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_remap_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.
desi() ran the designate code twice, first for adding up the cost,
then for changing designation. However, the checking pass already
changed the sector when that cost nothing. The checking pass also
suppressed messages. There was at least one message that never got
printed because it was suppressed in the checking pass, and the
condition for it was no longer true in the changing pass, due to the
premature sector change: when a deity changed a non-coastal sector to
harbor or bridge head.
The total cost check is of limited value: designate costing money is a
bad idea, and the stock game has no such sectors. Not enough value to
justify keeping and fixing this disgusting mess. Remove it instead.
Always charge land units at least as much mobility for assaulting from
non-landing ships as for landing ships. Before, marines lost all
mobility when assaulting from a non-landing ship, which could be less
than what the same assault costs from a landing ship (half an update's
worth).
Actually, this isn't just simplification. When mobility gain per
update was configured to be greater than 128, mobility could go from 1
to less than -127 when assaulting from a landing ship, and thus
overflow. Make it saturate at -127. Note that you can expect plenty
of trouble elsewhere with such a silly configuration.
With -s, LWP initializes thread stacks to track stack use. It did
that after makecontext(), with disastrous results on systems where
makecontext() writes something to the stack that swapcontext() expects
to find there. Makes FreeBSD 6.2 crash in swapcontext().
Factor stack allocation out of lwpNewContext() into lwpNewStack().
Move call of lwpStackCheckInit() into lwpNewStack().
A bridge (span or tower) must be splashed when it gets damaged below
SCT_MINEFF. Likewise when its last supporting sector (bridge head or
tower) gets damaged below SCT_MINEFF, unless EASY_BRIDGES is enabled.
We need to check this whenever a bridge head, span or tower gets
damaged. This is done in three places, and all of them screw up:
* checksect() ignores damage to bridge heads. It also leaves writing
back the sector it checks to the caller, which never happens when
it's called from sct_postread().
Note that checksect() drowns all planes on bridges it splashes.
Functions that need to exempt flying planes from such a fate have to
splash bridges themselves.
* sect_damage() ignores damage to bridge towers, and damage to bridge
spans unless EASY_BRIDGES is enabled. It then runs checksect(),
which compensates for these omissions, but happily drowns the planes
sect_damage() attempts to protect.
* eff_bomb() ignores damage to bridge heads. Collateral damage makes
sect_damage() run, which compensates for the omission.
This causes the following bugs:
* Efficiency damage going through sect_damage() can drown planes it
shouldn't. This affects pinpoint bombing when collateral damage
splashes a bridge, and strategic bombing. The drowned planes then
crash and burn when they attempt to land at their (just splashed)
base.
* Efficiency damage to bridge heads not going through sect_damage()
fails to collapse unsupported bridges. This affects pin-bombing
efficiency without collateral damage, and ground combat. Also deity
commands edit, setsector and add, but that could be regarded as a
feature.
* If the sector file somehow ends up with an inefficient bridge span,
it collapses on every read again and again, until it collapses on a
write. Related problems exist with other actions of checksect(),
and they're not addressed here.
* If the sector file somehow ends up with adjacent inefficient bridge
towers, checksect() on any of them recurses infinitely:
- checksect() inefficient tower T1
- knockdown() T1, but don't write that back to the sector file
- bridgefall() T1; this reads all adjacent sectors, including
inefficient towert T2
- checksect() T2
- knockdown() T2, but don't write that back to the sector file
- bridgefall() T1; this reads adjacent sectors including T1
- checksect() T1
...
This commit creates a new function bridge_damaged() to splash any
bridges that became inefficient or unsupported after damage to a
sector. To avoid the inifinite recursion, we call it in
sct_prewrite() instead of checksect().
No uses knockdown() outside bridgefall.c remain, so give it internal
linkage.
This feature was added in commit 70c03561bb but was
not added to the list of supported commands at the top
of file. No functional change just a documentation
correction.
Need to be in the empserver directory before applying
the "git apply" command, move the "cd empserver" to the
clone step. This change also fixes a bug when using
with an existing repository not being the correct
directory when executing the "git pull" command.
Fixes commit acd768e5e2 as it used git diff format.
The three previous commits touching path.c.in broke mingw.patch: the
relative vs absolute paths changes moved the initializers to be
patched to new variables, and the telfil removal messed up the
context.
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.
Mil are not required for building units since 4.0.0. l_mil was still
initialized to l_item[I_MILIT], and used instead of that in a couple
of places. Fix those, and remove the initialization.
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.
Commit 530deef2 failed to update .TH of fairland(6), files(6) and
pconfig(6).
Commit 530deef2 failed to update .TH of empire(6).
Commit eeb9d3cb created empsched(6) with the wrong .TH.
Until now, they tried to recover and continue (debug off). That's
appropriate only for the server. The server could be told to abort
instead (debug on, selected by option -d), but not the utility
programs.
Change debug to be on by default, and switch it off early in the
server's main(). No functional change for the server.
Create a local git repository on the SuSE nightly build machine.
Update local git repository from the Markus's public repository
at the beginnning of the nightlybuilds for SuSE and WIN32.
For the SuSE and WIN32 nightlybuilds drawn from the local
git repository on SuSE machine.
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.
The function that gave its name to this file is long gone, the file's
description is bogus, and it contains just one definition. Move that
to ../subs/border.c, and delete the file.
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.
Detect .git instead of CVS. Use git-ls-files with git, else
$(srcdir)/sources.mk. info/findsubj.pl now gets the info files as
arguments rather than from sources.mk.
Remaking info subjects doesn't quite work, but it was broken similarly
before.
