This commit refactors RepeatablePath to contain a list of tagged unions
containing "optional" and "required" variants. Both variants have a null
terminated file path as their payload, but the tag dictates whether the
path must exist or not. This implemenation is used to force consumers to
handle the optional vs. required distinction.
This also moves the parsing of optional file paths into RepeatablePath's
parseCLI function. This allows the code to be better unit tested. Since
RepeatablePath no longer contains a simple list of RepeatableStrings,
many other of its methods needed to be reimplemented as well.
Because all of this functionality is built into the RepeatablePath type,
other config options which also use RepeatablePath gain the ability to
specify optional paths as well. Right now this is only the
"custom-shaders" option. The code paths in the renderer to load shader
files has been updated accordingly.
In the original optional config file parsing, the leading ? character
was removed when paths were expanded. Thus, when config files were
actually loaded recursively, they appeared to be regular (required)
config files and an error occurred if the file did not exist. **This
issue was not found during testing because the presence of the
"theme" option masks the error**. I am not sure why the presence of
"theme" does this, I did not dig into that.
Now because the "optional" or "required" state of each path is tracked
in the enum tag the "optional" status of the path is preserved after
being expanded to an absolute path.
Finally, this commit fixes a bug where missing "config-file" files were
not included in the +show-config command (i.e. if a user had
`config-file = foo.conf` and `foo.conf` did not exist, then `ghostty
+show-config` would only display `config-file =`). This bug applied to
`custom-shaders` too, where it has also been fixed.
- move wuffs code from src/ to pkg/
- eliminate stray debug logs
- make logs a warning instead of an error
- remove initialization of some structs to zero
Improve the performance of Kitty graphics by switching to the WUFFS
library for decoding PNG images and for "swizzling" G, GA, and RGB data
to RGBA formats needed by the renderers.
WUFFS claims 2-3x performance benefits over other implementations, as
well as memory-safe operations.
Although not thorougly benchmarked, performance is on par with Kitty's
graphics decoding.
https://github.com/google/wuffs
Before, cells that were explicitly set to match the default bg color
were treated as if they did NOT have the default and extending would
occur. We now check the exact RGB of each cell.
Fixes#2099
This is another heuristic of sorts to make `window-padding-color=extend`
look better by default. If a fully covering glyph is used then we use
the fg color to extend rather than the background.
This doesn't account for fonts that may do this for whatever codepoints,
but I think that's a special scenario that we should just recommend
disabling this feature.
Before this fix, if vsync was on the GPU cells buffer could be cleared
for a frame while resizing the terminal down. This was due to the fact
that the surface sent messages for the resize to both the renderer and
the IO thread. If the renderer thread was processed first then the GPU
cells buffer(s) would be cleared and not rebuilt, because the terminal
state would be larger than the GPU cell buffers causing updateFrame to
bail out early, leaving empty cell buffers.
This fixes the problem by changing the origin of the renderer's resize
message to be the IO thread, only after properly updating the terminal
state, to avoid clearing the GPU cells buffers at a time they can't be
successfully rebuilt.
This happened to work in releases somehow but Xcode debug builds would
catch this as an assertion. Our cell bg pipeline now uses the "full
screen vertex shader" which takes no parameters, so we don't need a
vertex descriptor.
- Significant changes to optimize memory usage.
- Adjusted formatting of the metal shader code to improve readability.
- Normalized naming conventions in shader code.
- Abstracted repetitive code for attribute descriptors to a helper
function.
With a minimum contrast set, the colored glyphs that Powerline uses
would sometimes be set to white or black while the surrounding background
colors remain unchanged, breaking up contiguous colors on segments of
the Powerline.
This no longer happens with this patch as Powerline glyphs are now
special-cased and exempt from the minimum contrast adjustment.
There are scenarios where this configuration looks bad. This commit
introduces some heuristics to prevent it. Here are the heuristics:
* Extension is always enabled on alt screen.
* Extension is disabled if a row contains any default bg color. The
thinking is that in this scenario, using the default bg color looks
just fine.
* Extension is disabled if a row is marked as a prompt (using semantic
prompt sequences). The thinking here is that prompts often contain
perfect fit glyphs such as Powerline glyphs and those look bad when
extended.
This introduces some CPU cost to the extension feature but it should be
minimal and respects dirty tracking. This is unfortunate but the feature
makes many terminal scenarios look much better and the performance cost
is minimal so I believe it is worth it.
Further heuristics are likely warranted but this should be a good
starting set.
