Skip to content

Settings: test switches for a bug hunt; D-Modem's V.34 precoder switches - #30

Draft
Neo-Desktop wants to merge 10 commits into
mainfrom
claude/test-switches
Draft

Neo-Desktop wants to merge 10 commits into
mainfrom
claude/test-switches

Conversation

@Neo-Desktop

Copy link
Copy Markdown
Collaborator

Requested by Sydney · project thread

A draft for a field test of SmartUSB56 dial-outs to a USR V.92 modem. The USR asks us to precode, then never decodes our data. This PR's CI build is the APK for that test.

  • Settings › Line (USB): a "Test switches" field. It takes NAME=VALUE pairs, limited to D-Modem's own DSP_* and DMODEM_* names. They go into the environment at the modem's next start and are logged as test switches: .... Switches from the previous start are unset.
  • D-Modem bumped to a6b8703 (branch claude/v34-precoding-isi):
    • DSP_V34_TX_PRECODE=0 and DSP_V34_TX_NL=0 make the transmitter leave out the precoder or the non-linear encoder that the remote's MP asks for.
    • netsim gains NET_ISI, a filter on A's line, so a receiver asks for precoding.
  • Unit test: testSwitchesAreTheModemsOwnSettingsOnly.

Not built locally, because this container can't reach the Android SDK; CI builds it. D-Modem's quick ctest suite passes on the bumped commit.


Generated by Claude Code

… bug hunt; D-Modem at a6b8703 (V.34 test switches for the precoder)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01B82eRRWkXKKQdZ1QTqN43P
@Neo-Desktop Neo-Desktop self-assigned this Nov 10, 2026
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01B82eRRWkXKKQdZ1QTqN43P
…g typed in scrolls into view

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01B82eRRWkXKKQdZ1QTqN43P

Copy link
Copy Markdown
Collaborator Author

The Android build fails in PppTakeoverTest.pppOnTheCallThereIs (no answer to the UDP echo within 20 s, line 99), and then needsACall fails because the first test left the link up. A re-run failed the same way.

This PR doesn't touch PPP. Since the last green run, only SettingsScreen.kt has changed (one imePadding() modifier), and the earlier run of this branch, with the same settings and service changes, passed that test in 6 s. I can't run the Robolectric tests in my environment, so I haven't found the cause yet. I'm keeping the PR watched.


Generated by Claude Code

…rom the connect on)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01B82eRRWkXKKQdZ1QTqN43P
… (the line at 8000 Hz)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01B82eRRWkXKKQdZ1QTqN43P
…; no S after J fails in 4 s

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01B82eRRWkXKKQdZ1QTqN43P

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants