TUI: library console output corrupts alternate screen display #48520

Activity

1. github-actions

   assigned

   kommander

   on Sep 11, 2026
2. github-actions commented on Sep 11, 2026

   

   github-actionsBot

   on Sep 11, 2026 – with GitHub Actions

   Contributor

   More actions

   This issue might be a duplicate of existing issues. Please check:

   • TUI is overwritten by raw plugin or worker output #43064: "TUI is overwritten by raw plugin or worker output" — very similar: raw stdout/stderr writes from in-process plugins corrupt the TUI alternate screen display
   • Background hook stderr output can be painted into the input box until terminal repaint #31219: "Background hook stderr output can be painted into the input box" — same root symptom of unguarded terminal writes corrupting the TUI frame
   • v18 tui is broken (console logs broke everything) #37455: "v18 tui is broken (console logs broke everything)" — console output breaking the TUI display

   If your issue is distinct from these (e.g. the general-class tracking angle you describe vs. a specific instance), please add a note clarifying how it differs so maintainers can keep them appropriately linked.
3. thomas-bw commented on Sep 11, 2026

   

   thomas-bw

   on Sep 11, 2026

   Author

   More actions

   This issue is distinct from all three. It intentionally tracks the general class of "any console.* output from library dependencies corrupts the TUI's alternate screen", not a specific instance:

   • v18 tui is broken (console logs broke everything) #37455 and MCP schema warnings: non-standard "format" values (uint32/uint64) from schemars pollute TUI #31002 are concrete instances of this class (a user plugin's console.log, Ajv's unknown format "uint64" warning). MCP schema warnings: non-standard "format" values (uint32/uint64) from schemars pollute TUI #31002 was already fixed at the source in fix(mcp): silence Ajv unknown format warnings from MCP tool schemas #48331 (Ajv logger disabled via the MCP SDK's ClientOptions); TUI: library console output corrupts alternate screen display #48520 exists so the TUI is resilient against the next library that emits console output, whichever it is.
   • TUI is overwritten by raw plugin or worker output #43064 is the complementary case, not the same one: it concerns raw process.stdout/stderr.write from in-process plugins/workers and --print-logs output, and explicitly notes that console.* is captured there. This issue covers only the console.* path. TUI is overwritten by raw plugin or worker output #43064 should stay open to track the raw-stdio half of the problem.
   • Background hook stderr output can be painted into the input box until terminal repaint #31219 is a different symptom detail: background hook stderr painted into the input box until the next repaint (cursor-placement/repaint behavior), not console output corrupting the alternate screen.

   This issue is already linked to its fix: #48517 installs a ConsoleGuard in every TUI entrypoint (tui, attach, run) and the server worker, redirecting console.debug/info/log/warn/error into the opencode log file while respecting OPENCODE_LOG_LEVEL. Non-TUI entrypoints keep normal console behavior.
4. Rome-NotBeepBoop commented on Sep 28, 2026

   

   Rome-NotBeepBoop

   on Sep 28, 2026

   More actions

   Still happening for me on macOS with OpenCode 1.18.32 and cua-driver 0.30.2.

   I'm using cua-driver as a normal local stdio MCP with:

   cua-driver mcp

   When OpenCode starts, it prints a ton of these:

   unknown format "uint32" ignored in schema ...
   unknown format "uint64" ignored in schema ...
   

   They end up getting printed directly over the TUI and make the interface pretty hard to use.
   The MCP itself still seems to work fine, so it looks like it's just the warning output messing up the TUI.

   Also checked and cua-driver 0.30.2 is currently the latest version.
   Believe it or not, the image attached below is OpenCode in Ghostty 😭.

   
5. thdxr commented on Oct 11, 2026

   

   thdxr

   on Oct 11, 2026

   Member

   More actions

   Thanks for the report! OpenCode 2.0 is now out and development is focused on V2, so we're closing issues filed against OpenCode 1. Please upgrade (opencode upgrade --major, details: https://anoma.ly/notes/opencode-2-0), and if this is still relevant on 2.0, reopen this issue or open a new one with your 2.0 version.

   — from 𝕺𝖕𝖊𝖓𝕮𝖔𝖉𝖊
6. thdxr

   closed this as not plannedon Oct 11, 2026