Canvas_SortComponentPool now takes the raw canvas component id as it is
specialized to the canvas subpools.
Canvas_SetLen resizes the root view and then updates the hierarchy for
every canvas in the system.
Canvas_InitSys sets up the component system with the systems it needs
(canvas, view, text). This is required to ensure view_href is just past
the canvas components as it is needed for retrieving the actual canvas
component (and thus sub-pool range ids) from arbitrary views in the
canvas.
Entities are fetched with the correct offset from the pool entities.
This will make it easy for client code to set up data needed by the
console before the console initializes. It already separates console
cvar setup and initialization, which has generally been a good thing.
The flashing pink around the Q menu cursor was caused by vulkan command
buffer writes and draw queue population being out of phase, which was
fixed by the recent screen update changes (specifically,
42441e87d4).
Rather important for debugging 2d stuff (draw's lines are 2d-only).
Other than translucent console, this gets the vulkan draw api back to
full operation.
This needed either more font ids to be supported, or small lump pics (up
to 32 x 32) to be loaded into the atlas. I went with both. The menus
don't use Draw_TextBox, but quakeworld's netgraph does.
This makes use of slice rendering to achieve the effective scaling, but
the slice data is created only when needed so pics that never use slices
don't waste 16 vertices.
The goal is to get vulkan relying on the "renderpass" abstraction, but
this gets vulkan up and running again, and even fixes the rendering
issues (in the end, getting canvas working wasn't required, but is still
planned).
This is a bit of a hack to allow me to work on vulkan's screen update
"pipeline" without having to mess with the other renderers, since it
turns out they're (currently) fundamentally incompatible.
When a pic needs dynamic vertices (eg, for sub-pics), a descriptor set
is allocated and updated if one has not been created for the pic. This
is done each frame: the descriptor sets are recycled (there currently is
rarely a need for more than a small handful of dynamic descriptors, so
64 should be plenty for now).
Unfortunately, due to the order of operations issue between draw items
getting queued and submitting commands to vulkan (the cause of the pics
not rendering correctly per 8fff71ed4b),
the validation layers complain (correctly) about the command buffers
being executed with updated descriptor sets. Getting the canvas system
up and running will fix that.
The pic is scaled to fill the specified rect (then clipped to the
screen (effectively)). Done just for the console background for now, but
it will be used for slice-pics as well.
Not implemented for vulkan yet as I'm still thinking about the
descriptor management needed for the instanced rendering.
Making the conback rendering conditional gave an approximately 3% speed
boost to glsl with the GL stub (~12200fps to ~12550fps), for either
conback render method.
The wording might seem a little odd, but cl_screen is really the full 2D
client HUD while the console is completely independent of the client and
shouldn't know that the client even exists. Ideally, the resize events
would be handled by the canvas system, to which end this is a small
step.
This fixes the broken dynamic lighting in fisheye rendering. It does
mean that frustum culling of lit surfaces needed to be removed, but if
not doing frustum culling on lit surfaces was good enough for a P90,
it's probably good enough for an i7-6850K.
They are usually larger images (eg, the main menu graphic) and thus make
a mess of the atlas (thus, making them separate means a smaller atlas
can be used). All sorts of things are in a mess (descriptor management,
subpic rendering not supported, wrong alpha value for the transparent
pixel), but this gets the basic loader going.
This just takes advantage of the dynamic verts for doing subpics. It's
not really the most optimal code as it has to write both the vertices
(64 bytes per quad) and the instances (24 bytes per quad), but that's
still better than the old 128 bytes per quad (and having a single
pipeline is nice).
The problem was that I had mixed up the purpose of the per-frame vertex
buffers and used them for the core quad data when they were meant for
subpic and the like, and forgotten about the static vertex buffer.
This gets at least conchars working (pic in general not tested yet).
Any performance gains will be utterly swamped by the deferred renderer,
but it will allow better control of quad render order by any client
code (and should be slightly better for simpler renderers when I get
support for them working).
Right now, plenty is broken (much of the higher level draw functions are
disabled, and pics don't render correctly), but this gets at least the
basics in so I'm not bouncing diffs around as much.
It turns out the slice pipeline is compatible with the glyph pipeline in
that its vertex attribute data is a superset (just the addition of the
offset attributes). While the queues have yet to be merged, this will
eventually get glyphs, sliced sprites, and general (static) quads into
the one pipeline. Although this is slightly slower for glyph rendering
(due to the need to pass an extra 8 bytes per glyph), this should be
faster for quad rendering (when done) as it will be 24 bytes per quad
instead of 32 bytes per vertex (ie, 128 bytes per quad), but this does
serve as a proof of concept for doing quads, glyphs and sprites in the
one pipeline.
The main reason I had created in the first place was I hadn't thought of
using image view swizzles to handle coverage-alpha textures (for
monochrome glyphs), and for whatever reason also had the texture in a
different binding slot to the twod fragment shader. With both issues out
of the way, there's no reason to have an almost identical (just some
naming) shader just for glyphs.
With an eye towards merging the 2d pipelines as much as possible, I
found that the glyph and basic 2d quad texture descriptors were in
different slots for no reason I can think of. Having them in the same
slot would mean I could use the same fragment shader for all 2d
pipelines (though the plan is to get it down to two: (sliced) quads and
lines).
I hadn't noticed the problem until playing with early fragment tests for
the sprite fragment shaders, but passing data that expects triangle
strips to a pipeline that expects triangle lists doesn't work too well
when drawing quads.
This is the beginning of supporting 2d rendering in 3d space. The idea
is that a canvas can be in 2d orthographic space (not attached to any
entity with a 3d transform), or in 3d perspective space (attached to an
entity with a 3d transform, either as a child of the camera, or of some
object in 3d space).
It will replace the current HUD code when it's working.
I found I needed the subrange start as well as the end, but I liked that
the subpools themselves used only the end of the range, so switching to
just a unint32_t for the value and adding a function to return a tuple
made sense. I had kept the struct because I thought I might want to
store additional information (eg, the entity "owning" the subpool), but
found that I didn't need such information as the systems using subpools
that way would have access to the entity by other means.
Interestingly, the change found a bug in subpool creation: I really
don't know why things worked before, but they work better now :)
Subpools are for grouping components by some criterion. Any component
that has a rangeid callback will be grouped with other components that
return the same render id. Note that the ordering of components within a
group will be affected by adding a component into a group that comes
before that group (or removing a component).
Component pools can have multiple groups, added and removed dynamically,
but removing a group should (currently) be done only when empty.
While "set" is a tad strong (there's just the one component for now), I
had missed the changes when adding ECS systems. Fixes the segfault at
the end of demo1 (ie, when any center text is printed).
Instead of creating new entities for the text views. This approximately
halves the number of entities required to display flowed text, but also
tests the ability to have an entity in multiple hierarchies (the goal of
the ECS component and system changes).
Marking them as cached means that they'll be "uncached" instead of
destroyed when freed, which would not be a particularly good thing. I
have no memory as to how I found this as I found the change in my git
stash.
While this does require an extra call after registering components, it
allows for multiple component sets (ie, sub-systems) to be registered
before the component pools are created. The base id for the registered
component set is returned so it can be passed to the subsystem as
needed.
There's now a main ecs.h file that includes the sub-system headers,
removing the need to explicitly include several header files, but the
sub-systems are a less cluttered.
This means that the component id used for hierarchy references must be
passed to Hierarchy_New and Hierarchy_Copy, but does all an entity to
have more than one hierarchy, which is useful for canvases (hierarchies
of views) in the 3d world (the canvas root would have a 3d hierarchy
reference and a 2d (view) hierarchy reference).
It seems that the mouse escaping the barriers requires some combination
of hitting two at once, and holding your mouth just right (something
about sliding the mouse up and down one barrier near the other).
However, sending the mouse back to the center of the screen when it
touches a barrier makes such sliding impossible.
This seems to fix#38
I obviously need a better way to test legacy code because the fix for
unsigned-int behavior with clang broke mouse warping when using
XGrabPointer instead of XInput2's XIGrabEnter.
The separation now uses height above (right of) the base line, and depth
below (left of) the base line. This puts the text exactly where I want
it, but there's still the problem of uneven line spacing caused by
descenders and ascenders. However, I suspect that's more up to the
text/font handling code to get the boxes right (maybe set spaces to have
the right dimensions?).
The main problem was the confusion about the coordinates within a single
glyph, and thus the glyphs position within the view's box. With this,
flowed text works fairly well except for some issues with spacing
between lines (which I think is due to the flow code not having been
tested with offset boxes).
While Draw_Glyph does draw only one glyph at a time, it doesn't shape
the text every time, so is a major win for performance (especially
coupled with pre-shaped text).
Font cannot be overridden yet, but script attributes (language, script
type, direction) and features can be set at all three levels in a
passage. Attributes on the root level act as defaults for the paragraph
and word levels, and paragraph attributes act as defaults for the word
level.
Passage_Delete needs to check if the hierarchy is valid as no text may
have been added, which results in a null pointer for the hierarchy.
Text shaping needs to set language etc every time it resets the buffer.
This causes some problems with linking if libQFgui is linked with
libQFrenderer (which is necessary in the long run), but it seems
everything gets away with it for now (which, tbh, I don't like).
And add a function to process a passage into a set of views with glyphs.
The views can be flowed: they have flow gravity and their sizes set to
contain all the glyphs within each view (nominally, words). Nothing is
tested yet, and font rendering is currently broken completely.
Font and text handling is very much part of user interface and at least
partially independent of rendering, but does fit it better with GUI than
genera UI (ie, both graphics and text mode), thus libQFgui as well as
libQFui are built in the ui directory.
The existing font related builtins have been moved into the ruamoko
client library.
I had done the loader for the GPU renderers, so the CPU renderer didn't
draw the characters transparently. Fixes the pink block in my ruamoko
test scene (due to the notify text area).
While it doesn't really make any difference to the texture upload (8-bit
is 8-bit), and the sampler is in control of the interpretation, this
makes vulkan more consistent with the specification of the glyph
texture.
In theory, it supports all the non-palette formats, but only luminance
and alpha (tex_l and tex_a) have been tested. Fixes the rather broken
glyph rendering.
World scale can only be approximate if non-uniform scales and
non-orthogonal rotations are involved, but it is still useful
information sometimes.
However, the calculation is expensive (needs a square root), so remove
world scale as a component and instead calculate it on an as-needed
basis because it is quite expensive to do for every transform when it is
used only by the legacy-GL alias model renderer.
Thanks to the 3d frame buffer output being separate from the swap chain,
it's possible to have a different frame buffer size from the window
size, allowing for a smaller buffer and thus my laptop can cope (mostly)
with the vulkan renderer.
The escape was actually harmless as the buffers would not be read due to
the particle count being 0 (thus why the buffers were at the end of the
staging buffer: no space was allocated for them, only for the system
buffer, but their offsets were just past the system buffer). However,
the validation layers quite rightly did not like that. Thus, the two
buffers are pointed to the system buffer so all three descriptors are
always valid.
Where too far is 1024 units as that is the maximum supported, or the
radius. The change to using unsigned for the distances meant the simple
checks missed the effective max dist going negative, thus leading to a
segfault.
I had debated putting the blending in the compose subpass or a separate
pass but went with the separate pass originally, but it turns out that
removing the separate pass gains 1-3% (5-15/545 fps in a timedemo of
demo1).
viewstate's time is from cl.time which is not what's used to set
last_servermessage (that uses realtime). After careful investigation, I
found that cl.time is not at all suitable and that the original id code
used realtime (I think it was just me being lazy when I merged the
code). Fixes the stuck net icon.
quake changes rocket and grenade models to explosion models, but
quakeworld does not. This resulted in nq drawing two explosion sprites
instead of one. Separating the types allows nq to skip adding a sprite
for the explosion.
It's a bit flaky for particles, especially at higher frame rates, but
that's due to supporting only 64 overlapping pixels. A reasonable
solution is probably switching to a priority heap for the "sort" and
upping the limit.
This required making the texture set accessible to the vertex shader
(instead of using a dedicated palette set), which I don't particularly
like, but I don't feel like dealing with the texture code's hard-coded
use of the texture set. QF style particles need something mostly for the
smoke puffs as they expect a texture.
It doesn't want to work on my nvidia (or more recent sid?) and doesn't
seem to be necessary. The problem may be multiple event sets before the
first wait, but investigation can wait for now.
This is probably the biggest reason I had problems with particles not
updating correctly: the descriptors were generally point pointing to
where the data actually was in the staging buffer.
I don't yet know whether they actually work (not rendering yet), but the
system isn't locking up, and shutdown is clean, so at least resources
are handled correctly.
Although it works as intended (tested via hacking), it's not hooked up
as the current frame buffer handling in r_screen is not readily
compatible with how vulkan output is handled. This will need some
thought to get working.
This splits up render pass creation so that the creation of the various
resources can be tailored to the needs of the actual render pass
sub-system. In addition, it gets window resizing mostly working (just
some problems with incorrect rendering).
If the result object type pointer is null, then the parsed result type
and value pointers are written directly to the result object rather than
testing the parsed result type against the object type and copying the
parsed result value data to the location of the object value. It is then
up to the caller to check the type and copy the value data.
It turns out the semaphore used for vkAcquireNextImageKHR may be left in
a signaled state for VK_ERROR_OUT_OF_DATE_KHR. While it seems to be
possible to clear the semaphore using an empty queue submission,
destroying and recreating the semaphore works well.
Still have problems with the frame buffer after window resize, though.
Swap chain acquisition is part of final output handling. However, as the
correct frame buffers are required for the render passes, the
acquisition needs to be performed during the preoutput render pass.
Window resize is still broken, but this is a big step towards fixing it.
This is the minimum maximum count for sampled images, and with layered
shadow maps (with a minimum of 2048 layers supported), that's really way
more than enough.
I guess nvidia gives a non-srgb format as the first in the list, but my
laptop gives an srgb format first, thus the unexpected difference in
rendering brightness. Hard-coding BGRA isn't any better, but it will do
for now.
Things are a bit of a mess with interdependence between sub-module
initialization and render pass initialization, and window resizing is
broken, but the main render pass rendering to an image that is then
post-processed (currently just blitted) is working. This will make it
possible to implement fisheye and water warp (and other effects, of
course).
When working, this will handle the output to the swap-chain images and
any final post-processing effects (gamma correction, screen scaling,
etc). However, currently the screen is just black because the image
for getting the main render pass output isn't hooked up yet.
Now each (high level) render pass can have its own frame buffer. The
current goal is to get the final output render pass to just transfer the
composed output to the swap chain image, potentially with scaling (my
laptop might be able to cope).
While the HUD and status bar don't cut out a lot of screen (normally),
they might start to make a difference when I get transparency working
properly. The main thing is this is a step towards pulling the 2d
rendering into another render pass so the main deferred pass is
world-only.
Using swizzles in an image view allows the same shader to be used with
different image "types" (eg, color vs coverage).
Of course, this needed to abandon QFV_CreateImageView, but that is
likely for the best.
It turns out that nearest filtering doesn't need any offsets to avoid
texel leaks so long as the screen isn't also offset. With this, the 2d
rendering looks good at any scale (minus the inherent blockiness).
It seemed like a good idea at the time, but it exacerbates pixel leakage
in atlas textures that have no border pixels (even in nearest sampling
modes).
The rest of the system won't add one automatically (since entity
creation no longer does), but the alias and iqm rendering code expect
there to be one. Fixes a segfault when starting a scene (demo etc).
There's no API yet as I need to look into the handling of qpic_t before
I can get any of this into the other renderers (or even vulkan, for that
matter).
However, the current design for slice rendering is based on glyphs (ie,
using instances and vertex pulling), with 3 strips of 3 quads, 16 verts,
and 26 indices (2 reset). Hacky testing seems to work, but real tests
need the API.
I don't know why it didn't happen during the demo loop, but going from
the start map to e1m1 caused a segfault due to the efrags for a lava
ball getting double freed (however, I do think it might be because the
ball passed through at least two leafs, while entities in the demos did
not). The double free was because SCR_NewScene (indirectly) freed all
the efrags without removing them from entities, and then the client code
deleting the entities caused the visibility components to get deleted
and thus the efrags freed a second time. Using ECS_RemoveEntities on the
visibility component ensures the entities don't have a visibility
component to remove when they are later deleted.
While simple component pools can be cleared simply by zeroing their
counts, ones that have a delete function need that function to be called
for all the components in the pool otherwise leaks can happen.
It's currently used only by the vulkan renderer, as it's the only
renderer that can make good use of it for alias models, but now vulkan
show shirt/pants colors (finally).
This cuts down on the memory requirements for skins by 25%, and
simplifies the shader a bit more, too. While at it, I made alias skins
nominally compatible with bsp textures: layer 0 is color, 1 is emissive,
and 2 is the color map (emissive was on 3).
As the RGB curves for many of the color rows are not linearly related,
my idea of scaling the brightest color in the row just didn't work.
Using a masked palette lookup works much better as it allows any curves.
Also, because the palette is uploaded as a grid and the coordinates are
calculated on the CPU, the system is extendable beyond 8-bit palettes.
This isn't quite complete as the top and bottom colors are still in
separate layers but their indices and masks can fit in just one, but
this requires reworking the texture setup (for another commit).
For whatever reason, I had added an extra 4 bytes to the fragment
shader's push-constants. It took me a while to figure out why renderdoc
wouldn't stop complaining about me not writing enough data.
It turns out my approach to alias skin coloring just doesn't work for
the quake data due to the color curves not having a linear relationship,
especially the bottom colors.
It works on only one layer and one mip, and assumes the provided texture
data is compatible with the image, but does support sub-image updates
(x, y location as parameters, width and height in the texture data).
The bright end of the color map is actually twice the palette value, but
I didn't understand this when I came up with the shirt/pants color
scheme for vulkan. However, the skin texture can store only 0..1, so the
mapping to 0..2 needs to be done in the shader. It looks like it works
at least better: the gold key at the end of demo1 doesn't look as bleh,
though I do get some weird colors still on ogres etc.
Currently only for gl/glsl/vulkan. However, rather than futzing with
con_width and con_height (and trying to guess good values), con_scale
(currently an integer) gives consistent pixel scaling regardless of
window size.
Well, sort of: it's still really in the renderer, but now calling
R_AddEfrags automatically updates the visibility structure as necessary,
and deleting an entity cleans up the efrags automatically. I wanted this
over twenty years ago.
The support for the new vector types broke compiling code using
--advanced. Thus it's necessary to ensure vector constants are
float-type and vec3 and vec4 are treated as vector and quaternion, which
meant resurrecting the old vector expression code for v6p progs.
This fixes maplist showing only those maps in the user directory.
However, no checking is done for duplicate files due to earlier search
paths overriding later paths.
This involved disabling sigils for hipnotic and rogue (not used),
adjusting the number of items views, and moving the two keys views for
hipnotic. Rogue is not yet using the correct status bar pics.
The functionality of the hipnotic and rogue weapon power-ups is now done
by a various mappings instead of separate functions. In theory, this
should make things more flexible, but most importantly, there's a lot
less code duplication.
Sigils can't be flashed as they don't have any animations provided, and
they're not normally as critical. I don't know why items weren't
flashed, but since the pics are there, might as well use them (and the
flashing keys do look pretty good).
I think this makes the purpose of the functions more clear and makes the
protocol logic less dependent on the meaning of some of the updates.
Most of the update functions are not fully implemented yet.
I had forgotten that the cl structs in nq and qw were different layouts,
which resulted in qw's sbar/hud being quite broken. Rather than messing
with the structs, I decided it would be far better in the long run to
clean up sbar's access to the cl struct and the few other nq/qw specific
globals it used. There are still plenty of bugs to fix, but now almost
everything is in the one place.
In the end, it was removal of the old entries that corrupted the parent
indices. Very nicely, most of the fixes involved removing code. Taking
advantage of the ECS to debug the hierarchies was fun, and the resulting
colorized entity names helped no end.
Even 37 objects is a lot, but it's a whole lot better than 180. Most
importantly, it reproduces the problem, which seems to be not all parent
indices getting updated. The child indices seem to be working nice, as
do the reference object indices (ie, the entity components). I suspect
its the parent indices getting corrupted that cause problems on the
second switch of the hud/sbar cvar as the parent indices are used to
find the child indices that need to be updated.
This improves the behavior of hierarchies when self-inserting, but nq's
sbar still crashes when trying to do so. However, its tree is a fair bit
more complex than the test case (that does pass now), so I need to try
to replicate the important parts of the tree with fewer objects (180 is
too many to work with).
As expected, reparenting a sub-hierarchy such that it (and possibly its
children) move up the arrays fails (this is why sbar needs to first
remove the sub-hierarchy then insert it).
Since test_build_hierarchy2 already tested removal of a sub-hierarchy
(once fixed), it seems test_build_hierarchy3 testing parenting within
the same hierarchy would be a good idea. Reparenting such that
everything moves to later in the arrays works nicely (not very
surprising).
Its updates to the various indices were out, but this was missed due to
the tests being wrong. I wonder if I got interrupted while working on
them last and just assumed the removals were correct. This improves
sbar's behavior, but it's still wrong when pulling the armory view out
of the inventory. Very unsure what's going on, but the various indices
look ok, as do the view positions.
Ugh, things were quite bad, it turns out. It seems a lot of trouble
would have been saved if these tests had worked (however, something is
still not quite right as views are out of place).
This is the bug that sbar found when pulling a sub-hierarchy out of a
larger hierarchy: child indices not getting updated correctly for later
siblings and any niece objects.
The hierarchy-specific tests from the transform tests have been moved
into the ecs tests and the transform tests renamed appropriately. As
part of the process, hierarchies can now have a null type (ie, no
additional components maintained by the hierarchy). This should make
sorting out the issues highlighted by sbar a bit easier.
It should have always been here, but when I first created the hierarchy
and transform objects, I didn't know where things would go. Having two
chunks of code for setting an entity's parent was too already too much,
and I expect to have other hierarchy types. Doesn't fix the issues
encountered with sbar, of course.
The text object covering the whole passage was not being initialized,
thus center print tried to print rubbish when (incorrectly) printing the
entire message.
I'm not particularly happy with the way onresize is handled, but at this
stage a better way of dealing with resizing views and getting the child
views to flow correctly hasn't come to mind. However, the system should
at least be usable.
I'm not sure when things broke on my laptop (I thought I got warp and
fisheye working on my laptop), but it turns out things weren't quite
right, thus warp (and presumably fisheye) weren't working properly due
to GLSL errors that I only just noticed. This fixes water warp (and
probably fisheye).
This includes moving the related cvars from botn nq and qw into the
client hud code. In addition, the hud code supports update and
update-once function components. The update component is for updates
that occur every frame, but update-once components (not used yet) are
for one-shot updates (eg, when a value updates very infrequently).
Much of the nq/qw HUD system is quite broken, but the basic status bar
seems to be working nicely. As is the console (both client and server).
Possibly the biggest benefit is separating the rendering of HUD elements
from the updating of them, and much less traversing of invisible views
whose only purpose is to control the positioning of the visible views.
The view flow tests are currently disabled until I adapt the flow code
to ECS.
There seems to be a problem with view resizing in that some gravities
don't follow resizing correctly.
