Skip to content

BananaPad v0.1.0-preview.2

Pre-release
Pre-release

Choose a tag to compare

@chrissotraidis chrissotraidis released this 03 Sep 00:18
· 5 commits to main since this release
367a3da

BananaPad v0.1.0-preview.2

Developer Preview 2 is a focused iPhone and iPad update for the two issues
reported against Preview 1. It remains a ROM-free unsigned arm64 IPA for iOS
and iPadOS 15 or newer.

Assets

  • BananaPad-v0.1.0-preview.2-unsigned.ipa
  • BananaPad-v0.1.0-preview.2-unsigned.ipa.sha256
  • SHA-256: 2ab9265e0a9eb980c5be85cea829bec98c38a790e022b47361c8c9e60ad7c762
  • Size: 7,479,552 bytes
  • App executable SHA-256: 4849f0d5bcfc1e43ad5ad53452f28c4ef1b0e1b3f0206cc32bd81d98fd912689
  • Product-source SHA-256: e0cf86eca3f0113716b25055d991df31a4b8234181a45270a08a09b60ff8e400

The IPA is unsigned and must be re-signed with the user's own Apple Account.
It does not include Donkey Kong 64, a ROM, extracted assets, generated private
inputs, saves, logs, a provisioning profile, or signing material. BananaPad
does not require JIT.

Fixes

  • Added a default-on touch Z lock. Hold Z continuously for one second to lock
    it; hold it again for one second to unlock it. The locked button remains
    visibly pressed and provides light haptic feedback.
  • Added Hold Z to Lock under the three-dot menu's Settings sheet. Turning
    the setting off restores momentary-only Z behavior.
  • Corrected the inherited audio queue-duration calculation to use the actual
    output device rate. The exercised DK64 path supplies 22.05 kHz audio while
    iOS outputs at 48 kHz; the old mismatch could trigger sample dropping at
    about 45.94 ms of real queued output instead of the intended 100 ms.
  • Added bounded audio diagnostics for the requested/obtained format, queue
    depth, starvation, submission gaps, sample dropping, block boundaries, and
    conversion or queue errors.

Verification

The device app passed the arm64 iPhoneOS, iOS 15, both-device-family, privacy,
ROM/save/private-data, signing-residue, personal-path, system-dependency, and
no-dynamic-code audits. Two independent deterministic package runs were
byte-for-byte identical, and ZIP integrity passed.

An extended Simulator run before the release-number-only rebuild observed 81
two-second audio windows with no sample dropping, no conversion/queue errors,
and no sustained queue starvation. Every window crossed the old effective
45.94 ms threshold, providing direct evidence that the prior calculation could
damage otherwise normal buffering.

Retest requested

The Z-lock gesture and the audio fix still need reporter confirmation on a
physical iPhone. If an audible pop remains, reproduce it and immediately use
••• → Share Diagnostics & Logs… so the new audio telemetry can identify the
next failure mode. Review the report before attaching it, and never share ROMs,
saves, signing files, or other private data.