If input is cleared before, it can become impossible to cancel the video
when decoding and displaying a frame takes longer than a frame.
git-svn-id: https://svn.eduke32.com/eduke32@2833 1a8010ca-5511-0410-912e-c29ae57300e0
The library was built with GCC 4.6.1 and configured like
./configure --disable-vp8-encoder --disable-multithread --disable-spatial-resampling
On the aforementioned AMD system, this reduces the mean time for decoding a frame
from 18.2 to 15.4 ms, so IMO it's worth the somewhat hefty addition of 430 Kb.
git-svn-id: https://svn.eduke32.com/eduke32@2831 1a8010ca-5511-0410-912e-c29ae57300e0
On an AMD Phenom II X4 system with generic memory modules, this brings down
the mean time for this conversion from 16.5 to 10.5 ms.
(GCC 4.6.1, optimized build)
git-svn-id: https://svn.eduke32.com/eduke32@2830 1a8010ca-5511-0410-912e-c29ae57300e0
This changes a constant inside the panning calculation from 256 to 255 (making
e.g. panning 255 and 0 the same in the "San Andreas fault" sign in E1L5) and
uses the correct reference wall for the "do panning correction?" conditional.
Now, the problematic walls should look the same as in Polymost.
git-svn-id: https://svn.eduke32.com/eduke32@2822 1a8010ca-5511-0410-912e-c29ae57300e0
This makes the corrections conditional on where they appear in (under-, over-,
white or mask wall).
git-svn-id: https://svn.eduke32.com/eduke32@2821 1a8010ca-5511-0410-912e-c29ae57300e0
I'm willing to make this one concession to correctness in the name of backward-
compatibility. I think that one reason why this has never cause a crash is that
tiles in BUILD are allocated in Ken's big allocache buffer, so oob accesses were
dampened by that (though they hit uninitialized data).
git-svn-id: https://svn.eduke32.com/eduke32@2819 1a8010ca-5511-0410-912e-c29ae57300e0
The event is run after drawing the scene, but before the overlays. To
make a screenshot from a script, set DOSCRSHOT to non-zero. It will then
be scheduled to run once after the drawing but before the above-mentioned
event. The screenshots will be called mcapXXXX.{png,tga}.
git-svn-id: https://svn.eduke32.com/eduke32@2818 1a8010ca-5511-0410-912e-c29ae57300e0
This would only hit when polymost_printext256 erred out (mem alloc failure,
glGenTextures failure), i.e. "almost never".
Also, tweak a bound check in polymost_printext256.
git-svn-id: https://svn.eduke32.com/eduke32@2817 1a8010ca-5511-0410-912e-c29ae57300e0
The editor colors are the ones listed on page 9 of the Mapster32 built-in help.
git-svn-id: https://svn.eduke32.com/eduke32@2816 1a8010ca-5511-0410-912e-c29ae57300e0
The map iterator now has init/finish capability, making it possible to
write scripts that aggregate data over multiple map files. One such example
calculates some statistics, the other loads art metadata and looks for
red walls with non-pow2 ysize tiles.
git-svn-id: https://svn.eduke32.com/eduke32@2814 1a8010ca-5511-0410-912e-c29ae57300e0
Non-pow2 wall/mask drawing in classic introduced cases where walls are
drawn "incorrectly" because they were constructed with the old behavior
in mind. Polymer appears to "correct" for it partially, but doesn't cover
all cases. Specifically, now we have:
- E1L1 first inside secret room (5000, 50000): Polymer draws like Duke 1.5,
classic now draws with an offset.
- E3L2 near the vault (-20000, 25000): both classic and Polymer draw with
offsets compared to Duke 1.5, but they're different!
This means that more research is needed into what makes these two cases
diverge, even though both have the same root cause.
!!! Also, mappers should abstain from using non-power-of two textures on
walls until this issue is resolved in a satisfactory fashion !!!
git-svn-id: https://svn.eduke32.com/eduke32@2811 1a8010ca-5511-0410-912e-c29ae57300e0
This was introduced with r2771, which fixed e.g. AMC TC city_si's mirrors,
but instructed the base drawrooms inside yax_drawrooms to not correct the
passed sectnum. Therefore, stuff would get drawn wrongly when passing
sector boundaries, like from the platform to the rails in trueror1.map.
git-svn-id: https://svn.eduke32.com/eduke32@2810 1a8010ca-5511-0410-912e-c29ae57300e0
The latter is commented out, since it has to be *compiled* with
script_expertmode enabled.
git-svn-id: https://svn.eduke32.com/eduke32@2808 1a8010ca-5511-0410-912e-c29ae57300e0
Related to that, it looks like out-of-bounds accesses when drawing such walls/
maskwalls or *sprites* are fixed, too. Sprites still show a stray lines on some
occasions, but Valgrind doesn't complain then.
git-svn-id: https://svn.eduke32.com/eduke32@2805 1a8010ca-5511-0410-912e-c29ae57300e0
Most notably, -Wdeclaration-after-statement. This and -Wpointer-arith
give some warning on linux, but this is "harmless" as it's assumed that
we'll be always compiling with GCC or Clang there.
Also, erratum in the "Make ksqrt take uint32_t ..." commit:
hypotenuse -> squared length of the hypotenuse
git-svn-id: https://svn.eduke32.com/eduke32@2797 1a8010ca-5511-0410-912e-c29ae57300e0
-Wwrite-strings is useful to detect code where string literals and e.g. alloc'd
strings are used side-by-side, potentially creating dangerous situations, or to
find uses of old, non-constified APIs. However, enabling it would still flood
the log with too many warnings. Also, GCC wrongly warns for initializations of
char arrays.
git-svn-id: https://svn.eduke32.com/eduke32@2796 1a8010ca-5511-0410-912e-c29ae57300e0
Most differences are handled with function parameters, except that one instance
checked SpriteProjectile[i].spawns for being >0 instead of >=0. The factored
function always checks for >=0.
git-svn-id: https://svn.eduke32.com/eduke32@2794 1a8010ca-5511-0410-912e-c29ae57300e0
uhypsq calculates the hypotenuse using unsigned multiplication. This is
permissible since for arbitrary int32s a and b, the following holds in
two's complement arithmetic:
(int32_t)((uint32_t)a * b) == (int32_t)((int64_t)a * b)
("Signed and unsigned multiplication is the same on the bit level.")
This fixes various overflows where wall lengths for walls of length > 46340
are calculated, but does not rid us of other overflows in the same vein
(usually dot products between vectors where one point is a wall vertex and
the other a position in a sector).
git-svn-id: https://svn.eduke32.com/eduke32@2791 1a8010ca-5511-0410-912e-c29ae57300e0
The latter shows that "int32_t ksqrt(int32_t)" also copes with values in the
range INT32_MIN..-1, effectively interpreting them as uint32_t (i.e. adding
2**32). However, this should not be relied on from CON.
git-svn-id: https://svn.eduke32.com/eduke32@2790 1a8010ca-5511-0410-912e-c29ae57300e0
The bug was introduced with SAMESIZE_ACTOR_T enabling in r2208.
This fixes being unable to read the messages in A.Dream* by zykov eddy.
git-svn-id: https://svn.eduke32.com/eduke32@2788 1a8010ca-5511-0410-912e-c29ae57300e0
Arrays inside structs must not be accessible, since they're not bound-checked
by the FFI. Therefore, we flatten them into repeated scalar fields and need
to write accessor functions later.
git-svn-id: https://svn.eduke32.com/eduke32@2786 1a8010ca-5511-0410-912e-c29ae57300e0
This fixes an integer overflow when a distance is calculated later.
git-svn-id: https://svn.eduke32.com/eduke32@2785 1a8010ca-5511-0410-912e-c29ae57300e0
That is, "clang -ftrapv" builds don't abort almost immediately after entering
a level.
There are various classes of overflow bugs, needing different handling:
- Some texture mapping code was written with signed integer wrapping semantics
in mind. In some places, we're able to get away with unsigned casts.
- sometimes, we really need a wider range, like when calculating distances or
dot products
- negating INT_MIN. Here, we cast to int64_t temporarily. Note that if the
result is 32-bit wide, no 64-bit code may actually need to be generated.
- shifting into a signed integer's sign bit. We cast to uint32 here.
- in hitscan(), at the "abyss crash prevention code" comment, it's clearly
the other code that is better...
This is not merely done for pedantry, but rather makes it easier to track down
overflow bugs having a real impact on the game.
git-svn-id: https://svn.eduke32.com/eduke32@2784 1a8010ca-5511-0410-912e-c29ae57300e0
- hitscan & related types and constants
- profiling with gethitickms
- translator: eval the opening parts of block commands early
- fix getbunch
git-svn-id: https://svn.eduke32.com/eduke32@2779 1a8010ca-5511-0410-912e-c29ae57300e0
For SDL 1.2 and when building on linux, the code using
clock_gettime(CLOCK_MONOTINIC, ...) is taken over from SDL HG.
gethitickms() is a convenience function that return milliseconds as
doubles and isn't exposed in any header file, yet.
git-svn-id: https://svn.eduke32.com/eduke32@2778 1a8010ca-5511-0410-912e-c29ae57300e0
Doesn't work: indexed-color modes, gamma (at least for X11), mouse wheel,
special keys like ENTER or BACKSPACE in the OSD, probably more...
In build/Makefile.shared, we now have logic to autodetect an SDL2 installed
in /usr/local, however OS X and Wii builds follow other Makefile code paths,
it seems. Note that the matching SDL2_mixer must be used then, too.
In source/jaudiolib/src/driver_sdl.c, change the #includes from <SDL/SDL_xxx.h>
to "SDL_xxx.h". SDL wiki says this is the most portable way, hopefully this
doesn't break builds for anyone.
git-svn-id: https://svn.eduke32.com/eduke32@2777 1a8010ca-5511-0410-912e-c29ae57300e0
This was actually broken all the time except when ALL sectors were selected
(which was what I tested incidentally when I wrote the first "fix"). D'oh!
git-svn-id: https://svn.eduke32.com/eduke32@2773 1a8010ca-5511-0410-912e-c29ae57300e0
This is to the cvar of the same name as samples/aspect.map is to r_usenewaspect.
git-svn-id: https://svn.eduke32.com/eduke32@2769 1a8010ca-5511-0410-912e-c29ae57300e0
This exposes some problems in the default/in-the-wild CONs. As usual,
we'll have to retrofit sensible semantics :rolleyes:.
git-svn-id: https://svn.eduke32.com/eduke32@2765 1a8010ca-5511-0410-912e-c29ae57300e0
The latter is only for development, since the embedded version already has a
undeclared-var-reference handling similar to that. Also fix parm2memberpat.
git-svn-id: https://svn.eduke32.com/eduke32@2762 1a8010ca-5511-0410-912e-c29ae57300e0
It's time to replace some int32s with 64-bit ints in core engine functions.
The problem is that for example, the dot product is taken between vectors that
may be the difference between two arbitrary points in a sector, so even if one
sticks to the "no blue walls" rule, that doesn't guarantee freedom from
overflows.
git-svn-id: https://svn.eduke32.com/eduke32@2761 1a8010ca-5511-0410-912e-c29ae57300e0
This was exposed by Clang with -ftrapv, which apparently also traps those.
git-svn-id: https://svn.eduke32.com/eduke32@2757 1a8010ca-5511-0410-912e-c29ae57300e0
This is to not slow down the core drawing functions too much in debugging builds
and mimics the way things are on x86.
git-svn-id: https://svn.eduke32.com/eduke32@2755 1a8010ca-5511-0410-912e-c29ae57300e0
Bang! One more invalid internals exposure squashed. Maybe one day we'll be able
to call EDuke32 "stable"...
Actually the offending sprite IS drawn as face sprite with shade 32 and xrepeat
and yrepeat 255 for the convenience of the CON coder who will have to debug it.
git-svn-id: https://svn.eduke32.com/eduke32@2744 1a8010ca-5511-0410-912e-c29ae57300e0
Literals written out as if they were unsigned and having the highest-order
bit set are all too ubiquitous, like hitscan masks for example.
git-svn-id: https://svn.eduke32.com/eduke32@2742 1a8010ca-5511-0410-912e-c29ae57300e0
Those were happening when wall drawing in an upper/lower layer "closed" too
much of the "curtain" (umost/dmost) for subsequent drawing in that portal.
Example: looking down the sewer manhole at the beginning of Retaliation will
now render without HOM.
git-svn-id: https://svn.eduke32.com/eduke32@2739 1a8010ca-5511-0410-912e-c29ae57300e0
Also,
- don't strip const when passing the char ptr to it and a couple more
instances in game.c
- use g_netPort when -connect parameter doesn't have a port suffix (":XXXX"),
so that e.g.
eduke32 -port 1700 -connect localhost
is the same as
eduke32 -connect localhost:1700
(-port must come before -connect, unfortunately.)
git-svn-id: https://svn.eduke32.com/eduke32@2737 1a8010ca-5511-0410-912e-c29ae57300e0
This was supposed to go into the last commit, but I did
git commit --amend
without adding the changes to the index. Now I'm too lazy to sort
things out.
git-svn-id: https://svn.eduke32.com/eduke32@2736 1a8010ca-5511-0410-912e-c29ae57300e0
The GAMESTATE one has to be carried out twice, first on the data from
xd3_encode_memory, and then with the compressed data (because it might have
increased in size, though this is very unlikely). The MOVE check is
similar, though there, failure of the first check implies corrupt memory
(which is why we're Bassert'ing that condition).
Currently, the overflow on GAMESTATE sending happens when switching to a
different map.
git-svn-id: https://svn.eduke32.com/eduke32@2735 1a8010ca-5511-0410-912e-c29ae57300e0
These are element indices for a while now. Also eliminate some dead
code afterwards.
git-svn-id: https://svn.eduke32.com/eduke32@2731 1a8010ca-5511-0410-912e-c29ae57300e0
r2727 made old savegames incompatible, as an array with MAXVOLUMES*MAXLEVELS
is saved in Gv_WriteSave().
git-svn-id: https://svn.eduke32.com/eduke32@2730 1a8010ca-5511-0410-912e-c29ae57300e0
New Wii control defaults for the Wii Remote + Nunchuk and the Classic Controller. This includes new code added just so that the Home key brings up the menu in-game, reducing the need for a USB keyboard.
On the technical side, raw joystick access (comparable to what is available for keyboard and mouse) is now present in jmact, on the game side. (added: joystick.[ch])
Using this new raw joystick access, I replaced tueidj's hack to map A and B to LMB/RMB and D-Pad Up/Down to the scrollwheel.
I made the menus more friendly to mouse and joystick browsing by adding and unifying checks and clears for various buttons and gamefuncs. In fact, the majority of the time spent on this commit was tracking down problems that appeared with the factoring and trying to understand the menu system and the way input checks are precariously executed.
In addition, "Press any key or button to continue" now truly means what it says.
As a result of incorporating proper raw access into control.c instead of it directly accessing the implementaiton, the program *may* no longer be affected by joystick input when it is out of focus. This follows the pattern set by the mouse, and I think this is a positive change.
A small bonus: In the classic/old keyboard preset, the key for Show_Console has been changed from '`' to 'C' because '`' is taken by Quick_Kick.
git-svn-id: https://svn.eduke32.com/eduke32@2728 1a8010ca-5511-0410-912e-c29ae57300e0
NAM and Napalm can now share their con, def, and rts files if the one for their respective game is not present because the con and rts files are identical.
Also, decapitalize two string literals missed in r2540.
git-svn-id: https://svn.eduke32.com/eduke32@2726 1a8010ca-5511-0410-912e-c29ae57300e0
(LUNATIC build only.)
Also, a minor problem is identified. sizeof(actor_t) is 124 on 64-bit platforms,
while the expected size is 128 bytes. This needs to be corrected whenever the
next savegame version bump happens.
git-svn-id: https://svn.eduke32.com/eduke32@2724 1a8010ca-5511-0410-912e-c29ae57300e0
This introduces a Makefile variable EDUKE32_MY_DEVELOPER_ID, which is
expected to come from the environment, so that minor disagreements/
preferences can be handled -- don't let it go overboard though! -- to keep
the working directory clean of stuff that one doesn't actually work on.
My ID is 805120924, i.e. "helix".
git-svn-id: https://svn.eduke32.com/eduke32@2718 1a8010ca-5511-0410-912e-c29ae57300e0
Usage is from 3D mode only. When a sector (or wall) is committed to the
clipboard, it (or the wall's sector) is saved, and is subsequently used
when auto-aligning [;-ENTER] another sector's ceiling or floor against the
reference one's. This temp. sector is reset on any structurally modifying
operation except setting first walls; also, aligning extended ceilings or
floors is impossible as they use .*xrepeat internally.
The auto-alignment does not change picnums (this can be accomplished with
the ['+ENTER] combination), but copies the orientation bits 2^{2..6} to the
alignee. Afterwards, if the reference is relative-aligned, it tweaks them,
so that every case where the two firstwalls are either parallel or perpendi-
cular is handled correctly.
It does not yet calculate the panning values.
git-svn-id: https://svn.eduke32.com/eduke32@2715 1a8010ca-5511-0410-912e-c29ae57300e0
This is so that people know where to enable it. One day, all these
debugging switches ought to go into the Makefiles in one form or
another.
git-svn-id: https://svn.eduke32.com/eduke32@2714 1a8010ca-5511-0410-912e-c29ae57300e0
Also factor out the code (2x) setting the alignment bits for sprites.
git-svn-id: https://svn.eduke32.com/eduke32@2712 1a8010ca-5511-0410-912e-c29ae57300e0
It works perfectly fine without it for me, and including it for no reason
is just an annoyance to other people building from source.
git-svn-id: https://svn.eduke32.com/eduke32@2708 1a8010ca-5511-0410-912e-c29ae57300e0
Defines SPRITES_OF(Statnum, Iter) and SPRITES_OF_SECT(Sectnum, Iter)
[The first is so that STAT_* can be substituted for Iter and it reads nicely.]
Usage should be self-explanatory, but one thing to keep in mind is that
the sprites that are iterated over MUST NOT be deleted.
git-svn-id: https://svn.eduke32.com/eduke32@2707 1a8010ca-5511-0410-912e-c29ae57300e0
This fixes the glitches/corruption whereever showview is used from
EVENT_DISPLAYROOMS while the scene is actually rendered to a tile instead of the
frame buffer, for example because we have a tilted view or "pixel doubling"
enabled. Fixing it for real (i.e. so that the showview actually completes)
will require more effort.
git-svn-id: https://svn.eduke32.com/eduke32@2706 1a8010ca-5511-0410-912e-c29ae57300e0
vbit was being accessed with an int32 e.g. at address 3200 when only 3203 were
allocated. We simply alloc +1 byte.
git-svn-id: https://svn.eduke32.com/eduke32@2705 1a8010ca-5511-0410-912e-c29ae57300e0
This is done by kopen4loadfrommod'ing and immediately closing the file
(since the checker is still run from there when not ingame).
On my system, times were like 0.12 ms, 0.2 ms and 0.5 ms for each
openfrompath/findfrompath, openfrompath/Bopen and SHGetFileInfo
respectively, so it shouldn't have a too noticable impact on initial
loading times.
git-svn-id: https://svn.eduke32.com/eduke32@2704 1a8010ca-5511-0410-912e-c29ae57300e0
I experienced the following on Windows XP: a few openfrompath() -->
findfrompath() calls were taking enormously long (4.5 secs) to complete, having
been passed a file name like "//bla/qwe.asd". My guess is that Windows then
tried to interpret these as a network FS path in access(), and the 4.5 secs
represents a timeout value.
git-svn-id: https://svn.eduke32.com/eduke32@2703 1a8010ca-5511-0410-912e-c29ae57300e0
The first means that the expensive (~0.5 ms) SHGetFileInfo() calls won't inter-
fere with smooth gameplay, but files that are opened only at game-time like
sounds won't be checked. The second means that there are now less false
positives, i.e. warnings about files that would be found due to the
check-all-{upper,lower} hack.
git-svn-id: https://svn.eduke32.com/eduke32@2702 1a8010ca-5511-0410-912e-c29ae57300e0
This fixes the crash on Wii when looking sidewards (e.g. when dying), but
makes the tilted view look more pixelated.
The problem was that an attempt was made to render onto a 640x640 tile for
the rotated view, but the base engine arrays were too small for that.
git-svn-id: https://svn.eduke32.com/eduke32@2698 1a8010ca-5511-0410-912e-c29ae57300e0
Also, correct the loadboard() return value checks in premap.c to be
aware of a returned -2 ("wrong map version").
git-svn-id: https://svn.eduke32.com/eduke32@2694 1a8010ca-5511-0410-912e-c29ae57300e0
Most of the time, Windows file name lookup is case-insensitive. Reading the
docs (see MSDN's CreateFile help, for example), it seems like case-sensitivity
can be controlled on a per-file basis where applicable (NTFS), but people
should be concerned about matching case in the DEFs/on disk *especially* in
that case.
Also, note that this hack will not always help on systems with case-sensitive
lookup.
git-svn-id: https://svn.eduke32.com/eduke32@2693 1a8010ca-5511-0410-912e-c29ae57300e0
When a file from the local file system is opened, its real file name is gotten
with SHGetFileInfo() and compared against the one that was passed. In the case
they're not identical, a warning is issued.
This is one step towards eliminating mismatched file names in DEFs etc., which
cause trouble on systems that look them up case-sensitively. However, it's not
perfect because the issue is trickier than it appears on first sight.
For one thing, this will only check the last (i.e. file) part in the path,
falsely accepting mismatched directory names. However for these, it reports
them ruthlessly, even for those names where the try-other-case hack (try all
uppercase, all lowercase) would find the correctly-cased file.
git-svn-id: https://svn.eduke32.com/eduke32@2692 1a8010ca-5511-0410-912e-c29ae57300e0
