From a5058bef51f172c9b4d9d1e7a3653ae2280e72d8 Mon Sep 17 00:00:00 2001
From: Nilstrieb <48135649+Nilstrieb@users.noreply.github.com>
Date: Sun, 7 May 2023 12:13:13 +0200
Subject: [PATCH] qpath parsing

---
 Cargo.lock               |  15 +++--
 elven-forest/Cargo.toml  |   1 +
 elven-forest/src/main.rs |   3 +-
 elven-forest/src/size.rs | 117 ++++++++++++++++++++++++++++++++++++++-
 4 files changed, 129 insertions(+), 7 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index 5174ba8..2503a0d 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -109,6 +109,7 @@ dependencies = [
  "clap",
  "elven-parser",
  "memmap2",
+ "rustc-demangle",
  "tabled",
 ]
 
@@ -319,18 +320,18 @@ dependencies = [
 
 [[package]]
 name = "proc-macro2"
-version = "1.0.51"
+version = "1.0.56"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5d727cae5b39d21da60fa540906919ad737832fe0b1c165da3a34d6548c849d6"
+checksum = "2b63bdb0cd06f1f4dedf69b254734f9b45af66e4a031e42a7480257d9898b435"
 dependencies = [
  "unicode-ident",
 ]
 
 [[package]]
 name = "quote"
-version = "1.0.23"
+version = "1.0.26"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8856d8364d252a14d474036ea1358d63c9e6965c8e5c1885c18f73d70bff9c7b"
+checksum = "4424af4bf778aae2051a77b60283332f386554255d722233d09fbfc7e30da2fc"
 dependencies = [
  "proc-macro2",
 ]
@@ -359,6 +360,12 @@ version = "0.6.28"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "456c603be3e8d448b072f410900c09faf164fbce2d480456f50eea6e25f9c848"
 
+[[package]]
+name = "rustc-demangle"
+version = "0.1.23"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d626bb9dae77e28219937af045c257c28bfd3f69333c512553507f5f9798cb76"
+
 [[package]]
 name = "rustix"
 version = "0.36.8"
diff --git a/elven-forest/Cargo.toml b/elven-forest/Cargo.toml
index 8ead826..2670b38 100644
--- a/elven-forest/Cargo.toml
+++ b/elven-forest/Cargo.toml
@@ -10,4 +10,5 @@ anyhow = "1.0.69"
 clap = { version = "4.1.6", features = ["derive"] }
 elven-parser = { path = "../elven-parser" }
 memmap2 = "0.5.8"
+rustc-demangle = { version = "0.1.23", features = ["std"] }
 tabled = "0.10.0"
diff --git a/elven-forest/src/main.rs b/elven-forest/src/main.rs
index 64604ab..39aa4f0 100644
--- a/elven-forest/src/main.rs
+++ b/elven-forest/src/main.rs
@@ -1,10 +1,9 @@
-mod size;
-
 use std::{
     fmt::Display,
     fs::File,
     path::{Path, PathBuf},
 };
+mod size;
 
 use anyhow::Context;
 use clap::Parser;
diff --git a/elven-forest/src/size.rs b/elven-forest/src/size.rs
index 0943144..4c7fa7d 100644
--- a/elven-forest/src/size.rs
+++ b/elven-forest/src/size.rs
@@ -35,8 +35,123 @@ pub fn analyze_text_bloat(elf: ElfReader<'_>) -> Result<()> {
     symbol_sizes.reverse();
 
     for (sym, size) in symbol_sizes {
-        println!("{size} {sym}");
+        let components =
+            symbol_components(std::str::from_utf8(sym)?).with_context(|| sym.to_string())?;
+
+        println!("{size} {components}");
     }
 
     Ok(())
 }
+
+fn symbol_components(sym: &str) -> Result<String> {
+    let demangled = rustc_demangle::demangle(sym).to_string();
+
+    if demangled.starts_with('<') {
+        let qpath = parse_qpath(&demangled).context("invalid qpath")?;
+
+        // qpath
+        return Ok(demangled);
+    } else {
+        // normal path
+        let components = demangled.split("::").collect::<Vec<_>>();
+        let path = components.join(";");
+        return Ok(path);
+    }
+}
+
+#[derive(Debug, PartialEq)]
+struct QPath<'a> {
+    qself: &'a str,
+    trait_: &'a str,
+    pathy_bit: &'a str,
+}
+
+fn parse_qpath(s: &str) -> Result<QPath<'_>> {
+    let mut chars = s.char_indices().skip(1);
+    let mut angle_brackets = 1u64;
+
+    let mut result = None;
+    let mut as_idx = None;
+
+    while let Some((idx, char)) = chars.next() {
+        match char {
+            '<' => angle_brackets += 1,
+            '>' => {
+                angle_brackets -= 1;
+                if angle_brackets == 0 {
+                    result = Some(idx);
+                    break;
+                }
+            }
+            ' ' => {
+                if angle_brackets == 1 && as_idx == None {
+                    as_idx = Some(idx);
+                }
+            }
+            _ => {}
+        }
+    }
+
+    let q_close_idx = result.with_context(|| {
+        format!("qualified symbol `{s}` does not end qualified part with > properly")
+    })?;
+
+    let as_idx =
+        as_idx.with_context(|| format!("qualified symbol `{s}` does not contain ` as `"))?;
+
+    let q = &s[..q_close_idx];
+    let pathy_bit = &s[q_close_idx + 1..];
+    let pathy_bit = pathy_bit.strip_prefix("::").with_context(|| {
+        format!("path after qualification does not start with `::`: `{pathy_bit}`")
+    })?;
+
+    let qself = &q[1..as_idx];
+    let trait_ = &q[(as_idx + " as ".len())..];
+
+    Ok(QPath {
+        qself,
+        trait_,
+        pathy_bit,
+    })
+}
+
+#[cfg(test)]
+mod tests {
+    use crate::size::QPath;
+
+    use super::{parse_qpath, symbol_components};
+
+    #[test]
+    fn parse_qpaths() {
+        assert_eq!(
+            parse_qpath("<std::path::Components as core::fmt::Debug>::fmt").unwrap(),
+            QPath {
+                qself: "std::path::Components",
+                trait_: "core::fmt::Debug",
+                pathy_bit: "fmt",
+            }
+        );
+
+        assert_eq!(
+            parse_qpath("<<std::path::Components as core::fmt::Debug>::fmt::DebugHelper as core::fmt::Debug>::fmt").unwrap(),
+            QPath {
+                qself: "<std::path::Components as core::fmt::Debug>::fmt::DebugHelper",
+                trait_: "core::fmt::Debug",
+                pathy_bit: "fmt",
+            }
+        );
+    }
+
+    #[test]
+    fn path_debug_helper() {
+        // <<std::path::Components as core::fmt::Debug>::fmt::DebugHelper as core::fmt::Debug>::fmt::hc586615181f69e94
+        let sym = "_ZN106_$LT$$LT$std..path..Iter$u20$as$u20$core..fmt..Debug$GT$..fmt..DebugHelper$u20$as$u20$core..fmt..Debug$GT$3fmt17h4f87ac80fb33df05E";
+        let components = symbol_components(sym).unwrap();
+
+        assert_eq!(
+            components,
+            "std;path;Iter;fmt;DebugHelper;fmt;hc586615181f69e94"
+        )
+    }
+}