helix-plus/helix-loader/src/lib.rs

pub mod config;
pub mod grammar;

use anyhow::{anyhow, Result};
use etcetera::base_strategy::{choose_base_strategy, BaseStrategy};
use once_cell::sync::Lazy;
use std::{
    collections::{HashMap, HashSet},
    path::{Path, PathBuf},
};
use toml::Value;

pub const VERSION_AND_GIT_HASH: &str = env!("VERSION_AND_GIT_HASH");

static RUNTIME_DIRS: once_cell::sync::Lazy<Vec<PathBuf>> =
    once_cell::sync::Lazy::new(prioritize_runtime_dirs);

static CONFIG_FILE: once_cell::sync::OnceCell<PathBuf> = once_cell::sync::OnceCell::new();

pub fn initialize_config_file(specified_file: Option<PathBuf>) {
    let config_file = specified_file.unwrap_or_else(|| {
        let config_dir = config_dir();

        if !config_dir.exists() {
            std::fs::create_dir_all(&config_dir).ok();
        }

        config_dir.join("config.toml")
    });

    // We should only initialize this value once.
    CONFIG_FILE.set(config_file).ok();
}

/// A list of runtime directories from highest to lowest priority
///
/// The priority is:
///
/// 1. sibling directory to `CARGO_MANIFEST_DIR` (if environment variable is set)
/// 2. subdirectory of user config directory (always included)
/// 3. `HELIX_RUNTIME` (if environment variable is set)
/// 4. subdirectory of path to helix executable (always included)
///
/// Postcondition: returns at least two paths (they might not exist).
fn prioritize_runtime_dirs() -> Vec<PathBuf> {
    const RT_DIR: &str = "runtime";
    // Adding higher priority first
    let mut rt_dirs = Vec::new();
    if let Ok(dir) = std::env::var("CARGO_MANIFEST_DIR") {
        // this is the directory of the crate being run by cargo, we need the workspace path so we take the parent
        let path = PathBuf::from(dir).parent().unwrap().join(RT_DIR);
        log::debug!("runtime dir: {}", path.to_string_lossy());
        rt_dirs.push(path);
    }

    let conf_rt_dir = config_dir().join(RT_DIR);
    rt_dirs.push(conf_rt_dir);

    if let Ok(dir) = std::env::var("HELIX_RUNTIME") {
        rt_dirs.push(dir.into());
    }

    // fallback to location of the executable being run
    // canonicalize the path in case the executable is symlinked
    let exe_rt_dir = std::env::current_exe()
        .ok()
        .and_then(|path| std::fs::canonicalize(path).ok())
        .and_then(|path| path.parent().map(|path| path.to_path_buf().join(RT_DIR)))
        .unwrap();
    rt_dirs.push(exe_rt_dir);
    rt_dirs
}

/// Runtime directories ordered from highest to lowest priority
///
/// All directories should be checked when looking for files.
///
/// Postcondition: returns at least one path (it might not exist).
pub fn runtime_dirs() -> &'static [PathBuf] {
    &RUNTIME_DIRS
}

/// Find file with path relative to runtime directory
///
/// `rel_path` should be the relative path from within the `runtime/` directory.
/// The valid runtime directories are searched in priority order and the first
/// file found to exist is returned, otherwise None.
fn find_runtime_file(rel_path: &Path) -> Option<PathBuf> {
    RUNTIME_DIRS.iter().find_map(|rt_dir| {
        let path = rt_dir.join(rel_path);
        if path.exists() {
            Some(path)
        } else {
            None
        }
    })
}

/// Find file with path relative to runtime directory
///
/// `rel_path` should be the relative path from within the `runtime/` directory.
/// The valid runtime directories are searched in priority order and the first
/// file found to exist is returned, otherwise the path to the final attempt
/// that failed.
pub fn runtime_file(rel_path: &Path) -> PathBuf {
    find_runtime_file(rel_path).unwrap_or_else(|| {
        RUNTIME_DIRS
            .last()
            .map(|dir| dir.join(rel_path))
            .unwrap_or_default()
    })
}

pub fn config_dir() -> PathBuf {
    // TODO: allow env var override
    let strategy = choose_base_strategy().expect("Unable to find the config directory!");
    let mut path = strategy.config_dir();
    path.push("helix");
    path
}

pub fn cache_dir() -> PathBuf {
    // TODO: allow env var override
    let strategy = choose_base_strategy().expect("Unable to find the config directory!");
    let mut path = strategy.cache_dir();
    path.push("helix");
    path
}

pub fn config_file() -> PathBuf {
    CONFIG_FILE
        .get()
        .map(|path| path.to_path_buf())
        .unwrap_or_else(|| config_dir().join("config.toml"))
}

pub fn workspace_config_file() -> PathBuf {
    find_workspace().0.join(".helix").join("config.toml")
}

pub fn lang_config_file() -> PathBuf {
    config_dir().join("languages.toml")
}

pub fn log_file() -> PathBuf {
    cache_dir().join("helix.log")
}

/// Merge two TOML documents, merging values from `right` onto `left`
///
/// When an array exists in both `left` and `right`, `right`'s array is
/// used. When a table exists in both `left` and `right`, the merged table
/// consists of all keys in `left`'s table unioned with all keys in `right`
/// with the values of `right` being merged recursively onto values of
/// `left`.
///
/// `merge_toplevel_arrays` controls whether a top-level array in the TOML
/// document is merged instead of overridden. This is useful for TOML
/// documents that use a top-level array of values like the `languages.toml`,
/// where one usually wants to override or add to the array instead of
/// replacing it altogether.
pub fn merge_toml_values(left: toml::Value, right: toml::Value, merge_depth: usize) -> toml::Value {
    fn get_name(v: &Value) -> Option<&str> {
        v.get("name").and_then(Value::as_str)
    }

    match (left, right) {
        (Value::Array(mut left_items), Value::Array(right_items)) => {
            // The top-level arrays should be merged but nested arrays should
            // act as overrides. For the `languages.toml` config, this means
            // that you can specify a sub-set of languages in an overriding
            // `languages.toml` but that nested arrays like Language Server
            // arguments are replaced instead of merged.
            if merge_depth > 0 {
                left_items.reserve(right_items.len());
                for rvalue in right_items {
                    let lvalue = get_name(&rvalue)
                        .and_then(|rname| {
                            left_items.iter().position(|v| get_name(v) == Some(rname))
                        })
                        .map(|lpos| left_items.remove(lpos));
                    let mvalue = match lvalue {
                        Some(lvalue) => merge_toml_values(lvalue, rvalue, merge_depth - 1),
                        None => rvalue,
                    };
                    left_items.push(mvalue);
                }
                Value::Array(left_items)
            } else {
                Value::Array(right_items)
            }
        }
        (Value::Table(mut left_map), Value::Table(right_map)) => {
            if merge_depth > 0 {
                for (rname, rvalue) in right_map {
                    match left_map.remove(&rname) {
                        Some(lvalue) => {
                            let merged_value = merge_toml_values(lvalue, rvalue, merge_depth - 1);
                            left_map.insert(rname, merged_value);
                        }
                        None => {
                            left_map.insert(rname, rvalue);
                        }
                    }
                }
                Value::Table(left_map)
            } else {
                Value::Table(right_map)
            }
        }
        // Catch everything else we didn't handle, and use the right value
        (_, value) => value,
    }
}

/// Recursively load a TOML document, merging with any inherited parent files.
///
/// The paths that have been visited in the inheritance hierarchy are tracked
/// to detect and avoid cycling.
///
/// It is possible for one file to inherit from another file with the same name
/// so long as the second file is in a search directory with lower priority.
/// However, it is not recommended that users do this as it will make tracing
/// errors more difficult.
pub fn load_inheritable_toml(
    name: &str,
    search_directories: &[PathBuf],
    visited_paths: &mut HashSet<PathBuf>,
    default_toml_data: &HashMap<&str, &Lazy<Value>>,
    merge_toml_docs: fn(Value, Value) -> Value,
) -> Result<Value> {
    let path = get_toml_path(name, search_directories, visited_paths)?;

    let toml_doc = load_toml(&path)?;

    let inherits = toml_doc.get("inherits");

    let toml_doc = if let Some(parent_toml_name) = inherits {
        let parent_toml_name = parent_toml_name.as_str().ok_or_else(|| {
            anyhow!(
                "{:?}: expected 'inherits' to be a string: {}",
                path,
                parent_toml_name
            )
        })?;

        let parent_toml_doc = match default_toml_data.get(parent_toml_name) {
            Some(p) => (**p).clone(),
            None => load_inheritable_toml(
                parent_toml_name,
                search_directories,
                visited_paths,
                default_toml_data,
                merge_toml_docs,
            )?,
        };

        merge_toml_docs(parent_toml_doc, toml_doc)
    } else {
        toml_doc
    };

    Ok(toml_doc)
}

/// Returns the path to the TOML document with the given name
///
/// Ignores paths already visited and follows directory priority order.
fn get_toml_path(
    name: &str,
    search_directories: &[PathBuf],
    visited_paths: &mut HashSet<PathBuf>,
) -> Result<PathBuf> {
    let filename = format!("{}.toml", name);

    let mut cycle_found = false; // track if there was a path, but it was in a cycle
    search_directories
        .iter()
        .find_map(|dir| {
            let path = dir.join(&filename);
            if !path.exists() {
                None
            } else if visited_paths.contains(&path) {
                // Avoiding cycle, continuing to look in lower priority directories
                cycle_found = true;
                None
            } else {
                visited_paths.insert(path.clone());
                Some(path)
            }
        })
        .ok_or_else(|| {
            if cycle_found {
                anyhow!("Toml: cycle found in inheriting: {}", name)
            } else {
                anyhow!("Toml: file not found for: {}", name)
            }
        })
}

// Loads the TOML data as `toml::Value`
fn load_toml(path: &Path) -> Result<Value> {
    let data = std::fs::read_to_string(path)?;
    let value = toml::from_str(&data)?;

    Ok(value)
}

/// Returns the names of the TOML documents within a directory
pub fn read_toml_names(path: &Path) -> Vec<String> {
    std::fs::read_dir(path)
        .map(|entries| {
            entries
                .filter_map(|entry| {
                    let entry = entry.ok()?;
                    let path = entry.path();
                    (path.extension()? == "toml")
                        .then(|| path.file_stem().unwrap().to_string_lossy().into_owned())
                })
                .collect()
        })
        .unwrap_or_default()
}

#[cfg(test)]
mod merge_toml_tests {
    use std::str;

    use super::merge_toml_values;
    use toml::Value;

    #[test]
    fn language_toml_map_merges() {
        const USER: &str = r#"
        [[language]]
        name = "nix"
        test = "bbb"
        indent = { tab-width = 4, unit = "    ", test = "aaa" }
        "#;

        let base = include_bytes!("../../languages.toml");
        let base = str::from_utf8(base).expect("Couldn't parse built-in languages config");
        let base: Value = toml::from_str(base).expect("Couldn't parse built-in languages config");
        let user: Value = toml::from_str(USER).unwrap();

        let merged = merge_toml_values(base, user, 3);
        let languages = merged.get("language").unwrap().as_array().unwrap();
        let nix = languages
            .iter()
            .find(|v| v.get("name").unwrap().as_str().unwrap() == "nix")
            .unwrap();
        let nix_indent = nix.get("indent").unwrap();

        // We changed tab-width and unit in indent so check them if they are the new values
        assert_eq!(
            nix_indent.get("tab-width").unwrap().as_integer().unwrap(),
            4
        );
        assert_eq!(nix_indent.get("unit").unwrap().as_str().unwrap(), "    ");
        // We added a new keys, so check them
        assert_eq!(nix.get("test").unwrap().as_str().unwrap(), "bbb");
        assert_eq!(nix_indent.get("test").unwrap().as_str().unwrap(), "aaa");
        // We didn't change comment-token so it should be same
        assert_eq!(nix.get("comment-token").unwrap().as_str().unwrap(), "#");
    }

    #[test]
    fn language_toml_nested_array_merges() {
        const USER: &str = r#"
        [[language]]
        name = "typescript"
        language-server = { command = "deno", args = ["lsp"] }
        "#;

        let base = include_bytes!("../../languages.toml");
        let base = str::from_utf8(base).expect("Couldn't parse built-in languages config");
        let base: Value = toml::from_str(base).expect("Couldn't parse built-in languages config");
        let user: Value = toml::from_str(USER).unwrap();

        let merged = merge_toml_values(base, user, 3);
        let languages = merged.get("language").unwrap().as_array().unwrap();
        let ts = languages
            .iter()
            .find(|v| v.get("name").unwrap().as_str().unwrap() == "typescript")
            .unwrap();
        assert_eq!(
            ts.get("language-server")
                .unwrap()
                .get("args")
                .unwrap()
                .as_array()
                .unwrap(),
            &vec![Value::String("lsp".into())]
        )
    }
}

/// Finds the current workspace folder.
/// Used as a ceiling dir for LSP root resolution, the filepicker and potentially as a future filewatching root
///
/// This function starts searching the FS upward from the CWD
/// and returns the first directory that contains either `.git` or `.helix`.
/// If no workspace was found returns (CWD, true).
/// Otherwise (workspace, false) is returned
pub fn find_workspace() -> (PathBuf, bool) {
    let current_dir = std::env::current_dir().expect("unable to determine current directory");
    for ancestor in current_dir.ancestors() {
        if ancestor.join(".git").exists() || ancestor.join(".helix").exists() {
            return (ancestor.to_owned(), false);
        }
    }

    (current_dir, true)
}