Crash reports: replayable, bounded, local, reachable by agents (#384)

Every panic (in a command, in the kernel, anywhere else) and every unclean
exit (abort, stack overflow, kill) leaves a report in a local crash folder:
message, file:line, a trimmed backtrace, version, commit, system and renderer,
the command and recent commands, and when possible the design and replay.json
that reproduce it (solvecraft-cli crashes replay <id>).

- Cost: nothing until the app installs the hook (tests, CLI and web never
  do); then one small journal line per top-level command of the design on
  screen, taken back with a truncate when the command changed nothing.
- Unclean exits: each run holds a lock on its journal; a journal nobody holds
  at the next start belongs to an app that died. Autosaves restart the
  journal from the autosaved design.
- Bounded: a crash's signature (place + message, numbers ignored) is counted,
  not rewritten; at most 5 new reports per run, 20 reports, 50 MB, 90 days;
  recovered kernel panics are pruned first.
- Agents: diagnostics.crash_reports / crash_report / crash_reports.clear
  commands, the MCP crash_reports tool, solvecraft-cli crashes, solvecraft
  --crash-log; diagnostics.crash_test (with SOLVECRAFT_CRASH_TEST) checks it.
- People: the Crash Reports window opens after a crash (Help > Crash
  Reports…) with a reviewed issue draft per report; nothing is sent.
- Release builds keep line tables for file:line in backtraces.

Tests: child processes that really panic in a command, in the kernel, on a
thread, and abort; dedup counts, limits, hostile folders, journal parsing,
the UI window and draft, the MCP tool. docs/crash-reports.md.

Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: bherbruck <[email protected]>

apps/solvecraft-cli/src/crashes.rs

@@ -0,0 +1,103 @@
+//! `solidcraft-cli crashes`: the crash reports the desktop app wrote on this computer
+//! (docs/crash-reports.md): list them, show one, replay one, delete them.
+
+use std::sync::Mutex;
+
+use serde_json::{Value, json};
+use solvecraft_engine::crash;
+
+const USAGE: &str = "usage: solidcraft-cli crashes [list [--json] | show <id> | replay <id> | dir | clear [<id>]]";
+
+pub fn run(args: &[String]) -> Result<(), String> {
+    let dir = crash::default_dir().ok_or("no crash folder on this system")?;
+    let id = || args.get(1).map(String::as_str).ok_or_else(|| USAGE.to_string());
+    match args.first().map(String::as_str) {
+        None | Some("list") => {
+            let reports = crash::list(&dir);
+            if args.iter().any(|a| a == "--json") {
+                println!("{}", super::pretty(&json!({ "dir": dir.to_string_lossy(), "reports": reports })));
+                return Ok(());
+            }
+            println!("crash reports in {}", dir.display());
+            if reports.is_empty() {
+                println!("  (none)");
+            }
+            for r in reports {
+                let times = if r.count > 1 { format!(" ×{}", r.count) } else { String::new() };
+                let cmd = r.command.map(|c| format!(" in {c}")).unwrap_or_default();
+                let replay = if r.replayable { "  [replayable]" } else { "" };
+                println!("  {}  {}{times}{cmd}  v{}{replay}\n      {}", r.id, r.kind, r.version, r.message);
+            }
+            Ok(())
+        }
+        Some("show") => {
+            println!("{}", super::pretty(&crash::read(&dir, id()?)?));
+            Ok(())
+        }
+        Some("replay") => replay(&crash::read(&dir, id()?)?),
+        Some("dir") => {
+            println!("{}", dir.display());
+            Ok(())
+        }
+        Some("clear") => {
+            println!("deleted {} report(s)", crash::clear(&dir, args.get(1).map(String::as_str)));
+            Ok(())
+        }
+        _ => Err(USAGE.to_string()),
+    }
+}
+
+/// Panics seen while replaying: (location, message).
+static PANICS: Mutex<Vec<(String, String)>> = Mutex::new(Vec::new());
+
+/// Run a report's commands on its design and say whether the crash happens again.
+fn replay(report: &Value) -> Result<(), String> {
+    let id = report["id"].as_str().unwrap_or_default();
+    let replay_path = report["replay_path"].as_str().ok_or_else(|| format!("report {id} cannot be replayed (see its notes)"))?;
+    let script = super::read_json(replay_path)?;
+    let commands = script["commands"].as_array().cloned().unwrap_or_default();
+    let mut s = match report["design_path"].as_str() {
+        Some(d) => super::load(d)?,
+        None => solvecraft_engine::Session::default(),
+    };
+    std::panic::set_hook(Box::new(|info| {
+        let loc = info.location().map_or_else(|| "(unknown)".to_string(), |l| format!("{}:{}", l.file(), l.line()));
+        let msg = info.payload_as_str().unwrap_or("(not a text message)").to_string();
+        if let Ok(mut p) = PANICS.lock() {
+            p.push((loc, msg));
+        }
+    }));
+    if report["kind"] == "unclean_exit" {
+        println!("this crash ended the app without a report: if this process ends abruptly, it reproduced");
+    }
+    println!(
+        "replaying {} command(s) on {}",
+        commands.len(),
+        report["design_path"].as_str().map_or_else(|| "a new design".to_string(), |d| d.to_string())
+    );
+    for (i, c) in commands.iter().enumerate() {
+        let cmd = c["command"].as_str().unwrap_or_default();
+        let r = s.execute(cmd, c.get("params").unwrap_or(&Value::Null));
+        match r {
+            Ok(_) => println!("  {:>3}. {cmd}: ok", i + 1),
+            Err(e) => println!("  {:>3}. {cmd}: {e}", i + 1),
+        }
+    }
+    let _ = std::panic::take_hook();
+    let seen = PANICS.lock().map(|p| p.clone()).unwrap_or_default();
+    let want = report["signature"].as_str().unwrap_or_default();
+    let same = seen.iter().any(|(l, m)| format!("{:016x}", crash::signature(l, m)) == want);
+    if same {
+        println!(
+            "reproduced: the same internal error ({}) at {}",
+            report["message"].as_str().unwrap_or_default(),
+            report["location"].as_str().unwrap_or_default()
+        );
+    } else if let Some((l, m)) = seen.first() {
+        println!("an internal error happened, but not the reported one: {m} at {l}");
+        println!("(the report says: {} at {})", report["message"].as_str().unwrap_or_default(), report["location"].as_str().unwrap_or_default());
+    } else {
+        println!("did not reproduce: no internal error (fixed, or it depends on something the replay does not have)");
+    }
+    Ok(())
+}

apps/solvecraft-cli/src/main.rs

@@ -16,6 +16,7 @@
 #![deny(clippy::unwrap_used, clippy::expect_used, clippy::panic, clippy::unimplemented, clippy::todo, clippy::unreachable)]
 #![forbid(unsafe_code)]
 
+mod crashes;
 mod fuzz;
 mod oracle;
 mod recipe;
@@ -42,6 +43,7 @@ const USAGE: &str = "usage:
   solidcraft-cli oracle <case-dir>... [--json]
   solidcraft-cli fuzz [--from N] [--count N] [--jobs J] [--timeout S] [--steps N] [--minimise] [--out DIR] [--report FILE]
   solidcraft-cli mcp [--in design.solvecraft|script.json] [--connect 127.0.0.1:PORT]
+  solidcraft-cli crashes [list [--json] | show <id> | replay <id> | dir | clear [<id>]]
   solidcraft-cli licences
   solidcraft-cli icon --out icon.png [--size 256] | --out icon.ico
 ";
@@ -74,6 +76,7 @@ fn main() -> ExitCode {
         Some("fuzz") => fuzz::run(&args[1..]),
         Some("fuzz-one") => fuzz::one(&args[1..]),
         Some("mcp") => cmd_mcp(&args[1..]),
+        Some("crashes") => crashes::run(&args[1..]),
         Some("icon") => cmd_icon(&args[1..]),
         Some("licences" | "licenses") => {
             // Written, not print!ed: a closed pipe (`| head`) is not a crash.