(getnatp(us), them)..." by
"relations_with(us, them)...".
Markus Armbruster [Sat, 29 Jan 2011 08:57:21 +0000 (09:57 +0100)]
Rearrange uses of getrel() slightly
Just to make the next few commits easier to review.
Markus Armbruster [Sat, 29 Jan 2011 12:09:26 +0000 (13:09 +0100)]
Eliminate a few pointless relations variables
Just to make the next few commits easier to review.
Markus Armbruster [Tue, 25 Jan 2011 19:13:41 +0000 (20:13 +0100)]
New relations_with()
Relations checking with getrel() often needs a special case for "is
same country". If you forget, you get behavior appropriate for a
neutral foreign country, which is usually very wrong (see commit
16c68eb4 for an example).
Unlike getrel(), relations_with() considers countries allied to
themselves. Less dangerous. In fact, allied behavior is typically
just right, so the special case isn't even needed.
Markus Armbruster [Sun, 30 Jan 2011 17:37:02 +0000 (18:37 +0100)]
Permit sharebmap with yourself
Does nothing. Before, it failed with a mildly bogus "does not have
friendly relations towards you" message.
Markus Armbruster [Wed, 19 Jan 2011 20:05:44 +0000 (21:05 +0100)]
Make share_bmap() do nothing for sharing with oneself
Before, it overwrote '?', '.', ' ' in the bmap with the capitalized
country letter, but only for sectors the player owns. Pretty
harmless, just weird. It can't happen currently, because sharebmap
with self fails with "does not have friendly relations towards you".
Markus Armbruster [Sat, 5 Feb 2011 11:50:50 +0000 (12:50 +0100)]
Permit flash to yourself
Flashing yourself failed with a bogus "not logged on" message for
deities, and a mildly bogus "not a deity or friendly with us" message
for mortals.
Fix by simply permitting it. Not terribly useful, except perhaps for
empire-hub users, but why not.
Markus Armbruster [Sat, 5 Feb 2011 11:48:10 +0000 (12:48 +0100)]
Drop redundant chat(), sendmessage() parameter us
It can't be anything but the current player.
Markus Armbruster [Sat, 5 Feb 2011 10:30:32 +0000 (11:30 +0100)]
Clean up confusing control flow in sendmessage()
The second patch hunk fixes a latent bug. Before, rejected deity
flashes led to a bogus "not logged on" message, now they lead to a
"not accepting" message. But deity flashes can't be rejected, so this
doesn't matter.
Markus Armbruster [Sat, 5 Feb 2011 10:17:40 +0000 (11:17 +0100)]
Optimize sendmessage() use of getnatp()
Don't call it when its value must be TO anyway.
While there, oops when it fails instead of recovering silently.
Markus Armbruster [Sat, 5 Feb 2011 10:09:37 +0000 (11:09 +0100)]
Fix racy detection of ally rejecting flashes
sendmessage() checked NF_FLASH on two places: when deciding whether to
send the message, and later when telling the player why it didn't send
a flash. This can race with the toggle command as follows: if a flash
could not be sent because the recipient's NF_FLASH was off, and the
recipient toggled it on before the flag was checked again, the flash
command claimed the sender wasn't logged on.
Markus Armbruster [Sat, 5 Feb 2011 09:42:58 +0000 (10:42 +0100)]
Change flash not to print "Flash sent" for deities
It's annoying. It never did that for mortals.
Markus Armbruster [Sat, 5 Feb 2011 09:18:30 +0000 (10:18 +0100)]
Remove pointless variables from setrel()
Markus Armbruster [Sat, 5 Feb 2011 08:31:44 +0000 (09:31 +0100)]
Make setrel() refuse to change relations to self
No current caller actually attempts that, but let's make it obvious.
Markus Armbruster [Sat, 5 Feb 2011 08:30:12 +0000 (09:30 +0100)]
Change setrel(), setcont(), setrej() to return void
Nobody cares for their value anyway.
Markus Armbruster [Sat, 5 Feb 2011 08:28:05 +0000 (09:28 +0100)]
Oops on invalid arguments in setrel(), setcont(), setrej()
Before, they recovered silently.
Markus Armbruster [Mon, 31 Jan 2011 19:53:44 +0000 (20:53 +0100)]
Use feels_like_helping() in dosupport(), lnd_support()
feels_like_helping() case cn == foe is missing in the code it
replaces. No difference in behavior, because:
* cn == foe && cn == friend can't happen. Because you can't get into
ground combat against yourself (assault, attack and paradrop don't
let you), friend != foe for support.
* cn == foe && cn != friend behaves the same: no support.
feels_like_helping() returns 0 because of the explicit case. The
replaced code doesn't support because cn can't be at war with
itself.
Markus Armbruster [Mon, 31 Jan 2011 19:46:53 +0000 (20:46 +0100)]
Factor feels_like_helping() out of quiet_bigdef(), sd(), dd()
Markus Armbruster [Mon, 31 Jan 2011 06:56:44 +0000 (07:56 +0100)]
Plug memory leaks in mission execution code
Mission execution first builds lists of eligible units, one list per
country. These lists are passed to perform_mission() one by one,
where they get freed.
Bugs:
* unit_interdict() didn't pass the list for the submarine's owner, but
build_mission_list_type() built one. Any submarine movement within
own submarine interdiction mission op areas leaked memory.
* dosupport() passed only lists for countries that actually support
(ally at war with victim), but build_mission_list_type() built lists
for all countries hostile to the victim. Ground combat within
support mission op areas countries that are hostile to one of the
party without actually supporting the other leaked memory.
* perform_mission() failed to free missiles targeting units.
Fixing the latter is straightforward.
Fix the first two by deciding whether a country acts on a mission
trigger before building any lists, in ground_interdict(),
unit_interdict(), dosupport(). Remove the code dealing with that from
build_mission_list_type() and the loops around perform_mission().
Markus Armbruster [Mon, 31 Jan 2011 19:02:23 +0000 (20:02 +0100)]
Move code from def_support(), off_support() to dosupport()
Markus Armbruster [Sun, 30 Jan 2011 16:08:00 +0000 (17:08 +0100)]
Eliminate nav_loadship() variables landown, shipown
Code is clearer without them. Works because load_it() never changes
ship or sector owner.
Markus Armbruster [Sun, 30 Jan 2011 15:55:15 +0000 (16:55 +0100)]
Drop redundant nav_loadship() parameter cnum
Markus Armbruster [Sun, 30 Jan 2011 15:54:19 +0000 (16:54 +0100)]
Eliminate nav_ship() variable cnum
Code is clearer without it.
Markus Armbruster [Sun, 30 Jan 2011 15:12:14 +0000 (16:12 +0100)]
Oops on invalid actor and victim arguments in nreport()
Replaces the existing, silent recovery from invalid victim argument.
Markus Armbruster [Sat, 29 Jan 2011 09:40:59 +0000 (10:40 +0100)]
Clean up use of current player in march code
lnd_mar() and lnd_mar_one_sector() take an actor argument.
Nevertheless, they sometimes used player->cnum. Fortunately, they are
the same: all callers pass current player for actor. Normalize to
actor for consistency.
Markus Armbruster [Sun, 23 Jan 2011 18:39:48 +0000 (19:39 +0100)]
Fix land unit attack mobility cost out of allied sectors
Land units pay a mobility penalty when marching into a non-old-owned
sector without sector mobility, to slow them down in newly taken
sectors. Attacking land units pay this penalty regardless of sector
mobility.
When attacking out of an allied sector, the penalty was computed as if
the land unit was owned by that ally. Attacking sectors old-owned by
that ally was too cheap, and taking back one's own was too expensive.
Broken since attacking land units pay the "newly taken" mobility
penalty: commit
2e693275, v4.3.6.
Markus Armbruster [Thu, 20 Jan 2011 06:49:44 +0000 (07:49 +0100)]
Fix attack when attacking sector gets taken by ally
When an attacking sector got lost while the player was at a prompt,
and the new owner was allied to the player, the server got confused:
1. If the sector attacked with mil, the server let the ghost mil
attack, but not occupy.
2. If the sector was allied, the server reported the sector loss and
land units dropping out of the attack, but claimed the lost sector was
yours.
Fix 1. by dropping sectors from attack when they change owner away
from the player, regardless of relations. Side effect: also drops any
surviving land units there. Before, they dropped out only if the new
owner wasn't allied to the player. That change's okay.
Fix 2. the obvious way: change the messages.
Broken in 4.0.0.
Markus Armbruster [Sun, 16 Jan 2011 20:20:28 +0000 (21:20 +0100)]
Fix trade ships to pay off in own harbor
Broken in commit
3318e4e4, v4.3.17.
Markus Armbruster [Sun, 9 Jan 2011 18:03:38 +0000 (19:03 +0100)]
Remove option SLOW_WAR
SLOW_WAR has issues:
* The check whether the attacker old-owns the attacked sector is
broken, because att_abort() uses sect.sct_oldown uninitialized.
Spotted by the Clang Static Analyzer.
* Its implementation in setrel() is somewhat scary. It's actually
okay, because that part of setrel() only runs within decl(). Other
callers don't reach it: update_main() because player->god != 0
there, and the rest because they never pass a rel < HOSTILE.
* Documentation is a bit vague.
SLOW_WAR hasn't been used in a public game in years. Fixing it is not
worth it, so remove it instead.
Markus Armbruster [Sat, 15 Jan 2011 13:39:30 +0000 (14:39 +0100)]
Fix nightly's sandbox cleanup not to hang when stdin is a tty
"rm -r" prompts for read-only files when stdin is a tty. Probably
broken since we install builtin configuration read-only, in commit
b4161cd7, v4.3.0.
Move "done" message to the right place. Broken in commit
70c03561,
v4.3.12.
Markus Armbruster [Sat, 15 Jan 2011 12:47:16 +0000 (13:47 +0100)]
Fix nightly's check for existing sandbox
Broken in commit
70c03561, v4.3.12.
Markus Armbruster [Sat, 15 Jan 2011 12:46:43 +0000 (13:46 +0100)]
Clean up a useless use of cat
Markus Armbruster [Sun, 15 Aug 2010 09:18:31 +0000 (11:18 +0200)]
Fix make clean to remove info.ps
Missed in commit
9067d7a4.
Markus Armbruster [Sun, 18 Jul 2010 16:16:01 +0000 (18:16 +0200)]
Fix sail command to support full path length
Off-by-one in cmd_sail_ship() chopped off the last character of
full-length sail paths.
Markus Armbruster [Sun, 18 Jul 2010 13:14:41 +0000 (15:14 +0200)]
Remove unused plurize()
Unused since commit
44c36fa, v4.3.23.
Markus Armbruster [Sun, 29 Aug 2010 09:38:02 +0000 (11:38 +0200)]
Change ioq_dequeue() to return void
For symmetry with ioq_append().
Markus Armbruster [Sun, 29 Aug 2010 09:31:02 +0000 (11:31 +0200)]
Err, the race in io_output() doesn't double-free
The previous commit's message claims the race can lead to duplicated
output, use after free, then double-free. That's correct only up to
the use after free. There is no double-free.
Heap corruption (double-free?) has been observed in Changeling,
though. Player logged in (still in sanctuary), map #, crashed within
removecc()'s free(io->data). Partial backtrace:
raise () from /lib64/libc.so.6
abort () from /lib64/libc.so.6
__libc_message () from /lib64/libc.so.6
malloc_printerr () from /lib64/libc.so.6
removecc (ioq=0x251fd10, cc=468) at ../src/lib/gen/ioqueue.c:350
ioq_dequeue (ioq=0x251fd10, cc=468) at ../src/lib/gen/ioqueue.c:135
io_output (iop=0x251fc90, wait=1) at ../src/lib/empthread/io.c:231
recvclient (cmd=0x258d8e0 "", size=1024) at ../src/lib/player/recvclient.c:82
getcommand (combufp=0x2557068 "map #1") at ../src/lib/player/empdis.c:84
I haven't been able to reproduce.
To hopefully catch ioqueue going south earlier, make ioq_dequeue()
oops when it can't dequeue as many bytes as requested.
Markus Armbruster [Sun, 4 Jul 2010 15:30:33 +0000 (17:30 +0200)]
Fix race in io_output() that can lead to double-free
Move call of ioq_makeiov() to its use, because calling it before
empth_select() is racy, as follows.
Player thread flushes output by calling io_output(player->iop, 1).
io_output() sets up iov[] to point to queued output. empth_select()
blocks on output.
Another thread sends a C_FLASH or C_INFORM message to this player.
This calls io_output(p->iop, 0). The output file descriptor has
become writable since the player thread blocked on it, so some output
gets written and dequeued.
The player thread resumes, writes out iov[] and dequeues. Any output
already written by the other thread gets duplicated. If the other
thread's dequeue operation freed struct io buffers, there's use after
free followed by double-free.
Markus Armbruster [Sun, 4 Jul 2010 15:30:03 +0000 (17:30 +0200)]
Document xundump() memory leak more clearly
Markus Armbruster [Sun, 27 Jun 2010 07:33:24 +0000 (09:33 +0200)]
Planes get to sweep and sonar only after flak and interception
If defenders get to shoot before bombs are dropped, they surely get to
shoot before time-consuming missions like sweep and sonar.
Sweep and sonar used to happen after air defense, but before flak and
interception. Air defense existed from Chainsaw 3 to v4.3.19.
Markus Armbruster [Sun, 18 Jul 2010 08:37:43 +0000 (10:37 +0200)]
Neater plane sonar contact output
plane_sona() prints an empty line to make sonar contacts stand out.
Move it so the contact is visually "attached" to the right sector,
like this:
flying over sea at 15,-3
Sonar contact in 15,-3
sub #3 13,-3
flying over sea at 13,-3
Before:
flying over sea at 15,-3
Sonar contact in 15,-3
sb submarine (#3) 13,-3
flying over sea at 13,-3
Markus Armbruster [Sun, 27 Jun 2010 09:35:08 +0000 (11:35 +0200)]
Don't write garbage to unused trade destination in trade file
struct trdstr members trd_x, trd_y are used only for teleporting
trades. For others, trad() wrote garbage coordinates to the trade
file. They weren't used except by xdump. Fortunately, even there
they're visible only to deities.
Write invalid coordinates instead. Do that in set() as well, so that
coordinates are valid only when we have a teleport destination.
Spotted by the Clang Static Analyzer.
Markus Armbruster [Sun, 27 Jun 2010 08:09:30 +0000 (10:09 +0200)]
Clean up dead store in skyw()
Dead since commit
38befcb4. Spotted by the Clang Static Analyzer.
Markus Armbruster [Sun, 20 Jun 2010 16:34:58 +0000 (18:34 +0200)]
Make disabled recursive supply in s_commod() compile again
Untested. Broken in commit
98f24d5c, v4.3.20.
Markus Armbruster [Sun, 20 Jun 2010 16:33:42 +0000 (18:33 +0200)]
Remove unused variables in disabled lnd_fort_interdiction()
Markus Armbruster [Sun, 20 Jun 2010 15:58:30 +0000 (17:58 +0200)]
Clean up unobvious coordinate system use in doland()
Use the obviously correct newown instead of sect->sct_own. They're
actually equal here.
Markus Armbruster [Sun, 20 Jun 2010 15:57:55 +0000 (17:57 +0200)]
Clean up unobvious coordinate system use in detonate()
Use the obviously correct player->cnum instead of own. They're
actually equal here.
Markus Armbruster [Sun, 20 Jun 2010 15:57:17 +0000 (17:57 +0200)]
Clean up unobvious coordinate system use in pln_airbase_ok()
Use the obviously correct player->cnum instead of pp->pln_own.
They're actually equal here.
Markus Armbruster [Sun, 20 Jun 2010 15:53:17 +0000 (17:53 +0200)]
Clean up suspicious coordinate system use in unit_put()
It showed unit coordinates in unit's coordinate system instead of the
actor's. Fortunately, they're the same, since it is reachable only
for non-zero actor, only shp_nav_one_sector(), lnd_mar_one_sector()
and sail_nav_fleet() pass that, and even deities can't navigate
foreign ships or march foreign land units.
Markus Armbruster [Sun, 20 Jun 2010 14:34:08 +0000 (16:34 +0200)]
Factor rad_char() out of radmap() and rad_map_set()
Code was duplicated in commit
0d477e5d.
Markus Armbruster [Sun, 20 Jun 2010 14:19:22 +0000 (16:19 +0200)]
Compute radar range in one place, rad_range()
Before, a part was duplicated in radmap() and rad_map_set(), and
another part in their callers.
Markus Armbruster [Sun, 20 Jun 2010 12:55:30 +0000 (14:55 +0200)]
Don't limit radar command's range to fit into world map
Limited since Chainsaw 2 so that the radar map fits into a world map
without clipping, i.e. its diameter neither exceeds WORLD_X / 2 nor
WORLD_Y. Maybe range exceeding that triggered bugs then. It doesn't
now, and it makes no sense.
The limit never applied to automatic bmap update from ship radar.
Markus Armbruster [Sun, 20 Jun 2010 11:55:52 +0000 (13:55 +0200)]
Simplify radmap() and radmap2()
radmap() is now radmap2()'s only caller. Inline radmap2() and
simplify. This cleans up a suspicious-looking use of xyas(): it
relied on the fact that owner == player->cnum if pr_flag.
Markus Armbruster [Sun, 20 Jun 2010 11:43:16 +0000 (13:43 +0200)]
Simplify automatic bmap update from ship radar
Inline radmap2() into radmapupd() and simplify. Drop unused parameter
seesub. Rename to rad_map_set().
Markus Armbruster [Sun, 20 Jun 2010 11:32:03 +0000 (13:32 +0200)]
Remove radmapnopr(), use radmapupd() instead
Markus Armbruster [Sun, 20 Jun 2010 08:47:37 +0000 (10:47 +0200)]
Clean up output destinations in navigation code
shp_nav() and shp_nav_one_sector() printed both to their actor
argument and to ship owner. shp_nav_one_sector()'s use of xyas()
looked particularly suspicious: it passed actor, then printed the
result to the ship owner. Fortunately, actor and ship owner are the
same, since even deities can't navigate foreign ships. Normalize to
actor for consistency.
While there, rename shp_mess() to shp_stays().
Markus Armbruster [Sun, 20 Jun 2010 08:42:34 +0000 (10:42 +0200)]
Clean up output destinations in march code
lnd_mar(), lnd_sweep() and lnd_mar_one_sector() printed to the current
player, their actor argument, and to land unit owner.
lnd_mar_one_sector()'s use of xyas() looked particularly suspicious:
it passed actor, then printed the result to the current player or land
unit owner. Fortunately, all three are the same: all callers pass
current player for actor, and land unit owner is the same, since even
deities can't march foreign land units. Normalize to actor for
consistency.
While there, rename lnd_mess() to lnd_stays().
Markus Armbruster [Sun, 20 Jun 2010 08:40:00 +0000 (10:40 +0200)]
Clean up confusing use of def->own in move_in_land()
It passed def->own to lnd_sweep(), which looks like a bug. But it's
actually player->cnum there, because take_def() already set def->own
to player->owner: take_def() first changes the owner of the attacked
sector by calling takeover(), then updates def->own from that in
att_get_combat().