fix: Improved appearance for health screen

[tilden]

This change fixes the poor rendering of the rotated 4-point square in the HRM UI, and improves the contrast by adding a background color to the heart icon.

Background
In #850, the health screen contrast was improved by changing the background color to white, and by replacing the glitched square with a circle. However, this causes the heart rate screen to no longer visually mesh with the rest of the health UI (which is dark, and uses angular shapes.)

After discussion, @hellcp determined that the root cause for the glitched rectangle was incorrect metrics. By correcting these, it is now possible to display a rectangle that does not glitch:

Furthermore, the low-contrast heart graphic appears to be caused by the heart having a transparent, as opposed to white, background color. This commit should also correct this:
Screencast From 2026-02-18 08-57-23.webm

Thanks to @hellcp for figuring all this out and implementing the changes over in pebble-dev/pebble-firmware#319

[gmarull]

I prefer the current color scheme; it's just more readable. Happy to change circle back to the fixed segments, though.

[tilden]

I prefer the current color scheme; it's just more readable. Happy to change circle back to the fixed segments, though.

@gmarull Gotcha! Just pushed a change such that the colors should now match what they were (white bg, dark red foreground). Let me know if you need additional changes!

[gmarull]

I prefer the current color scheme; it's just more readable. Happy to change circle back to the fixed segments, though.

@gmarull Gotcha! Just pushed a change such that the colors should now match what they were (white bg, dark red foreground). Let me know if you need additional changes!

Thanks! please fixup original commit, we do not merge intermediate work ;-)

[gmarull]

Note we're thinking to make all the health menus lighter to make things consistent and overall more readable

[tilden]

Thanks! please fixup original commit, we do not merge intermediate work ;-)

Done! Wasn't really sure what the guideline on that was, so thanks for clarifying!

[gmarull]

Since this seems to be derived work from pebble-dev/pebble-firmware#319, can the original author confirm it's Apache-2.0? pebble-firmware messed up licenses a while ago, so I want to be sure.

[tilden]

Note we're thinking to make all the health menus lighter to make things consistent and overall more readable

Ah I see! I think consistency could be good there. Might be interesting to play with the colors a bit; perhaps it could borrow from the old OS 3.x health color scheme 🤔

[tilden]

Since this seems to be derived work from pebble-dev/pebble-firmware#319, can the original author confirm it's Apache-2.0? pebble-firmware messed up licenses a while ago, so I want to be sure.

Sure! I have confirmation from the author on Discord that it's Apache-2.0:

before anyone asks, feel free to copy this into the core repo, and also show me what that looks like
(...)
the file I modified was an apache file and I didn't touch the license there

https://discord.com/channels/221364737269694464/1337430537132048435/1473538292879458357
https://discord.com/channels/221364737269694464/1337430537132048435/1473538474433843211

[gmarull]

pls fix CI issues

[tilden]

Dang, seems like the tests are failing:

  1) Failure:
test::health::hr::summary::card::render_current_bpm (no description) [../../tests/fw/apps/system_apps/health/test_health_hr_summary_card.c:82] [-t0]
  Expression is not true: gbitmap_pbi_eq(&s_ctx.dest_bitmap, namecat(__func__, ".pbi"))

  2) Failure:
test::health::hr::summary::card::render_no_data (no description) [../../tests/fw/apps/system_apps/health/test_health_hr_summary_card.c:73] [-t1]
  Expression is not true: gbitmap_pbi_eq(&s_ctx.dest_bitmap, namecat(__func__, ".pbi"))

  3) Failure:
test::health::hr::summary::card::render_timestamp (no description) [../../tests/fw/apps/system_apps/health/test_health_hr_summary_card.c:94] [-t2]
  Expression is not true: gbitmap_pbi_eq(&s_ctx.dest_bitmap, namecat(__func__, ".pbi"))

I'm guessing they're slightly different due to the changed metrics. Is there a way to extract the generated test images from the run? In the meantime I'll try regenerating locally.

[gmarull]

yes, run tests and look into build/test/tests/failed

[tilden]

Unfortunately, seems like there might be issues running the tests on MacOS at the moment; while I can build the firmware fine, the tests always fail with mysterious linker errors:

ld: pointer not aligned in '_s_shadow_bottom_bitmap' (/Users/dan/Development/Pebble/official/PebbleOS/build/test/8f3f6b6b920de209330d4bd28ec690a0/src/fw/applib/ui/shadows.o)
clang: error: linker command failed with exit code 1 (use -v to see invocation)

Seems someone else on the Discord was also unable to get the tests building on Mac. Is there a way to pull them out of the CI runner directly? Might be easier than fighting with it locally.

Edit: managed to get it working!

[tilden]

Got it! Turns out it's possible to get the tests to run on Mac if you disable some various checks in the wscript:

diff --git a/tests/wscript b/tests/wscript
index 8b12f6cd..63773d8e 100644
--- a/tests/wscript
+++ b/tests/wscript
@@ -374,6 +374,12 @@ def build(bld):
         bld.env.append_value('CFLAGS', '-fprofile-arcs')
         bld.env.append_value('CFLAGS', '-ftest-coverage')
         bld.env.append_value('LINKFLAGS', '--coverage')
+
+    import sys
+    if sys.platform == 'darwin':
+        bld.env.append_value('LINKFLAGS', ['-Wl,-no_fixup_chains'])
+        bld.env.append_value('DEFINES', 'DUMA_DISABLED') 
     test_wscript_dirs = [os.path.dirname(f.abspath()) for f in bld.path.ant_glob('**/wscript')]
     for dir in test_wscript_dirs:
         bld.recurse(dir)

I uploaded the new test images, so CI should pass now, hopefully.