 jacqueline
|
912f9c0757
|
Free the i2s dma channel when it's not in use
It turns out this is why bluetooth playback keeps running out of memory
/facepalm
|
2023-10-19 13:01:28 +11:00 |
|
|
jacqueline
|
782e8dc8c2
|
Add better controls for queue manipulation
|
2023-10-18 14:35:28 +11:00 |
|
|
jacqueline
|
2eb7eaa2a6
|
Go straight to the playback screen after 'play now'
|
2023-10-18 09:41:01 +11:00 |
|
|
jacqueline
|
bb91724b94
|
Codec memory tweaks
- Ensure mad and ogg structs are allocated internally
- Free old codec before making a new one
|
2023-10-16 16:15:48 +11:00 |
|
|
jacqueline
|
7523772886
|
Decouple play/pause from output on/off
I think this was the cause of toggling play/pause making audio go away.
Or at least I can't repro that bug anymore, anyway.
|
2023-10-16 13:52:43 +11:00 |
|
|
jacqueline
|
96ea6cef88
|
Fix incorrect track lengths
|
2023-10-16 13:07:52 +11:00 |
|
|
jacqueline
|
f03093e5c2
|
Remove early-exit when there's no battery
SAMD is currently returning incorrect states, so this is the cause of
the battery meter often reading 0 when not plugged in
|
2023-10-16 08:50:50 +11:00 |
|
|
jacqueline
|
f27670a112
|
acquire spi lock in the opus tag parser
|
2023-10-14 11:59:34 +11:00 |
|
|
jacqueline
|
ac320e376e
|
Merge branch 'main' of git.sr.ht:~jacqueline/tangara-fw
|
2023-10-14 11:47:00 +11:00 |
|
|
jacqueline
|
c2dddba30a
|
Add very basic db versioning
|
2023-10-14 11:46:42 +11:00 |
|
|
jacqueline
|
542e6f5d90
|
The tweaking of max internal alloc size will continue until performance improves
|
2023-10-14 11:30:34 +11:00 |
|
|
jacqueline
|
5a459ac6e9
|
Remove accidental debug thing
|
2023-10-13 20:41:31 +11:00 |
|
|
jacqueline
|
ceba508593
|
Add modified time to TrackData
|
2023-10-13 15:29:06 +11:00 |
|
|
jacqueline
|
afbf3c31f4
|
Use libcppbor for much much nicer db encoding
|
2023-10-13 15:05:49 +11:00 |
|
|
jacqueline
|
20d1c280a7
|
Remove templating of Continuation
|
2023-10-13 11:13:20 +11:00 |
|
|
Robin Howard
|
b284eb4b21
|
hack: "fix" the .env loading by just using pwd instead of dirname $_ etc
|
2023-10-12 14:15:17 +11:00 |
|
|
jacqueline
|
53cf476876
|
Remove mV from top bar, now that we know it's not the issue
|
2023-10-12 11:51:14 +11:00 |
|
|
jacqueline
|
936e327974
|
Claw back a little more memory
|
2023-10-12 11:20:59 +11:00 |
|
|
jacqueline
|
f0d3a27dd9
|
Ensure StaticTask_t allocs are internal
|
2023-10-12 09:02:19 +11:00 |
|
|
jacqueline
|
ddcaa967fd
|
Update battery mV even with the samd doesn't respond
|
2023-10-12 08:42:51 +11:00 |
|
|
jacqueline
|
b0d85fc0d6
|
Skip ID3 tags in the mad decoder; mad itself sometimes chokes on them
|
2023-10-10 13:13:41 +11:00 |
|
|
jacqueline
|
6c20eafd05
|
First pass at having some kind of design and theming
|
2023-10-05 14:24:00 +11:00 |
|
|
jacqueline
|
bf1fc5a2a0
|
Guard audio pipeline file accesses
|
2023-10-05 14:23:48 +11:00 |
|
|
jacqueline
|
ee8e523456
|
Add a basic database reindex screen
|
2023-10-04 15:38:18 +11:00 |
|
|
jacqueline
|
28633e857f
|
Implement UI for enqueuing instead of replacing the current track
|
2023-10-04 14:47:30 +11:00 |
|
|
jacqueline
|
7a7fafdd92
|
Guard leveldb file access with the SPI lock too
Same freeze as was previously ocurring, only this one was a fair bit
rarer.
|
2023-10-04 13:43:43 +11:00 |
|
|
jacqueline
|
5c04e2ad8d
|
Neaten up the various kinds of inputs, and move volumes to The New Way
|
2023-10-04 10:18:31 +11:00 |
|
|
jacqueline
|
c851b789fa
|
Write up the new input modes to NVS and settings
|
2023-10-03 15:00:28 +11:00 |
|
|
jacqueline
|
09f129662e
|
Add scroll velocity + more input methods
|
2023-10-03 14:33:36 +11:00 |
|
|
jacqueline
|
7d5536e2ab
|
Generalise the lvgl input driver in preparation for more input methods
|
2023-10-03 10:37:29 +11:00 |
|
|
jacqueline
|
f2bad894cd
|
Show battery voltage in top bar
|
2023-10-03 09:52:30 +11:00 |
|
|
jacqueline
|
3df4cc9e3d
|
Make accessors in RelativeWheel const
|
2023-10-03 09:36:28 +11:00 |
|
|
jacqueline
|
95618239e6
|
Fix tasks config flag guarding
|
2023-09-29 15:35:00 +10:00 |
|
|
jacqueline
|
ba940baa0a
|
Add a lock around the SPI bus
This seems to have been the cause of recurring deadlocks that have been
difficult to repo.
|
2023-09-29 15:17:32 +10:00 |
|
|
jacqueline
|
c53802f308
|
Add a core dump command
|
2023-09-29 12:34:27 +10:00 |
|
|
jacqueline
|
c876c00654
|
add a fun and cute little loading state :)
|
2023-09-29 10:40:16 +10:00 |
|
|
jacqueline
|
6a47edcd35
|
Use databinding for the top bar. It's so nice now!
|
2023-09-28 10:43:48 +10:00 |
|
|
jacqueline
|
f09ba5ffd5
|
Use bindey for databinding instead of hand rolling ui updates
|
2023-09-28 08:29:55 +10:00 |
|
|
jacqueline
|
f168bfab76
|
Don't post events from ISRs
|
2023-09-27 14:39:19 +10:00 |
|
|
jacqueline
|
ce57b236c4
|
Move every alloc over 256 bytes into PSRAM
|
2023-09-26 21:46:05 +10:00 |
|
|
jacqueline
|
252f685ef1
|
Make NVS access synchronous again
Everything relevant is back in internal ram, and likely to stay there.
|
2023-09-26 21:07:44 +10:00 |
|
|
jacqueline
|
96252973d9
|
Make task tracing tools conditionally compiled
|
2023-09-26 20:47:07 +10:00 |
|
|
jacqueline
|
c36bbe4e7d
|
Claw back some more IRAM from unused components
|
2023-09-26 20:36:17 +10:00 |
|
|
jacqueline
|
abc626c1da
|
Place a2dp stream callback into IRAM
|
2023-09-26 20:36:05 +10:00 |
|
|
jacqueline
|
8608f9367f
|
Tune buffer sizes and locations for I2S
|
2023-09-26 17:23:34 +10:00 |
|
|
jacqueline
|
4d99d22e10
|
std::string -> std::pmr::string in psram
|
2023-09-26 13:36:07 +10:00 |
|
|
jacqueline
|
f6d0642109
|
Prepare for PMR with a memory_resource that understands heap_caps_malloc
|
2023-09-26 11:23:53 +10:00 |
|
|
jacqueline
|
2849399d54
|
Force large c++ arrays into PSRAM
This is mostly targetted at leveldb_ which is rude and does not support
custom allocators
|
2023-09-26 10:18:33 +10:00 |
|
|
jacqueline
|
a87790a424
|
Add basic leak detection tooling
|
2023-09-26 10:14:07 +10:00 |
|
|
jacqueline
|
a2eb0e8961
|
Fix incorrect default theme
|
2023-09-25 22:37:46 +10:00 |
|