clang rejects 0x80000000 as a non-type template argument of type `int`
due to narrowing. Use `INT32_MIN` instead, which has the same numeric
value, and should be accepted without requiring narrowing.
Reported-by: Havner <https://github.com/dxx-rebirth/dxx-rebirth/issues/353>
Fixes: 88832e3679 ("Use constexpr integral_constant for various magic numbers")
Global variable `grd_curcanv` is set to a variety of canvases, some of
which are local stack variables. Use of global variables in this way is
fragile, but works as long as the global is not used beyond the life of
the backing local.
Unfortunately, some existing uses do access the canvas beyond the
lifetime of the backing local. Playing movies sets the font of the
current canvas. If the current canvas is an expired stack variable,
setting the font overwrites other stack data. This data corruption
causes various symptoms, such as inability to play the escape tunnel
movie.
Prior to 03cca2b3dc, the corruption on
playing the endlevel movie had no user-visible effect. That commit
created a large local variable, which changed stack layout. Starting
with that commit, the corruption causes the movie to play as all black.
Fix this, and protect against some other data corruption possiblities,
by clearing the global when the local goes out of scope.
Reported-by: Havner <https://github.com/dxx-rebirth/dxx-rebirth/issues/345> (only as cutscene failure to play, not as the underlying corruption issue)
Rather than use an inline wrapper and rely on the compiler optimizer to
redirect gr_set_current_canvas(nullptr) to gr_set_default_canvas,
rewrite all relevant calls directly in the source.
git grep -l 'gr_set_current_canvas' | xargs sed -i -e 's:gr_set_current_canvas(\s*NULL\|nullptr\s*);:gr_set_default_canvas();:'
Defining SLEW_ON breaks the build due to use of member variable names
not used since before the D2X import in 2001.
`unifdef -USLEW_ON -m -- similar/main/endlevel.cpp`
This code was part of a feature abandoned before retail. It cannot be
usefully used in campaigns. Remove it to reduce code size and simplify
later changes.
Descent 1 bosses could always teleport, but were only placed in large
areas where free teleporting was always permitted.
Descent 2 boss 1 was placed in a confined segment and not permitted to
teleport out of it until it was opened. This was implemented by a
two-part change relative to Descent 1 rules:
- Descent 1 bosses were always permitted to teleport to any teleport
destination segment. Descent 2 bosses were only permitted to teleport
if the player was visible or the boss had recently been hit.
- When computing the permitted list of teleport destination segments,
Descent 1 used directly connected accessible segments, then stopped.
Descent 2 started with this rule, but if the list had at most 1 entry,
then it would assume this is the confined boss and recompute the list
with the first wall ignored. This recomputed list allowed the boss to
teleport to any segment in the larger arena outside its starting
segment.
After D2X-Rebirth support for emulating Descent 1 bosses was enhanced in
28bd4c1650, Descent 1 bosses gained the ability to teleport out of a
confining cube early, but only in D2X-Rebirth when emulating Descent 1.
In D1X-Rebirth, when a boss is placed in a confining cube, it can always
teleport, but only to that confining cube. In D2X-Rebirth, Descent 1
bosses retain the always-teleport rule of Descent 1, but gained the
Descent 2 rule for expanding its search to the surrounding arena. It
should only use the expanded search when it also abides by the Descent 2
restriction not to teleport before the first wall is opened. It did not
abide by that rule. This commit adds that restriction for Descent 1
bosses.
Reported-by: ef314159 <https://github.com/dxx-rebirth/dxx-rebirth/issues/348>
Fixes: 28bd4c1650 ("Enable D1 boss behavior in d2x build. So we get correct boss behavior when emulating D1, and 3rd party mn2s can include D1 bosses.")
Commit b1b5de4 switched from enabling select physics flags to setting
those flags and clearing all others. Unfortunately, flag PF_USES_THRUST
was omitted from the enabled list, so it was disabled on reset. It is
required to let player ships move.
Reported-by: Sottises <https://github.com/dxx-rebirth/dxx-rebirth/issues/347>
Fixes: b1b5de4297 ("Additional safeguard for bug #306")
Editor groups write `struct segment` to a file in raw form, but had no
code to enforce that this raw form remained stable over time. Various
changes to `struct segment` have repeatedly changed its internal
structure. Each change created an incompatible dialect of the editor
group file, and all the dialects share the same version number.
According to
```
git log -p -L'/struct segment {/,/};/:common/main/segment.h' HEAD --not 0.58.1-d1x 0.58.1-d2x --
```
`struct segment` changed layout in:
* d1c6b89f17 ("Move dsx::segment -> dcx::segment") [D1 only]
* 596ecbb38d ("Rename segment::value to segment::station_idx") [D1 only]
* 6f10a67c09 ("Move segment::sides to end")
* c53b734abb ("Compute slide segments early") [D2 only]
* 40e90fea22 ("Move Light_subtracted[] to Segments[].light_subtracted")
* a65d774c83 ("Improve packing of struct segment")
* c70c6c98b3 ("Remove obsolete segment::degenerated flag")
Mark editor groups as broken to avoid making the mess worse. If anyone
cares about group support, it needs to be rewritten not to depend on the
internal layout of `struct segment`.
Descent 2, but not Descent 1, had a useless test in render_side:
type == QUAD ? 0 :
type == TRI_13 ? 1 :
0
This is useless since, if type is not QUAD, the second expression will
apply. If type is QUAD, then the type is not TRI_13, so the second
expression would choose the same result as the first. The extra
comparison does not save any work, so it is useless. Remove it.
Commit 1335af4b51 restructured briefing number parsing, but
unintentionally changed the parser not to drop the newline after a
number. The briefing parsing code is very sensitive to minor changes,
so this broke parsing some briefing screens, including Vertigo briefing
screens after the first. After the change, clicking through the Vertigo
briefing aborts on failure to load a PCX file.
Fix this by consuming the newline before returning. This matches the
previous semantics.
Reported-by: Havner <https://github.com/dxx-rebirth/dxx-rebirth/issues/343>
Fixes: 1335af4b51 ("Simplify get_message_num")
Like many things in the main game loop, lavafall handling was
historically done on a per-frame basis, then its effects were scaled to
FrameTime to normalize the results. Ignoring rounding errors, this
produced roughly equivalent damage for high framerate users (who
experienced many but small damage hits) and low framerate users (who
experienced few but large damage hits). However, the randomized
movement was not scaled to FrameTime, which caused differing results for
high framerate users versus low framerate users. Commit b36c6f20c7
tried to fix this by forcing scrape_player_on_wall to run at a capped
maximum effective frame rate, then scaling the damage in
check_volatile_wall accordingly, so that high framerate users would
experience fewer damage hits, but the ones they received were larger,
thus maintaining approximately the same damage as before.
Prior to b36c6f20c7, damage was always scaled to FrameTime. In
b36c6f20c7 and later, damage scaled to max(FrameTime,
DESIGNATED_GAME_FRAMETIME), causing users with a high frame rate (and
thus low FrameTime) to take more damage on each pass. This damage
increase was balanced by an added hack in scrape_player_on_wall to limit
the frequency of the scrape so that high framerate users would skip some
scrapes, giving them a virtual frame rate appropriate to
DESIGNATED_GAME_FRAMETIME. However, the damage is only balanced if the
new governor is used consistently. It is not used consistently, so it
caused a regression for passable lava surfaces. Scraping on a solid
lava surface goes through scrape_player_on_wall and respects the
governor. Touching a passable lava surface (only available in Descent
2) bypasses scrape_player_on_wall and jumps directly to
check_volatile_wall, thereby bypassing the governor. This allows high
framerate users touching a passable lava surface to take many hits (as
they always did), but not receive the full benefit of downward scaling
the damage (as they once did) due to the new max() expression. Thus,
they are damaged frequently, but still take damage consistent with being
damaged infrequently.
Fix this by moving the hack from scrape_player_on_wall to
check_volatile_wall, so that both solid lava surfaces and passable lava
surfaces respect the governor. The governor is still an ugly hack that
should not be global, but this is a spot fix for the immediate problem.
Fixes: b36c6f20c7 ("Made scrape_player_on_wall() based on a timer. Due to the player being pushed away from the lava/water surface in every frame in a random vector (wrong, too), player movement per frame was not enough to counter this on FPS rates > ~120 which made damage scaling per frame nonsensical in these situations. Instead, execute scrape results in intevals based on DESIGNATED_GAME_FRAMETIME (or per frame if FrameTime>DESIGNATED_GAME_FRAMETIME) which fixes the issues and generally works much better for the purpose of this function.")
Commit f4b21088a0 ("Track vulcan ammo explicitly") fixed an original
retail bug that prevented the thief from stealing energy weapons,
because the thief could only steal weapons for which the player had ammo
and energy weapons never have ammo. This went unremarked for several
years, until a recent report of the new semantics as a game-breaking
regression because the thief is now "ridiculously potent".
Address this report, as well as an intermittently raised issue from
various users over time, by adding two new knobs to both the single
player "Gameplay" menu and the multiplayer setup screen: "Remove Thief
at level start" and "Prevent Thief Stealing Energy Weapons".
"Remove Thief" deletes the thief object during level load. It has no
impact on save games, and changing it after entering a level has no
effect on any thief already in the level.
"Prevent Thief Stealing" is checked at the moment of theft and, when
enabled, prevents stealing primary weapons other than Vulcan/Gauss.
This can be changed at will in single player and is immediately
effective. In multiplayer, this option can only be changed by the game
host in the pre-game setup.
For both knobs, there is one pair of checkboxes to control this as a
player preference, which applies in single player games. There is a
second pair of checkboxes in the multiplayer setup, which applies only
to multiplayer games. Therefore, in multiplayer, the host chooses thief
settings and all clients use the host's choice. The host may configure
the thief differently in multiplayer from how the host plays in single
player.
For users who wanted to remove the thief, no specific tally has been
kept for who requested it or when. Now that the code is being updated,
this is thrown in as an easy addition.
Reported-by: MegaDescent <http://forum.dxx-rebirth.com/showthread.php?tid=980> (for the thief stealing energy weapons as a game-breaking regression)
When reporting use of SDL_mixer, report its loaded version. When
running in verbose mode, report both the compile-time and load-time
versions of PhysFS, libSDL, and SDL_mixer.
Some Linux libraries print their own messages to stdout/stderr,
particularly in case of severe errors. Decorate messages generated by
Rebirth to distinguish them from library generated messages.
For specific blacklisted renderers, add a message informing the user
that the blacklist matched and changed settings.
Only one caller exists, and that caller alway passes a non-nullptr
value. Switch to a reference and remove the unused special case to
handle a nullptr input.
Most uses pass an orientation matrix. All sites are deterministic about
whether a matrix is passed. Make the matrix mandatory for sites that
passed it, and split out a separate version of g3_start_instance_matrix
for the 2 sites which do not provide orientation.
Function bm_load does not exist in current tip, 0.58.1-d2x, 0.51-d2x,
nor various early snapshots of D2X (not D2X-Rebirth, just D2X).
Attempting to enable STATION_ENABLED fails without this function. Since
it seems to have been absent for at least 10 years (0.51-d2x tagged Feb
2007), and possibly more than 16 years (D2X snapshots from 2001), it
seems reasonable to assume no one has used this code in any open source
release. Remove it.
This backports into Descent 1 the special-case that loads a Vulcan
cannon when it is dropped. It also fixes a longstanding bug that a
robot that drops more than one of the powerup would fill only one of
them.
Many of these locals are wasteful, since they are always sized to the
biggest buffer required. This is the minimal and safe solution. Future
work will tune them to the correct size.
(a+i) form assumes that a can be used as the left hand side of
operator+(). This is true for pointers, and for C arrays that degrade
into pointers, but not for C++ std::array. Switch to the explicit
syntax.
git grep -lz 'multibuf+' | xargs -0 sed -i -e 's/(\s*multibuf\s*+\s*\(\w\+\),/(\&multibuf[\1],/g'
Little endian systems stored sound as a `uint32_t`, but sent only the
lowest byte. Store only that byte, since the destination was unaligned
and was the last field in the message.
Players keep reporting errors that trace to the game rejecting invalid
data from user-made maps. In this case, asymmetric connectivity in
_Descent Vignettes_[1] level _Stalactite Volcano_ (level 15 in Descent,
level 14 in Descent 2) caused an exception when trying to render part of
the first big room.
The debug build shipped with function `check_segment_connections` to
detect some types of connectivity errors. Expose this function to
release builds, and call it for user-made maps. It already checked for
asymmetric segment links. Extend that to hotfix those links by breaking
the asymmetric connection. Log an URGENT class message when this
happens, so that users are aware that the level was modified at load.
Move the `NDEBUG` guards into `check_segment_connections` so that
non-debug builds check only for segment symmetry, but do not check for
normals and face count errors.
[1] http://www.enspiar.com/dmdb/viewMission.php?id=724
```
sha1sum vignette.hog vignette.msn
3cd659e6dd5927b41157dfb7d1dd87d90e781f01 vignette.hog
1f7d140ffab11816364040dd6da71a1568393a16 vignette.msn
stat -c '%s %Y %n' vignette.hog vignette.msn
5717889 1251643308 vignette.hog
956 1250212930 vignette.msn
```
Reported-by: MegaDescent <http://forum.dxx-rebirth.com/showthread.php?tid=970>
Revert to the user's chosen resolution, not to 0x0, which would then be
coerced to the minimum resolution. When the text says revert, it should
revert, not just pick a size.
Change the "OK" label to say "Revert to NxM" for the resolution it is
reverting to.
Tighten input validation so that a blank width or blank height produce
the correct error message, rather than being coerced to 0, then
rejecting 0 as too small.
d2x::segment has all members of d1x::segment, as well as two new members
exclusive to d2x::segment. Structure layout is such that d1x::segment
requires the same size allocation, and places anonymous pad members in
the locations that become named members in d2x::segment. Thus, reusing
d2x::segment for d1x::segment does not change the size of the structure
nor the offsets of any members used. This reuse may enable some
functions to be better shared by the dsx project.
Instead of blacklisting two specific bad N_save_pof_names values, ignore
all N_save_pof_names that would overrun the Save_pof_names array.
This allows Rebirth to load broken levels such as "Hazard Zone"[1]. In
the case of Hazard Zone, the POF count is simply missing, so the first
two characters of the first POF name were taken as a count. This led to
a massive overrun of the Save_pof_names array.
[1]:
```
sha1sum hazard.msn hazard.rdl
25abe7ba1aca91f0bd09551e65cfadeabfcb73df hazard.msn
f5e6761b674b595550b27733fd83eeb2eff5e8f8 hazard.rdl
```
`value` is generic and unclear. It is always meant to be used as an
index into the Station array, so rename it `station_idx` to show this.
Define and consistently use `station_none` to represent that no station
is assigned.
Delete stub "compiler-type_traits.h" header. Redirect all uses to the
standard <type_traits> header.
git grep -wlz 'compiler-type_traits.h' -- '*.cpp' '*.h' | xargs -0 perl -p -i <<EOF
BEGIN {
$i = 0;
}
if (($i == 1 && $_ eq "\n") || ($i < 2 && /^#include "/)) {
# First blank line or first user-include after a system-include.
# Print, then never again for this file.
print "#include <type_traits>\n";
$i = 2;
} elsif ($i == 0) {
$i = 1 if (/^#include </);
} elsif ($_ eq "#include \"compiler-type_traits.h\"\n") {
# Remove this line if found.
$_ = '';
}
# Reset state machine when moving to next file.
$i = 0 if eof;
EOF
All supported compilers have an acceptable <type_traits>. Commit
4cb3d46148 ("Move <type_traits> test to Cxx11RequiredFeature") made
<type_traits> support mandatory in August and no one has objected.
Remove the indirection and use namespace std directly for type_traits
members.
Early unification efforts missed this one because it was in gamemine.c
for Descent 1, but segment.c for Descent 2. Move it to gamemine.cpp for
both, so that it can be static for both games.
gcc-7 warns:
similar/main/mglobal.cpp:119:5: error: this use of "defined" may not be portable [-Werror=expansion-to-defined]
#if DXX_VALPTRIDX_ENABLE_FULL_TEMPLATE_INSTANTIATION
Fix this by using the long form of the definition.
Previously, valptridx used PREFIX for allow-invalid+mutable, c#PREFIX
for allow-invalid+const, v#PREFIX for require-valid+mutable, vc#PREFIX
for require-valid+const. Convert the types, factories, and all usage
sites to specify a qualifier for all four combinations:
im#PREFIX -> allow-invalid+mutable
ic#PREFIX -> allow-invalid+const
vm#PREFIX -> require-valid+mutable
vc#PREFIX -> require-valid+const
Changes to common/include/valptridx.h and common/include/fwd-valptridx.h
are manual. All other changes are generated by:
git grep -lz -e '\(obj\|seg\|clwall\|wall\|actdoor\|trg\)\(ptridx\|ptr\|idx\)\(_t\)\?\>' | xargs -0 sed -i -e 's/\<\(v\?\)\(\(obj\|seg\|clwall\|wall\|actdoor\|trg\)\(ptridx\|ptr\|idx\)\(_t\)\?\)\>/\1m\2/g'
for the 'm' prefix and:
git grep -lz -e '\(obj\|seg\|clwall\|wall\|actdoor\|trg\)\(ptridx\|ptr\|idx\)\(_t\)\?\>' | xargs -0 sed -i -e 's/\<\([cm]\(obj\|seg\|clwall\|wall\|actdoor\|trg\)\(ptridx\|ptr\|idx\)\(_t\)\?\)\>/i&/g'
for the 'i' prefix.
Past releases had a debug-only assertion that 25 was sufficient, but
then dynamically allocated enough storage for larger models as needed.
Commit 22a34809ee ("Move interpreter g3s_lrgb onto stack") added a
hard limit of 25, but did not detect attempts to exceed this. Custom
models that exceeded 25 would cause a stack buffer overwrite and likely
crash. Raise the limit to 64 and add sanity checking to refuse to
render models that exceed MAX_POINTS_PER_POLY.
Fixes: 22a34809ee ("Move interpreter g3s_lrgb onto stack")
Disable gcc's -Wformat-nonliteral for this one call since unchecked
format strings are (1) under the user's control, (2) unable to cause
problems, and (3) very difficult to fully support with the warning
enabled.
Reported-by: parkerlreed <https://github.com/dxx-rebirth/dxx-rebirth/issues/338>
For switch cases where existing comments or code flow logic obviously
intended to fall through, add a gcc-7 /*-fallthrough*/ comment to
silence warnings about this. Some cases which are less obvious are not
converted, so the code does not yet compile clean with
-Wimplicit-fallthrough.
Reported-by: parkerlreed <https://github.com/dxx-rebirth/dxx-rebirth/issues/338>
Commit 829e95b dropped a test for game_mode_hoard when computing light.
Since hoard orbs are in a union which overlaps the score in mission
mode, this caused a bug where a player in mission mode with non-zero
score would pulse light like a player carrying orbs in hoard mode.
Restore the test for hoard mode.
Fixes: 829e95b6f8 ("Separate hoard/proximity tracking")
The Windows implementation of inet_ntop incorrectly omits the `const`
qualifier on the input address. This broke the Windows build, since
Rebirth passes a const-qualified input address, as permitted by POSIX.
Some Windows cross-compilers, such as mingw32, lack a definition of
inet_ntop entirely. For such environments, fall back to inet_ntoa and
disallow building with IPv6.
Rework the formatting and add a Windows-specific test for whether
inet_ntop is available. As inet_ntop is specified by POSIX, assume all
modern non-Windows platforms support it.
Commit b32298df5a ("Rewrite powerup cap
code to centralize logic") centralized powerup cap code in the
powerup_cap_state class.
Commit 901a554e96 ("New powerup management
code: Addeed functions and packet type to ...") removed all use of
powerup_cap_state, but left the dead implementation present.
Commit 479f5ed584 ("Fix 'format specifies
type 'unsigned short' but the argument has type 'unsigned char''
warning") switched the already dead (but still compiled) code from %hu
to %hhu to fix a warning on OS X. Although the commit was written by
Chris, it was my suggestion to use %hhu. I neglected to test Windows
before suggesting it, so the change went in even though Windows does not
accept %hhu; this broke the Windows build. Fortunately, the code had
been dead for 11 months when the change was made, so the fix for Windows
is to remove the long dead code.