Add errdefer cleanup for codepoints and glyphs hash maps in init().
Previously, if ensureTotalCapacity or reloadMetrics() failed after
allocating these maps, they would leak.
Add tripwire test to verify all failure points in init().
Add errdefer to remove cache entry after getOrPut if subsequent
operations fail (getPresentation, atlas.grow, renderGlyph). Without
this, failed renders would leave uninitialized/garbage entries in
the glyph cache, potentially causing crashes or incorrect rendering.
Add tripwire test to verify the rollback behavior.
Use fast hash function on key for better distribution.
Direct compare glyph in eql to avoid Packed.from() if not neccessary.
16% -> 6.4% reduction during profiling runs.
Move size adjustment logic out of `Entry`, I understand the impulse to
put it there but it results in passing a lot of stuff around which isn't
great.
Rework `add(...)` in to `add(...)` and `addDeferred(...)`, faces are
passed directly now instead of passing an entry, and an options struct
is used instead of positional arguments for things like style, fallback,
and size adjustment.
Change size adjustment test back to a half pixel tolerance instead of 5%
because the previous commit (allowing fractional pixel sizes) fixed the
root cause of large differences.
There are two main improvements being made here. First, we move away from using autohash and instead
use a one-shot strategy similar to the Style hashing. Since the GlyphKey includes the Metrics struct,
which contains quite a few fields, autohash was performing expensive and unnecessary repeated updates.
The second improvement is actually just, not hashing Metrics. By ignoring the Metrics field, we can
fit the rest of the GlyphKey into a 64-bit packed struct and just return that as the hash! It
ends up being unique for each GlyphKey in renderGlyph, and is nearly a zero-cost operation.
This ends up boosting the performance (on my machine at least), from around 560fps to 590fps on the
DOOM-fire benchmark.
Also changes color atlas to always use an sRGB internal format so that
the texture reads automatically linearize the colors.
Renames the misleading `rgba` atlas format to `bgra`, since both
FreeType and CoreText are set up to draw color glyphs in bgra.
This commit changes a LOT of areas of the code to use decl literals
instead of redundantly referring to the type.
These changes were mostly driven by some regex searches and then manual
adjustment on a case-by-case basis.
I almost certainly missed quite a few places where decl literals could
be used, but this is a good first step in converting things, and other
instances can be addressed when they're discovered.
I tested GLFW+Metal and building the framework on macOS and tested a GTK
build on Linux, so I'm 99% sure I didn't introduce any syntax errors or
other problems with this. (fingers crossed)
This sets the stage for dynamically adjusting the sizes of fallback
fonts based on the primary font's face metrics. It also removes a lot of
unnecessary work when loading fallback fonts, since we only actually use
the metrics based on the parimary font.
Unify grid metrics calculations by relying on shared logic mostly based
on values directly from the font tables, this deduplicates a lot of code
and gives us more control over how we interpret various metrics.
Also separate metrics for underlined, strikethrough, and overline
thickness and position, and box drawing thickness, so that they can
individually be adjusted as the user desires.
Fixes#2364
This adds the bold, italic, and bold italic variants of JB Mono so it is
built-in. This also fixes up the naming convention for the embedded font
files across tests and removes redundant embedded font files.
Fixes#1722
Previously, SharedGrid.getIndex would properly lock any access to search
for and load the face metadata for a font face that contains a
codepoint. However, that face may be "deferred" (metadata loaded but the
actual face not loaded).
Later, outside of the SharedGrid write lock, a deferred face may be
initialized and cause a data race if two threads are doing this at the
same time, sometimes loading to a crash.
This commit fixes the issue by always preloading font indexes in
getIndex because the usage of getIndex implies a very near term future
use of getFace.
