Write initial code for parsing stats into chunks

Cargo.lock

@@ -14,12 +14,28 @@ version = "0.1.0"
 dependencies = [
  "anyhow",
  "codepage-437",
+ "compact_str",
  "insta",
  "nom",
  "peg_macro",
  "pest",
 ]
 
+[[package]]
+name = "castaway"
+version = "0.2.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0abae9be0aaf9ea96a3b1b8b1b55c602ca751eba1b1500220cea4ecbafe7c0d5"
+dependencies = [
+ "rustversion",
+]
+
+[[package]]
+name = "cfg-if"
+version = "1.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9555578bc9e57714c812a1f84e4fc5b4d21fcb063490c624de019f7464c91268"
+
 [[package]]
 name = "codepage-437"
 version = "0.1.0"
@@ -29,6 +45,20 @@ dependencies = [
  "csv",
 ]
 
+[[package]]
+name = "compact_str"
+version = "0.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3fdb1325a1cece981e8a296ab8f0f9b63ae357bd0784a9faaf548cc7b480707a"
+dependencies = [
+ "castaway",
+ "cfg-if",
+ "itoa",
+ "rustversion",
+ "ryu",
+ "static_assertions",
+]
+
 [[package]]
 name = "console"
 version = "0.15.11"
@@ -201,6 +231,12 @@ dependencies = [
  "proc-macro2",
 ]
 
+[[package]]
+name = "rustversion"
+version = "1.0.21"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8a0d197bd2c9dc6e53b84da9556a69ba4cdfab8619eb41a8bd1cc2027a0f6b1d"
+
 [[package]]
 name = "ryu"
 version = "1.0.20"
@@ -233,6 +269,12 @@ version = "2.7.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "bbbb5d9659141646ae647b42fe094daf6c6192d1620870b449d9557f748b2daa"
 
+[[package]]
+name = "static_assertions"
+version = "1.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a2eb9349b6444b326872e140eb1cf5e7c522154d69e7a0ffb0fb81c06b37543f"
+
 [[package]]
 name = "syn"
 version = "2.0.101"

Cargo.toml

@@ -12,6 +12,7 @@ peg_macro = { path = "./peg_macro"}
 insta = "1.39.0"
 nom = "7.1.3"
 pest = "2.7.15"
+compact_str = "0.9.0"
 
 [dev-dependencies]
 insta = "1.39.0"

src/labels/mod.rs

@@ -1 +1 @@
-pub mod labels;
+pub mod parse;

src/labels/parse.rs

@@ -1,4 +1,85 @@
-use crate::{peg::ParseState, world::Stat};
+use std::ops::Range;
+
+use compact_str::CompactString;
+use grammar::Tag;
+
+use crate::{
+    peg::{Capture, ParseState},
+    world::Stat,
+};
+
+type ParsedStat = Vec<Chunk>;
+enum Chunk {
+    Verbatim(String),
+    Label(LabelName),
+    Reference(LabelName),
+}
+
+struct LabelName {
+    namespace: CompactString,
+    name: CompactString,
+    local: Option<CompactString>,
+}
+
+fn parse_stat_labels(stat: &Stat) -> ParsedStat {
+    let code = &stat.code;
+    let mut parser = ParseState::new(code);
+    assert!(
+        grammar::program(&mut parser),
+        "Couldn't parse code: {:?}",
+        code
+    );
+
+    // Split code along capture group boundaries
+    let mut result = vec![];
+    let mut offset = 0;
+
+    let mut push_label = |cap: Capture<_>| {
+        let Range { start, end } = cap.span();
+        if offset < start {
+            result.push(Chunk::Verbatim(code[offset..start].into()))
+        }
+        let mut label = LabelName {
+            namespace: CompactString::with_capacity(0),
+            name: CompactString::with_capacity(0),
+            local: None,
+        };
+        for child in cap.children() {
+            match child.kind() {
+                Tag::Namespace => {
+                    label.namespace = child.text().into();
+                }
+                Tag::Global => {
+                    label.name = child.text().into();
+                }
+                Tag::Local => label.local = Some(child.text().into()),
+                _ => unimplemented!(),
+            }
+        }
+        result.push(Chunk::Label(label));
+        offset = end;
+    };
+
+    for cap in parser.captures() {
+        match cap.kind() {
+            Tag::Label => {
+                push_label(cap);
+            }
+            Tag::Reference => {
+                for child in cap.children() {
+                    if child.kind() == Tag::Label {
+                        push_label(child);
+                    }
+                }
+            }
+            _ => {}
+        }
+    }
+    if offset < code.len() {
+        result.push(Chunk::Verbatim(code[offset..code.len()].into()));
+    }
+    result
+}
 
 pub fn print_labels(b: &Stat) {
     let code = &b.code;
@@ -10,7 +91,7 @@ pub fn print_labels(b: &Stat) {
     );
 
     for capture in parser.captures() {
-        if capture.kind() == grammar::Tag::Label {
+        if capture.kind() == Tag::Label {
             println!("- {}", capture.text());
         }
     }
@@ -106,7 +187,7 @@ mod grammar {
             "n" / "s" / "e" / "w" / "i"                     // short forms
         ) eow;
 
-        // Labels (defined locations in the code)
+        // Labels
         // Examples: foo, namespace~foo, foo.local, .local, @
         label = #Label:(namespace? (label_name / #Anon:"@"));
         namespace = #Namespace:label_word "~";
@@ -115,9 +196,9 @@ mod grammar {
         label_local = "." #Local:label_word;
         label_word = word_char+; // labels can start with 0-9
 
-        // Messages (references to labels)
+        // References to labels
         // Examples: foo, all:namespace~bar.baz, @b, @f
-        message = #Message:(recipient? message_name);
+        message = #Reference:(recipient? #Label:message_name);
         recipient = #Recipient:word ":";
         message_name = namespace? (label_name / #Anon:anon_message);
         anon_message = "@" ("b" / "f");
@@ -216,7 +297,7 @@ mod test {
             let before = &input[last_index..group.span().start];
             let inner = match group.kind() {
                 Tag::Label => format!("({})", group.text()),
-                Tag::Message => format!("[{}]", group.text()),
+                Tag::Reference => format!("[{}]", group.text()),
                 _ => group.text().into(),
             };
             result.push_str(before);

src/labels/snapshots/zasm__labels__parse__test__label_detection.snap

@@ -1,5 +1,5 @@
 ---
-source: src/labels/labels.rs
+source: src/labels/parse.rs
 expression: result
 ---
 '

src/main.rs

@@ -5,7 +5,7 @@ mod preprocess;
 mod world;
 
 use anyhow::anyhow;
-use labels::labels::print_labels;
+use labels::parse::print_labels;
 use preprocess::eval::Context;
 use std::{env, error::Error, fs, path::PathBuf, process::exit};
 use world::World;