buildeff() rounds work and money up. Until recently, fractions could
only occur on tear-down, but with customized costs they can now also
occur on build-up.
The previous commit changed unit building to round money and work
randomly. Before, money was rounded down, and work was rounded up.
Round them randomly for sectors as well, for consistency.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
shiprepair() limits the efficiency gain to how much the workers can
build, rounding randomly. It charges work, money and materials for
the efficiency actually gained, rounding work up, money down, and
materials randomly. Same for planerepair() and landrepair(). Has
always been that way.
If you get lucky with the random rounding, you may get a bit of extra
work done for free.
The budget command runs the update code, and can be off by one due to
different random rounding.
Sector production used to have the same issue, only more serious,
because a single unit of tech production matters much more for the
budget than a single point of unit efficiency gain. I fixed it in
commit 6f7c93c, v4.3.31.
Fix it for unit building the same way: limit efficiency gain to the
amount the workers can produce (no rounding). Work becomes a hard
limit, not subject to random fluctuations. Randomly round work and
money charged for actual gain, like we do for materials. On average,
this charges exactly the work and money that's used.
This lets budget predict how much gets built a bit more accurately.
It's still not exact, as the amount of work available for building
remains slightly random, and the build cost is randomly rounded.
The old rounding of work for ships carries the comment "I didn't use
roundavg here, because I want to penalize the player with a large
number of ships." Likewise for planes. Rounding work up rather than
randomly increases the work cost by 0.5 per ship, plane or land unit
on average. I could keep the penalty by adding 0.5 before random
rounding. Not worth it, since the effect is actually pretty trivial.
Let's examine a fairly extreme case: an airfield with 600 available
work repairing a huge number of lightly damaged planes, say f2 with
81% average efficiency. The old code lets the airfield repair roughly
600 / 6.5 = ~92 planes, the new code 600 / 6 = 100.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Update code shared with budget uses the bp map instead of the sector,
so that budget can track materials and work available in sectors for
ship, plane and land unit building without updating the sector file.
Unfortunately, the bp map can become stale during the update.
prepare_sects() doesn't update the bp map for sea sectors, unlike
budget's calc_all(). Instead, we rely on calloc()'s initialization.
Works, but is a bit unclean.
prepare_sects() updates the bp map after fallout, but neglects to
update it for any of the later sector updates (steps 1b to 1f in info
Update-sequence). Che can destroy materials and available work, and
the plague can kill military. The bp map stays out of date until
produce_sect() updates it again.
Since we deal with sector production and countries in increasing order
of country number, foreign ships, planes and land units owned by
countries with lesser numbers get built before their sector produces.
Building uses the stale bp map then, and can use materials and
available work destroyed by che or the plague. The update test
demonstrates the former case.
For stopped sectors or when the owner is broke, produce_sect() updates
only materials in the bp map, not available work. Nothing builds in a
stopped sector, but allies may build in your sectors even when you're
broke. They can use available work destroyed by che then.
Screwed up when Empire 3 made the update code work for budget.
Note that budget bypasses the flawed code: it prepares its bp map
itself instead of calling prepare_sects().
Rather than fixing prepare_sects(), use a null bp map for the update:
writes become no-ops, and reads read from the underlying sector. Not
only does this remove the possibility of the bp map going stale during
the update, it saves a bit of memory, too.
calloc()'s initialization is now dead. Switch to malloc().
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
produce_sect() skips sectors without civilians, military and land
units. These are unowned. Any uw there are frozen in time: they
don't eat, procreate or produce. Has always been broken. Don't skip
such sectors.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
prepare_sects() caches the sector owner's getnatp() across
guerrilla(), do_plague() and populace(). This is wrong, because the
owner may change. The mistake can be traced back all the way back to
BSD Empire 1.1.
If the sector revolts or reverts to deity, the ex-owner still receives
taxes and bank interest. The update test demonstrates this bug.
If the sector revolts, we use the ex-owner's instead of the owner's
tech and research for plague, and we use the ex-owners happiness and
required happiness instead of the owner's for loyalty update and civil
unrest.
Change do_plague() and populace() to call getnatp() themselves. Call
it in prepare_sects() only after we're done messing with the sector
owner.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
We maintain a few sector invariants in sct_prewrite(). Since the
update bypasses sct_prewrite(), it needs to maintain them itself. The
two should be consistent.
The update reverts deserted sectors to deity in three places:
do_plague(), populace() and produce_sect(). None of them is
consistent with sct_prewrite().
populace() can revert unowned sectors to deity. This creates bogus
entries in the "lost" file. Harmless; messed up when the lost items
were added in 4.0.7. Visible in tests/smoke/final.xdump.
populace() fails to revert when there are only uw left. If PLAGUE is
enabled, do_plague() already reverted. Else, produce_sect() will.
This is the only case where they add value to populace(). Can be
traced back all the way to BSD Empire 1.1.
All three neglect to clear mobility. Harmless.
Fix populace()'s condition for reverting to deity, and make it clear
mobility. Drop the reverting from do_plague() and produce_sect().
populace() also resets state that applies to civilians when there are
none: work percentage, loyalty and old owner. However, it resets on
different conditions than sct_prewrite(). Messed up in Chainsaw;
before, populace() didn't reset at all.
For sectors without military, populace() fails to reset. This can
happen when the update wipes out civilians and military, say by plague
or fallout. The now bogus work percentage, loyalty and old owner
persist until sct_prewrite() runs on the next non-update sector
update. Except old owner is reset correctly by populace() when the
sector reverts to deity. It doesn't when the owner has a land unit
there.
Most of the time, this doesn't matter, as moving civilians into a
sector without civilians overwrites the sector's work percentage,
loyalty and old owner. However, airlifting and unloading civilians
fail when the old owner differs from the owner. Else they adopt the
sector's loyalty and work percentage (bug#49 and bug#255).
Fix populace() to reset any sector without civilians, like
sct_prewrite().
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
We maintain a few sector invariants in sct_prewrite(). Since the
update bypasses sct_prewrite(), it needs to maintain them itself. The
two should be consistent.
sct_prewrite() resets work percentage of owned sectors to 100% when
there are no civilians. The update's populace() resets it for unowned
sectors as well, if they have military.
Change sct_prewrite() to reset sct_work = 100 regardless of owner.
Also change sct_oninit() to initialize sct_work = 100, so it doesn't
change on first write. Update tests/smoke/fairland.xdump for the same
reason.
The massive test output differences are all due to sct_work.
Inconsistencies with the update remain. They will be fixed next.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
In Empire, even babies work.
neweff and production compute the projected population's work,
discarding fractions.
The update first computes the adults' work (discarding fractions),
then newborns' work (discarding fractions), then adds them together.
Double rounding. Moreover, it uses the old work percentage for the
adults' work, and the new one for the newborns' work. Broken in
Empire 3.
Fix by recomputing work after grow_people(). This is how things
worked before the regression. Also restores a bug: growfood()'s work
use is ignored. Harmless, because fcrate and fgrate are too low for
growfood() to produce anything, and nobody customizes them. Mark
FIXME anyway.
Update test output changes as expected: available work differs in
sectors where double rounding discards work, an in sectors with
changing work percentage.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Civilians, military and uw work only up to their sector's population
limit. The population limit depends on the sector type's maximum
population, research if RES_POP is enabled, and the sector's
efficiency for big cities.
The population limit may decrease between computation of work in
do_feed() and the end of the update:
* Research declines (only relevant with RES_POP). Work is not
corrected. The declined research will apply at the next update.
Since levels age after production is done, any work corrections
could only affect leftover available work. Wouldn't make sense.
The effect is negligible anyway. Even with an insanely fast decline
of 60% (level_age_rate = 1, etu_per_update = 60), the population
limit decreases by less than 10% in the worst case.
* upd_buildeff() changes sector type and efficiency. Work is
corrected only when this changes the sector type from big city to
not big city.
It isn't corrected on other sector type changes. These can affect
maximum population since the sector type's maximum became
configurable in commit 153527a (v4.2.20). Sane configurations don't
let players redesignate sectors to a type with different maximum
population. The server doesn't enforce this, though.
It isn't corrected when a big city's efficiency decreases, but
sector type change isn't achieved. Harmless, because tearing down a
city takes very little work (25 for 100%), so efficiency decrease
without type change means the work we have must be safely below any
sane population limit's work.
Good enough. However, the code implementing the work correction for
big cities is unclean. Get rid of it by tweaking the rules: a big
city's extra population does not work. City slickers, tsk, tsk, tsk.
At least they still pay their taxes.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Tests need repeatable pseudo-random numbers to yield repeatable
results. Commit 73f1ac8 (v4.3.33) reseeds the PRNG with the count of
commands right before executing a command when running_test_suite is
on. This doesn't help the update: whenever update code exercised by a
test is changed to consume fewer or more PRNs, all subsequent users
get different numbers regardless. The ensuing test result changes are
extremely tedious to review.
To address this problem, reseed the PRNG in the update's two most
important loops with the iteration count when running_test_suite.
This way, the effect of perturbing the PRN sequence lasts only until
the next iteration.
There are many more loops, but reseeding in all of them seems
impractical.
Perturbs test results across the board. Hopefully, that'll happen
less frequently now.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
Notable gaps in its coverage are fallout, most of guerrilla, delivery,
distribution, ALL_BLEED and LOSE_CONTACT.
Signed-off-by: Markus Armbruster <armbru@pond.sub.org>
