| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Speeds up rendering a lot in case of many offscreen actors, for example
when a heavy particle effect is running somewhere.
|
|
Sometimes I've used CamelCase and sometimes SNAKE_CASE for these values.
Since "enum class" values are always prefixed with the enum name, which
uses CamelCase, I find it more fitting to use it for the values as well.
Also fixes compilation on Windows where 'ERROR' was conflicting with a
define.
|
|
All ResourceManager functions that load resources now return respective
ResourceRef values, which helps to make sure resources are properly
cleaned up.
The Sound class was cleaned up and now also allows SoundEffect resources
to be unloaded.
The Animation class now keeps its ImageSet loaded only as long as
necessary. Previously, SimpleAnimation and ParticleEmitter would keep
the ImageSet loaded indefinitely by never decreasing its reference
count.
Reduced duplicated animation loading code between SimpleAnimation and
ParticleEmitter.
|
|
This makes accessing the config values much faster, since it no longer
needs to do a lookup nor string conversion, which means we could remove
some needless copying of the values.
Overall it makes it easier to find out where settings are used and it
puts the defaults along with the declaration.
Options with default values are no longer saved to the config file. This
does not include unrecognized options, which are kept around to provide
some compatibility with older clients.
While most basic options have kept the same name, more complicated
settings like window geometry, shortcuts, outfits, etc. now have their
own XML elements. Older clients will ignore these and erase them when
saving the configuration.
|
|
* Removed `Actor::getNumberOfLayers`, since we really only cared about
whether they should be visible when behind other stuff.
* Removed `CompoundSprite::drawnWhenBehind`, since this property was
only relevant for actors.
|
|
The logic update now uses Time::deltaTimeMs() where needed to make it
framerate-independent. This means there will no longer be multiple logic
calls per frame (as was usually the case with logic ticking at 100 fps
whereas the game would generally run at 60 fps).
At the same time, the game can be more precise at higher framerates and
should now run smoother at 144 Hz, for example. Previously the game would
sometimes skip logic ticks at that rate.
This change affects:
* Updating of animations
* Being movement speed
* More moving of manual time variables to Timer
Notoriously, the particle system still does 100 ticks/second.
|
|
The Timer is efficient because it does not depend on incrementing a
counter to keep track of time, nor does it call SDL_GetTicks every time
its state is checked (this happens once per frame instead).
Along with global functions Time::absoluteTimeMs() and
Time::deltaTimeMs(), this replaces previous globals tick_time, cur_time
and get_elapsed_time().
For now, there is still a fixed 100 times per second logic call rate,
but the new Time::deltaTimeMs() function should allow getting rid of
this.
|
|
Automatic reference counting of images is now used by Item, Icon,
AnimatedSprite, ImageSprite, ParticleEmitter, Minimap, Desktop and
Emote.
Since ResourceManager::get automatically adds a reference, it needs to
be explicitly subtracted when the resource is managed by ResourceRef.
This is taken care of by the new ResourceManager::getImageRef.
Also removed the apprently unused and duplicate "mDrawImage" from Item
(which also didn't get decRef called on it).
Fixes cleanup of emote ImageSet and ImageSprite instances, as well as
particle images.
|
|
* Removed some unused includes
* Removed unused ListBox::mFont
* Removed wrong cast to SDL_Scancode
* Removed superfluous .c_str()
* Removed superfluous explicit std::string construction
* Removed unused variable
* Use more emplace_back
* Turned FindBeingFunctor into a lambda
* Avoid needless pointer references for ambient layers and use a vector
|
|
I had broken this in 264be2108c51837fa92085f6c839e66aebbcfc9e by no
longer reading out the property.
Made some further simplifications and two more fixes:
* When the original image's size was not 800x600, it wouldn't get scaled
as expected due to the integer division being performed before the
multiplication.
* I had changed the default resolution to 1280x720, while mKeepRatio
expects to scale the original image as if it was displayed on an
800x600 screen. So now it's hardcoded on 800x600 rather than using
the defaultScreenWidth/Height variables.
* Removed the confusing x/y parameters from AmbientLayer::draw, which
were actually set to the graphics width/height.
* Changed SDLGraphics::drawRescaledImagePattern to not repeatedly
set the source origin.
|
|
Rewrote the tile animation loading code based on XML tags, replacing
the code that loaded tile animations from tile properties.
Also made a number of code simplifications and optimizations:
* Replaced a number of pointer members with value members.
* Pass around Animation and TileAnimation by value, using std::move to
avoid allocating copies.
* push -> emplace
* push_front -> emplace_front
* push_back -> emplace_back
* Use range-based for loops
* Use std::vector instead of std::list for storing affected tiles
(less fragmentation)
* Avoid string copies and allocations while parsing CSV layer data.
* Replaced xmlNodeGetContent with directly accessing 'content'.
|
|
|
|
* Don't needlessly store or return raw pointers in BeingInfo
* Less copying, more moving
* Less else after return
* Make AddDEF a template instead of a macro
* Removed some unused includes
* Use range-based for loops
|
|
modernize-loop-convert
modernize-deprecated-headers
|
|
A custom "Mask" property on a layer or a "foregroundXmask" property on a
map can now be used in combination with the SMSG_MAP_MASK to dynamically
disable certain map layers from the server.
Feature previously seen on ManaPlus and implemented for Mana client for
compatibility.
Also added a ResourceRef class for automating the Resource reference
counting.
Closes #44
|
|
modernize-use-auto
modernize-use-nullptr
modernize-use-override
modernize-use-using
|
|
An optimization in the tile layer rendering code meant for drawing
repeated tiles faster was not taking into account the case where the
tile width does not match the width of the tile grid.
Reviewed-by: Stefan Beller
|
|
Maybe it used to be necessary for these methods to be virtual, but this
is no longer the case. Hence we can avoid wasting CPU ticks searching
through virtual function tables, especially for frequently used methods
like getWidth, getHeight and setAlpha.
MapLayer::getTile was inlined.
Reviewed-by: Erik Schilling
|
|
It makes no sense to mark a return value as const when it is returned by
value. This only makes sense if the return value is passed by reference, in
order to prevent the receiver from modifying the value.
Also made some other small adjustments. A std::string does not need to be
initialized to "" explicitly and getSafeUtf8String could take its parameter by
reference.
Reviewed-by: Yohann Ferreira
|
|
On the fringe layer, actors are drawn 'sorted' together with the tiles. When
the sorting order was separated from the actual Y position of the actor, the
tile drawing loop was not adapted to take this drawOrder into account rather
than the plain Y position.
Also, ActorSprite::draw was applying a half-tile offset to position the sprite
at the bottom while the logical position of the actor is at the center of the
tile. However, it failed to override getDrawOrder to account for this offset,
causing actors to get drawn earlier than they should (and thus being overlapped
by fringe layer tiles when they actually shouldn't).
This fixes drawing glitches with the paths around Hurnscald and reduces the
glitches when walking up/down through grass.
Reviewed-by: Erik Schilling
|
|
This reverts commit 92fd074aa85e2357bfe1ab642209dd5a0d87e4d5.
Reverting because these checks should be done server-side and because having
them in the client would currently require wrapping them up in checks if
there is even a current map due to the reliance on tile size (we need to
handle the case where the current map didn't load properly without crashing).
Conflicts:
src/map.cpp
src/net/manaserv/beinghandler.cpp
Reviewed-by: Erik Schilling
|
|
Better to show non-rotated/flipped tiles than no tile at all. This also
fixes interpretation of collision tiles that happen to be flipped.
Also interpret the gid as an unsigned number, since that's how they are
written in the TMX file since the introduction of these flags.
Reviewed-by: Yohann Ferreira
|
|
|
|
Their name is a bit more clear with DEBUG prefix rather than MAP
prefix. They're already scoped in the Map class anyway.
MAP_NORMAL was changed to DEBUG_NONE to represent no debug flags.
Acked-by: Bertram
|
|
Conflicts:
CMakeLists.txt
src/map.cpp
src/winver.h
|
|
The layer rendering code was not prepared to handle tiles that were
wider than the tile width of the map.
This commit also fixes the initialization of the maximum tile height,
which was based on the map height rather than the tile height. This
could slightly reduce overdraw for some maps.
Reviewed-by: Stefan Beller <stefanbeller@googlemail.com>
|
|
This means that the order point of the sprites relative to the
particles is no longer the lowest point of the image but
instead a point which is approximately between the feet of
the characters.
The intent of the latest commits to treat sprites as perpendicular to the
ground instead of perpendicular to the view line is retained by this
approach.
I tested this with various particle effects and it results in exactly the
expected behavior. Note that this does NOT fix the current problems on TMW
with the snail slime effect, because the TMW content team accidently placed
this one 10px in the air. Sorry, garbage in, garbage out.
getDrawPixelY was re-renamed to getPixelY to be consistent with getPixelX,
while getPixelY was renamed to getDrawOrder, to make its purpose clear.
Further, particles are no longer drawn when behind other objects. This
is implemented by adding a drawnWhenBehind member function to Actor,
which currently returns true for everything except particles and compound-
sprites with more than one sub-sprites (the later case is consistent with
the previous behavior of them).
An exception are text particles which are excempt from this exception and
whose drawing order is also biased by 16 pixels south instead of north
to make them more visible.
Plus some minor changes from Bertram.
Reviewed-by: Bertram.
Resolves: Mana-Mantis #362.
|
|
|
|
- Made the map teleport distance fixed for manaserv.
- Small cleanups.
The branch is considered reviewed by: Cody.
Resolves Mana-Mantis: #74.
|
|
|
|
Reviewed-by: Bjorn
|
|
Reviewed-by: thorbjorn
|
|
Reviewed-by: Bertram
|
|
This is fixng many issues and (hopefully) will make the movement
rendering much smoother.
Merge branch 'master' of gitorious.org:~bertram/mana/mana-movement-code-merge
Conflicts:
src/being.cpp
src/net/manaserv/beinghandler.cpp
Resolves: TMW-Mantis #946.
Reviewed-by: Thorbjorn.
|
|
- Removed unnecessary comments
- Removed a useless contains() check in getTileCenter()
- Fix the above function documentation
- Don't permit FloorItem to be created without a map object.
Reviewed-by: Thorbjorn Lindeijer.
|
|
I have to admit I contributed a large part of these. Sorry for that.
Less empty space, more attention to the code.
Acked-by: Jared Adams
|
|
and made use of it through the tA code.
I also remove the obsolete TODO in the flooritem class.
|
|
|
|
I made it so that the behaviour can be changed with only
a boolean setting in the playerhandler.
|
|
Previously, the walkmask was hardcoded to handling walls.
Trivial.
|
|
This prevent some weird things happening in path finding when
playing for a very long time.
Reviewed-by: Thorbjorn.
|
|
It could lead to (even small) curious desyncs making things worse
to debug.
Trivial.
|
|
Conflicts:
src/being.cpp
src/client.cpp
src/commandhandler.cpp
src/gui/setup_video.cpp
src/gui/socialwindow.cpp
src/gui/viewport.cpp
src/gui/widgets/browserbox.cpp
src/gui/widgets/itemcontainer.cpp
src/imageparticle.cpp
src/localplayer.cpp
src/localplayer.h
src/map.cpp
src/net/tmwa/beinghandler.cpp
src/particle.cpp
src/particle.h
src/player.cpp
src/player.h
|
|
- Now OpenGL and the transparency disabling are set at startup and
not read again for displaying graphics, preventing graphic errors
before startup.
- We also agreed long time ago that SDL specific functions
should have a SDL prefix. The header has been rearranged a bit
to do so.
- Also fixed a possible discrepancy in the hasAlphaChannel() function.
Reviewed-by: CodyMartin.
Resolves: Mana-Mantis: #260.
|
|
This disables the Image::setAlpha() function, which
uses 60% of the client's CPU cycles.
When enabled, visual quality is slightly decreased,
especially with the particle system.
Toggling this setting On from an Off state requires
a client restart or the graphics look quite funny.
Bertram's addition:
- Renamed 'Low CPU' to 'Disable transparency (Low CPU)' in the gui
for better understanding.
- Removed the sprite display with 30% opacity when disabling
transparency since it made monsters and drops be drawn above all layers
at full opacity.
- Made the OpenGL mode disable the 'low CPU mode'.
- Fixed the GUI logic as much as possible. Please note that the GUI opacity
slider stays enabled when transparency is disabled even if told to be disabled
in that case.
Reviewed-by: CodyMartin, 4144, MadCamel.
|
|
Please note that I didn't turned all the getValue() call into new ones,
simply because I have to have config object initiated which is not
forcefully the case the branding file.
Resolves: Manasource Mantis #170.
|
|
Reviewed-by: Bertram
|
|
Conflicts:
src/gui/itempopup.cpp
src/item.cpp
src/monster.cpp
src/net/manaserv/playerhandler.cpp
src/net/tmwa/partyhandler.cpp
src/npc.cpp
src/player.cpp
src/resources/itemdb.cpp
src/resources/monsterdb.cpp
src/resources/monsterinfo.cpp
src/resources/npcdb.cpp
src/resources/spritedef.cpp
|
|
This change allows extending tilesets without causing issues.
Reviewed-by: Freeyorp
|
|
Conflicts:
data/graphics/CMakeLists.txt
data/graphics/Makefile.am
src/client.cpp
|
