use shared stage0 parser from `build_helper`

Signed-off-by: onur-ozkan <work@onurozkan.dev>

7 files changed, 120 insertions, 141 deletions

diff --git a/Cargo.lock b/Cargo.lock
index c034cbae991..aeefa89a608 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -380,10 +380,10 @@ name = "bump-stage0"
 version = "0.1.0"
 dependencies = [
  "anyhow",
+ "build_helper",
  "curl",
  "indexmap",
  "serde",
- "serde_json",
  "toml 0.5.11",
 ]
 
diff --git a/src/bootstrap/src/core/config/config.rs b/src/bootstrap/src/core/config/config.rs
index ed45bc30362..5550b2047f5 100644
--- a/src/bootstrap/src/core/config/config.rs
+++ b/src/bootstrap/src/core/config/config.rs
@@ -181,7 +181,7 @@ pub struct Config {
     pub test_compare_mode: bool,
     pub color: Color,
     pub patch_binaries_for_nix: Option<bool>,
-    pub stage0_metadata: Stage0Metadata,
+    pub stage0_metadata: build_helper::stage0_parser::Stage0,
     pub android_ndk: Option<PathBuf>,
     /// Whether to use the `c` feature of the `compiler_builtins` crate.
     pub optimized_compiler_builtins: bool,
@@ -350,34 +350,6 @@ pub struct Config {
     pub paths: Vec<PathBuf>,
 }
 
-#[derive(Default, Deserialize, Clone)]
-pub struct Stage0Metadata {
-    pub compiler: CompilerMetadata,
-    pub config: Stage0Config,
-    pub checksums_sha256: HashMap<String, String>,
-    pub rustfmt: Option<RustfmtMetadata>,
-}
-#[derive(Default, Deserialize, Clone)]
-pub struct CompilerMetadata {
-    pub date: String,
-    pub version: String,
-}
-
-#[derive(Default, Deserialize, Clone)]
-pub struct Stage0Config {
-    pub dist_server: String,
-    pub artifacts_server: String,
-    pub artifacts_with_llvm_assertions_server: String,
-    pub git_merge_commit_email: String,
-    pub git_repository: String,
-    pub nightly_branch: String,
-}
-#[derive(Default, Deserialize, Clone)]
-pub struct RustfmtMetadata {
-    pub date: String,
-    pub version: String,
-}
-
 #[derive(Clone, Debug, Default)]
 pub enum RustfmtState {
     SystemToolchain(PathBuf),
@@ -1296,13 +1268,13 @@ impl Config {
                 Some(p) => PathBuf::from(p),
                 None => git_root,
             };
-            // If this doesn't have at least `stage0.json`, we guessed wrong. This can happen when,
+            // If this doesn't have at least `stage0`, we guessed wrong. This can happen when,
             // for example, the build directory is inside of another unrelated git directory.
             // In that case keep the original `CARGO_MANIFEST_DIR` handling.
             //
             // NOTE: this implies that downloadable bootstrap isn't supported when the build directory is outside
             // the source directory. We could fix that by setting a variable from all three of python, ./x, and x.ps1.
-            if git_root.join("src").join("stage0.json").exists() {
+            if git_root.join("src").join("stage0").exists() {
                 config.src = git_root;
             }
         } else {
@@ -1320,9 +1292,7 @@ impl Config {
             .to_path_buf();
         }
 
-        let stage0_json = t!(std::fs::read(config.src.join("src").join("stage0.json")));
-
-        config.stage0_metadata = t!(serde_json::from_slice::<Stage0Metadata>(&stage0_json));
+        config.stage0_metadata = build_helper::stage0_parser::parse_stage0_file();
 
         // Read from `--config`, then `RUST_BOOTSTRAP_CONFIG`, then `./config.toml`, then `config.toml` in the root directory.
         let toml_path = flags
diff --git a/src/bootstrap/src/core/download.rs b/src/bootstrap/src/core/download.rs
index 75e0f646da6..a074d53aa36 100644
--- a/src/bootstrap/src/core/download.rs
+++ b/src/bootstrap/src/core/download.rs
@@ -9,9 +9,9 @@ use std::{
 };
 
 use build_helper::ci::CiEnv;
+use build_helper::stage0_parser::VersionMetadata;
 use xz2::bufread::XzDecoder;
 
-use crate::core::config::RustfmtMetadata;
 use crate::utils::helpers::{check_run, exe, program_out_of_date};
 use crate::{core::build_steps::llvm::detect_llvm_sha, utils::helpers::hex_encode};
 use crate::{t, Config};
@@ -408,7 +408,7 @@ impl Config {
     /// NOTE: rustfmt is a completely different toolchain than the bootstrap compiler, so it can't
     /// reuse target directories or artifacts
     pub(crate) fn maybe_download_rustfmt(&self) -> Option<PathBuf> {
-        let RustfmtMetadata { date, version } = self.stage0_metadata.rustfmt.as_ref()?;
+        let VersionMetadata { date, version } = self.stage0_metadata.rustfmt.as_ref()?;
         let channel = format!("{version}-{date}");
 
         let host = self.build;
@@ -606,7 +606,7 @@ impl Config {
             DownloadSource::Dist => {
                 let dist_server = env::var("RUSTUP_DIST_SERVER")
                     .unwrap_or(self.stage0_metadata.config.dist_server.to_string());
-                // NOTE: make `dist` part of the URL because that's how it's stored in src/stage0.json
+                // NOTE: make `dist` part of the URL because that's how it's stored in src/stage0
                 (dist_server, format!("dist/{key}/{filename}"), true)
             }
         };
@@ -616,7 +616,7 @@ impl Config {
         // this on each and every nightly ...
         let checksum = if should_verify {
             let error = format!(
-                "src/stage0.json doesn't contain a checksum for {url}. \
+                "src/stage0 doesn't contain a checksum for {url}. \
                 Pre-built artifacts might not be available for this \
                 target at this time, see https://doc.rust-lang.org/nightly\
                 /rustc/platform-support.html for more information."
diff --git a/src/tools/build_helper/src/lib.rs b/src/tools/build_helper/src/lib.rs
index 575f3677155..6a4e86eb1df 100644
--- a/src/tools/build_helper/src/lib.rs
+++ b/src/tools/build_helper/src/lib.rs
@@ -2,3 +2,4 @@ pub mod ci;
 pub mod git;
 pub mod metrics;
 pub mod util;
+pub mod stage0_parser;
diff --git a/src/tools/build_helper/src/stage0_parser.rs b/src/tools/build_helper/src/stage0_parser.rs
new file mode 100644
index 00000000000..ff05b116989
--- /dev/null
+++ b/src/tools/build_helper/src/stage0_parser.rs
@@ -0,0 +1,76 @@
+use std::collections::BTreeMap;
+
+#[derive(Default, Clone)]
+pub struct Stage0 {
+    pub compiler: VersionMetadata,
+    pub rustfmt: Option<VersionMetadata>,
+    pub config: Stage0Config,
+    pub checksums_sha256: BTreeMap<String, String>,
+}
+
+#[derive(Default, Clone)]
+pub struct VersionMetadata {
+    pub date: String,
+    pub version: String,
+}
+
+#[derive(Default, Clone)]
+pub struct Stage0Config {
+    pub dist_server: String,
+    pub artifacts_server: String,
+    pub artifacts_with_llvm_assertions_server: String,
+    pub git_merge_commit_email: String,
+    pub git_repository: String,
+    pub nightly_branch: String,
+}
+
+pub fn parse_stage0_file() -> Stage0 {
+    let stage0_content = include_str!("../../../stage0");
+
+    let mut stage0 = Stage0::default();
+    for line in stage0_content.lines() {
+        let line = line.trim();
+
+        if line.is_empty() {
+            continue;
+        }
+
+        // Ignore comments
+        if line.starts_with('#') {
+            continue;
+        }
+
+        let (key, value) = line.split_once('=').unwrap();
+
+        match key {
+            "dist_server" => stage0.config.dist_server = value.to_owned(),
+            "artifacts_server" => stage0.config.artifacts_server = value.to_owned(),
+            "artifacts_with_llvm_assertions_server" => {
+                stage0.config.artifacts_with_llvm_assertions_server = value.to_owned()
+            }
+            "git_merge_commit_email" => stage0.config.git_merge_commit_email = value.to_owned(),
+            "git_repository" => stage0.config.git_repository = value.to_owned(),
+            "nightly_branch" => stage0.config.nightly_branch = value.to_owned(),
+
+            "compiler_date" => stage0.compiler.date = value.to_owned(),
+            "compiler_version" => stage0.compiler.version = value.to_owned(),
+
+            "rustfmt_date" => {
+                stage0.rustfmt.get_or_insert(VersionMetadata::default()).date = value.to_owned();
+            }
+            "rustfmt_version" => {
+                stage0.rustfmt.get_or_insert(VersionMetadata::default()).version = value.to_owned();
+            }
+
+            dist if dist.starts_with("dist") => {
+                stage0.checksums_sha256.insert(key.to_owned(), value.to_owned());
+            }
+
+            unsupported => {
+                println!("'{unsupported}' field is not supported.");
+            }
+        }
+    }
+
+    stage0
+}
diff --git a/src/tools/bump-stage0/Cargo.toml b/src/tools/bump-stage0/Cargo.toml
index 4680a7ab661..de5d821133d 100644
--- a/src/tools/bump-stage0/Cargo.toml
+++ b/src/tools/bump-stage0/Cargo.toml
@@ -7,8 +7,8 @@ edition = "2021"
 
 [dependencies]
 anyhow = "1.0.34"
+build_helper = { path = "../build_helper" }
 curl = "0.4.38"
 indexmap = { version = "2.0.0", features = ["serde"] }
 serde = { version = "1.0.125", features = ["derive"] }
-serde_json = { version = "1.0.59", features = ["preserve_order"] }
 toml = "0.5.7"
diff --git a/src/tools/bump-stage0/src/main.rs b/src/tools/bump-stage0/src/main.rs
index 1999859ff98..f9c1699eb4c 100644
--- a/src/tools/bump-stage0/src/main.rs
+++ b/src/tools/bump-stage0/src/main.rs
@@ -1,4 +1,7 @@
+#![deny(unused_variables)]
+
 use anyhow::{Context, Error};
+use build_helper::stage0_parser::{parse_stage0_file, Stage0Config, VersionMetadata};
 use curl::easy::Easy;
 use indexmap::IndexMap;
 use std::collections::HashMap;
@@ -8,7 +11,7 @@ const COMPILER_COMPONENTS: &[&str] = &["rustc", "rust-std", "cargo", "clippy-pre
 const RUSTFMT_COMPONENTS: &[&str] = &["rustfmt-preview", "rustc"];
 
 struct Tool {
-    config: Config,
+    config: Stage0Config,
 
     channel: Channel,
     date: Option<String>,
@@ -34,73 +37,9 @@ impl Tool {
             .try_into()
             .map_err(|_| anyhow::anyhow!("failed to parse version"))?;
 
-        // let existing: Stage0 = serde_json::from_slice(&std::fs::read(PATH)?)?;
-        let existing = Self::parse_stage0_file()?;
+        let existing = parse_stage0_file();
 
-        Ok(Self {
-            channel,
-            version,
-            date,
-            config: existing.config,
-            checksums: IndexMap::new(),
-        })
-    }
-
-    fn parse_stage0_file() -> Result<Stage0, Error> {
-        let stage0_content = include_str!("../../../stage0");
-
-        let mut stage0 = Stage0::default();
-
-        for line in stage0_content.lines() {
-            let line = line.trim();
-
-            if line.is_empty() {
-                continue;
-            }
-
-            // Ignore comments
-            if line.starts_with('#') {
-                continue;
-            }
-
-            let key_value: Vec<&str> = line.splitn(2, '=').collect();
-            let (key, value) = (*key_value.get(0).unwrap(), *key_value.get(1).unwrap());
-
-            match key {
-                "dist_server" => stage0.config.dist_server = value.to_owned(),
-                "artifacts_server"
-                | "artifacts_with_llvm_assertions_server"
-                | "git_merge_commit_email"
-                | "git_repository"
-                | "nightly_branch" => {
-                    stage0.config.other.insert(key.to_owned(), value.to_owned());
-                }
-
-                "compiler_date" => stage0.compiler.date = value.to_owned(),
-                "compiler_version" => stage0.compiler.version = value.to_owned(),
-
-                "rustfmt_date" => {
-                    let mut rustfmt = stage0.rustfmt.unwrap_or(Stage0Toolchain::default());
-                    rustfmt.date = value.to_owned();
-                    stage0.rustfmt = Some(rustfmt);
-                }
-                "rustfmt_version" => {
-                    let mut rustfmt = stage0.rustfmt.unwrap_or(Stage0Toolchain::default());
-                    rustfmt.version = value.to_owned();
-                    stage0.rustfmt = Some(rustfmt);
-                }
-
-                dist if dist.starts_with("dist") => {
-                    stage0.checksums_sha256.insert(key.to_owned(), value.to_owned());
-                }
-
-                unsupported => {
-                    println!("'{unsupported}' field is not supported.");
-                }
-            }
-        }
-
-        Ok(stage0)
+        Ok(Self { channel, version, date, config: existing.config, checksums: IndexMap::new() })
     }
 
     fn update_stage0_file(mut self) -> Result<(), Error> {
@@ -115,11 +54,25 @@ impl Tool {
             "#;
 
         let mut file_content = HEADER.to_owned();
-        file_content.push_str(&format!("\ndist_server={}", self.config.dist_server));
 
-        for (key, value) in &self.config.other {
-            file_content.push_str(&format!("\n{}={}", key, value.as_str()));
-        }
+        let Stage0Config {
+            dist_server,
+            artifacts_server,
+            artifacts_with_llvm_assertions_server,
+            git_merge_commit_email,
+            git_repository,
+            nightly_branch,
+        } = &self.config;
+
+        file_content.push_str(&format!("\ndist_server={}", dist_server));
+        file_content.push_str(&format!("\nartifacts_server={}", artifacts_server));
+        file_content.push_str(&format!(
+            "\nartifacts_with_llvm_assertions_server={}",
+            artifacts_with_llvm_assertions_server
+        ));
+        file_content.push_str(&format!("\ngit_merge_commit_email={}", git_merge_commit_email));
+        file_content.push_str(&format!("\ngit_repository={}", git_repository));
+        file_content.push_str(&format!("\nnightly_branch={}", nightly_branch));
 
         file_content.push_str("\n");
 
@@ -149,7 +102,7 @@ impl Tool {
     // On the master branch the compiler version is configured to `beta` whereas if you're looking
     // at the beta or stable channel you'll likely see `1.x.0` as the version, with the previous
     // release's version number.
-    fn detect_compiler(&mut self) -> Result<Stage0Toolchain, Error> {
+    fn detect_compiler(&mut self) -> Result<VersionMetadata, Error> {
         let channel = match self.channel {
             Channel::Stable | Channel::Beta => {
                 // The 1.XX manifest points to the latest point release of that minor release.
@@ -160,7 +113,7 @@ impl Tool {
 
         let manifest = fetch_manifest(&self.config, &channel, self.date.as_deref())?;
         self.collect_checksums(&manifest, COMPILER_COMPONENTS)?;
-        Ok(Stage0Toolchain {
+        Ok(VersionMetadata {
             date: manifest.date,
             version: if self.channel == Channel::Nightly {
                 "beta".to_string()
@@ -179,14 +132,14 @@ impl Tool {
     /// We use a nightly rustfmt to format the source because it solves some bootstrapping issues
     /// with use of new syntax in this repo. For the beta/stable channels rustfmt is not provided,
     /// as we don't want to depend on rustfmt from nightly there.
-    fn detect_rustfmt(&mut self) -> Result<Option<Stage0Toolchain>, Error> {
+    fn detect_rustfmt(&mut self) -> Result<Option<VersionMetadata>, Error> {
         if self.channel != Channel::Nightly {
             return Ok(None);
         }
 
         let manifest = fetch_manifest(&self.config, "nightly", self.date.as_deref())?;
         self.collect_checksums(&manifest, RUSTFMT_COMPONENTS)?;
-        Ok(Some(Stage0Toolchain { date: manifest.date, version: "nightly".into() }))
+        Ok(Some(VersionMetadata { date: manifest.date, version: "nightly".into() }))
     }
 
     fn collect_checksums(&mut self, manifest: &Manifest, components: &[&str]) -> Result<(), Error> {
@@ -220,7 +173,11 @@ fn main() -> Result<(), Error> {
     Ok(())
 }
 
-fn fetch_manifest(config: &Config, channel: &str, date: Option<&str>) -> Result<Manifest, Error> {
+fn fetch_manifest(
+    config: &Stage0Config,
+    channel: &str,
+    date: Option<&str>,
+) -> Result<Manifest, Error> {
     let url = if let Some(date) = date {
         format!("{}/dist/{}/channel-rust-{}.toml", config.dist_server, date, channel)
     } else {
@@ -253,31 +210,6 @@ enum Channel {
     Nightly,
 }
 
-#[derive(Debug, Default, serde::Serialize, serde::Deserialize)]
-struct Stage0 {
-    config: Config,
-    compiler: Stage0Toolchain,
-    rustfmt: Option<Stage0Toolchain>,
-    checksums_sha256: IndexMap<String, String>,
-}
-
-#[derive(Debug, Default, serde::Serialize, serde::Deserialize)]
-struct Config {
-    dist_server: String,
-    /// There are other fields in the configuration, which will be read by src/bootstrap or other
-    /// tools consuming stage0 file. To avoid the need to update bump-stage0 every time a new field
-    /// is added, we collect all the fields in `IndexMap<String, String>` and serialize them back with the
-    /// same order and structure they were deserialized in.
-    #[serde(flatten)]
-    other: IndexMap<String, String>,
-}
-
-#[derive(Debug, Default, serde::Serialize, serde::Deserialize)]
-struct Stage0Toolchain {
-    date: String,
-    version: String,
-}
-
 #[derive(Debug, serde::Serialize, serde::Deserialize)]
 struct Manifest {
     date: String,