profile_xpath_slow_unit/
profile_xpath_slow_unit.rs

1//! Profile / reduce the XPath eval slow-unit captured by the fuzz
2//! corpus.  Times the original artifact + a battery of reductions
3//! to isolate which sub-pattern dominates the step budget.
4//!
5//! Usage:
6//!     cargo run --release --example profile_xpath_slow_unit -- \
7//!         crates/core/fuzz/artifacts/fuzz_xpath_eval/slow-unit-<hash>
8
9use std::time::Instant;
10use sup_xml_core::{parse_str, ParseOptions};
11use sup_xml_core::xpath::XPathContext;
12
13const FIXTURE: &str = r#"<?xml version="1.0" encoding="UTF-8"?>
14<catalog xmlns:ex="urn:example" xml:lang="en">
15  <book id="b1" price="9.99">
16    <title>The Pragmatic Programmer</title>
17    <author>Hunt</author>
18    <author>Thomas</author>
19    <year>1999</year>
20    <tags><tag>classic</tag><tag>craft</tag></tags>
21  </book>
22  <book id="b2" price="42">
23    <title lang="en">Compilers</title>
24    <author>Aho</author>
25    <author>Sethi</author>
26    <author>Ullman</author>
27    <year>2006</year>
28    <ex:rating>5</ex:rating>
29    <!-- a comment node for comment() tests -->
30    <?xml-stylesheet href="x.xsl"?>
31    <desc><![CDATA[<not really xml>]]></desc>
32  </book>
33  <book id="b3" price="0">
34    <title/>
35    <year>-1</year>
36    <empty></empty>
37    <nested><a><b><c><d>deep</d></c></b></a></nested>
38  </book>
39  <unicode>café αβγ 中文 𝛼</unicode>
40</catalog>"#;
41
42fn time_one(ctx: &XPathContext, label: &str, expr: &str) {
43    let mut best = std::time::Duration::MAX;
44    let mut last_result = String::new();
45    for _ in 0..3 {
46        let t = Instant::now();
47        let r = ctx.eval(expr);
48        let dt = t.elapsed();
49        if dt < best { best = dt; }
50        last_result = match &r {
51            Ok(_)  => "Ok".to_string(),
52            Err(e) => format!("Err: {}", e.message.chars().take(60).collect::<String>()),
53        };
54    }
55    println!("{:>9.2?}  [{:>4} B]  {label}   →  {last_result}",
56             best, expr.len());
57}
58
59fn main() {
60    let path = std::env::args().nth(1)
61        .expect("usage: profile_xpath_slow_unit <artifact-path>");
62    let bytes = std::fs::read(&path).expect("read artifact");
63    let full = std::str::from_utf8(&bytes).expect("artifact is UTF-8");
64
65    let doc = parse_str(FIXTURE, &ParseOptions::default()).expect("fixture parses");
66    let ctx = XPathContext::new(&doc);
67
68    // Report the predicate-nesting depth of the original artifact so we
69    // can pick a sensible MAX_PREDICATE_NESTING_DEPTH.
70    use sup_xml_core::xpath::ast::max_predicate_nesting;
71    use sup_xml_core::xpath::parse_xpath;
72    match parse_xpath(full) {
73        Ok(e)  => println!("# original artifact parses; predicate-nesting depth = {}", max_predicate_nesting(&e)),
74        Err(e) => println!("# original artifact parse-error: {}",
75                           e.message.chars().take(80).collect::<String>()),
76    }
77
78    println!("# best-of-3 timing of slow-unit reductions");
79    println!("{:>9}  {:>8}  {}", "time", "len", "expression  →  result");
80    println!("{}", "-".repeat(78));
81
82    // ── baseline ──────────────────────────────────────────────────────────
83    time_one(&ctx, "FULL artifact (1018 B)", full);
84
85    // ── strip from the right, halving each time ──────────────────────────
86    time_one(&ctx, "first 512 B",  &full[..full.len().min(512)]);
87    time_one(&ctx, "first 256 B",  &full[..full.len().min(256)]);
88    time_one(&ctx, "first 128 B",  &full[..full.len().min(128)]);
89    time_one(&ctx, "first  64 B",  &full[..full.len().min(64)]);
90
91    // ── isolate specific patterns ────────────────────────────────────────
92    time_one(&ctx, "just nested substring chains",
93        "r*substring(/*substring(/*substring(/*substring(/*substring(/*substring(/*substring(//book[author='Hu'])))))))");
94    time_one(&ctx, "deep //*/*/* descent chain",
95        "//*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*");
96    time_one(&ctx, "nested predicates (the worst case)",
97        "//*[//*[//*[//*[//*[//*[//*[.='x']]]]]]]");
98    time_one(&ctx, "nested //*/* in predicate",
99        "//*[/*//*//*//*//*//*//*//*//*//*]");
100    time_one(&ctx, "wildcard arithmetic chain",
101        "//*/*/*/* * * * * * * * * * //*/*/*");
102    time_one(&ctx, "many top-level *",
103        "* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *");
104
105    // ── single-step costs for comparison ─────────────────────────────────
106    time_one(&ctx, "trivial root",                                 "/");
107    time_one(&ctx, "trivial wildcard",                             "/*");
108    time_one(&ctx, "all elements",                                 "//*");
109    time_one(&ctx, "all elements with trivial predicate",          "//*[true()]");
110    time_one(&ctx, "all elements with attribute access",           "//*[@id]");
111    time_one(&ctx, "books filtered by author",                     "//book[author='Hunt']");
112    time_one(&ctx, "substring on string literal",                  "substring('hello', 2, 3)");
113    time_one(&ctx, "substring on node text",                       "substring(//title, 2, 3)");
114}
profile_xpath_slow_unit/profile_xpath_slow_unit.rs

profile_xpath_slow_unit/
profile_xpath_slow_unit.rs
