Previously, `height` was derived from a manually chosen glyph (typically
"I", Standard Cap Height) using FontBuilder, and `lineSpacing` was used
inconsistently for layout logic as a height.
Now, with the FreeType-based system:
- `height` uses `face->size->metrics.height`, which includes the
recommended line height with internal leading/line gap as defined by
the font designer.
- `lineSpacing` was removed
- `GetCapHeight` uses the standard cap height + ascender to have a
visual virtual alignment
This change standardizes font metric usage:
- Use `height` for vertical layout and line progression.
- Use `GetCapheight` in layout engines like `CGUIText` or `CGUIString`,
This ensures better alignment across fonts and consistent spacing in
multiline text rendering.
Fixes: #7962
The renderer instance is created with new, so use delete.
Direct leak of 80 byte(s) in 1 object(s) allocated from:
#0 0x7fd3c932870b in operator new(unsigned long) (/usr/lib/gcc/x86_64-pc-linux-gnu/15/libasan.so.8+0x12870b)
#1 0x55bbedf25fb0 in InitGraphics(CmdLineArgs const&, int, std::vector<CStr8, std::allocator<CStr8> > const&, ScriptContext&, ScriptInterface&) ../../../source/ps/GameSetup/GameSetup.cpp:654
#2 0x55bbed2bb95e in RunGameOrAtlas ../../../source/main.cpp:684
#3 0x55bbed271b28 in main ../../../source/main.cpp:759
#4 0x7fd3c665f3ed in __libc_start_call_main (/lib64/libc.so.6+0x263ed)
Fixes: #7951
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
This commit enhances text rendering by leveraging FreeType and a dynamic
font atlas. Previously, GUI scaling relied on bitmap fonts, which led to
blurry and distorted text when scaling up. Now, we scale the font size
directly using FreeType, resulting in much sharper and more readable
text at any GUI scale.
Key improvements:
- Replaced bitmap font scaling with true font size scaling via FreeType.
- Reduced glyph cache dependency on fixed positions, relying on shader
scaling instead.
- Switched to float-based glyph positioning for more accurate rendering
and scaling.
These changes ensure clean and consistent text appearance across
different GUI scales.
This fixes a crash when pyrogenesis got compiled with certain
compilers, whenever a command is entered in the console, while no
console.txt log exists.
Fixes#7964
Shrinking GCs dump the JITted code, which leads to redundant recompilations, lowers performance, and makes profiling JS more difficult.
They may still happen if the runtime is at risk of OOM.
Theere is now an `Engine.startAtlas` property that will start Atlas
when it's returned from the root page. The `Engine.RestartInAtlas`
function is removed.
Users may generate key presses that don't map to a distinct SDL scancode
and will be mapped SDL_SCANCODE_UNKNOWN instead. Using the same ID for
unmapped hotkeys means any such key press will execute unset hotkeys. As
luck would have it in #7644 this leads to calling "Custom exit to
desktop" if the hotkey is unbound as is the default.
So simply use a code for unused hotkeys that doesn't map to any SDL
scancodes.
Fixes: #7644
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
When the root page gets closed the engine quits instead or crashing.
The root pages are changed to use that mechanism to quit the engine.
This removes the need of `Engine.Exit` for the GUI.
Either a return value or an exception has to be stored. With
`std::variant` it's more explicit.
Also the return value is never empty, so there is no need for
`std::optional` anymore.
This avoids collisions in the user report, fixes#7174.
Update the user report version to account for the new build version
format, fixes#7173.
The build version displayed in the GUI is kept at 5 characters for
main menu clutter concerns.
With curl 8.10 [1] calling curl_multi_remove_handle retruns
CURLM_BAD_EASY_HANDLE iff the handle was removed already and no other
easy handle were registered. This resulted in ENUSRE triggering a
segfault during shutdown as reported. This was fixed upstream after
8.11.1 [2].
Do a runtime version check and for affected version allow
CURLM_BAD_EASY_HANDLE.
[1] curl ba235ab269080dc66e35835c829f7ac4290dbc1d
[2] curl 713182bd196bba014ba77f71176fea3de2236724
Fixes: #7295
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
Only the sender and the recipients receive the chat messages.
This commit only has an affecto on messages where the addressee(s) are
selected through the dropdown. Addressee(s) selected with a "/" command
are still sent to evevyone and filteret by the receiver.
`Engine.SetViewedPlayer` and `Engine.SetPlayerID` could be used to
reveal the map from GUI scripts and the in game console.
This is prevented by querying the simulation whether this player is
allowed to call thous functions.
These two vulnerabilities were introduced with their respective
features:
20e7d2224a introduced SetPlayerID to allow controlling other players
using the developer overlay.
a2f7d4d82a introduced SetViewedPlayer to allow observers to change the
perspective.
This replaces all links pointing to trac with their corresponding links
to gitea. Also replace http with https while at it.
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
We use a shader with 64 bones by default to consume less uniform memory.
But if we meet bigger skeletons we batch and handle them afterwards.
In the future we need to sort the input models to have less changes for
bound buffers. Also we might want to skin up to 4 models per a single
dispatch.
- Shallow-freezing of objects is never used in our codebase, so remove
that code path.
- Deep-freeze is bugged in recent versions of SpiderMonkey (see bug
https://bugzilla.mozilla.org/show_bug.cgi?id=1930258). Until a fix
and/or a better API is provided, reimplement this feature by
recusively freezing object properties.
- Add tests for the deepfreeze function.
Some return values can be assigned to a `const auto`, somewhere `dir` is
mutated so the whole result has to be assigned to a non `const` variable
and once only dir is used, so structured binding isn't used there.
In `Future` there is a notion of cancelation / stop-request. The task
callback doesn't have such a notion.
Some tasks (like the map-generation) are stopable. It did that in a
thread unsave way.
A task is canceled when the future is destroied or when `CancelOrWait`
is called on it.
To disable tests we carry a patch which allows disabling test by
appending DISABLED to the test function name. Instead just do as
upstream says and prepend the test function so it won't be recognized as
a test any longer.
The docs suggest to prepend x but anything will do. Continue to use
DISABLED_ but as prefix which is actually already in use in one case.
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
Stan