1
0
Fork 0
mirror of https://github.com/denoland/deno.git synced 2024-12-14 19:37:59 -05:00
denoland-deno/cli/npm/managed/resolvers/common/lifecycle_scripts.rs
Nathan Whitaker c371b2a492
fix(install): re-setup bin entries after running lifecycle scripts (#26752)
Fixes #26677

Some packages (like supabase) declare bin entries that don't exist until
lifecycle scripts are run. For instance, the lifecycle script downloads
a binary file which serves as a bin entrypoint.

Unfortunately you can't just defer setting up the bin entries until
after lifecycle scripts have run, because the scripts may rely on them.

I looked into this, and PNPM just re-links bin entries after running
lifecycle scripts. I think that's about the best we can do as well.

Note that we'll only re-setup bin entries for packages whose lifecycle
scripts we run. This should limit the performance cost, as typically a
given project will not have many lifecycle scripts (and of those, many
of them probably don't have bin entries to set up).
2024-11-12 09:23:39 -08:00

397 lines
13 KiB
Rust

// Copyright 2018-2024 the Deno authors. All rights reserved. MIT license.
use super::bin_entries::BinEntries;
use crate::args::LifecycleScriptsConfig;
use crate::task_runner::TaskStdio;
use crate::util::progress_bar::ProgressBar;
use deno_core::anyhow::Context;
use deno_npm::resolution::NpmResolutionSnapshot;
use deno_runtime::deno_io::FromRawIoHandle;
use deno_semver::package::PackageNv;
use deno_semver::Version;
use std::borrow::Cow;
use std::collections::HashSet;
use std::rc::Rc;
use std::path::Path;
use std::path::PathBuf;
use deno_core::error::AnyError;
use deno_npm::NpmResolutionPackage;
pub trait LifecycleScriptsStrategy {
fn can_run_scripts(&self) -> bool {
true
}
fn package_path(&self, package: &NpmResolutionPackage) -> PathBuf;
fn warn_on_scripts_not_run(
&self,
packages: &[(&NpmResolutionPackage, PathBuf)],
) -> Result<(), AnyError>;
fn has_warned(&self, package: &NpmResolutionPackage) -> bool;
fn has_run(&self, package: &NpmResolutionPackage) -> bool;
fn did_run_scripts(
&self,
package: &NpmResolutionPackage,
) -> Result<(), AnyError>;
}
pub struct LifecycleScripts<'a> {
packages_with_scripts: Vec<(&'a NpmResolutionPackage, PathBuf)>,
packages_with_scripts_not_run: Vec<(&'a NpmResolutionPackage, PathBuf)>,
config: &'a LifecycleScriptsConfig,
strategy: Box<dyn LifecycleScriptsStrategy + 'a>,
}
impl<'a> LifecycleScripts<'a> {
pub fn new<T: LifecycleScriptsStrategy + 'a>(
config: &'a LifecycleScriptsConfig,
strategy: T,
) -> Self {
Self {
config,
packages_with_scripts: Vec::new(),
packages_with_scripts_not_run: Vec::new(),
strategy: Box::new(strategy),
}
}
}
pub fn has_lifecycle_scripts(
package: &NpmResolutionPackage,
package_path: &Path,
) -> bool {
if let Some(install) = package.scripts.get("install") {
// default script
if !is_broken_default_install_script(install, package_path) {
return true;
}
}
package.scripts.contains_key("preinstall")
|| package.scripts.contains_key("postinstall")
}
// npm defaults to running `node-gyp rebuild` if there is a `binding.gyp` file
// but it always fails if the package excludes the `binding.gyp` file when they publish.
// (for example, `fsevents` hits this)
fn is_broken_default_install_script(script: &str, package_path: &Path) -> bool {
script == "node-gyp rebuild" && !package_path.join("binding.gyp").exists()
}
impl<'a> LifecycleScripts<'a> {
pub fn can_run_scripts(&self, package_nv: &PackageNv) -> bool {
if !self.strategy.can_run_scripts() {
return false;
}
use crate::args::PackagesAllowedScripts;
match &self.config.allowed {
PackagesAllowedScripts::All => true,
// TODO: make this more correct
PackagesAllowedScripts::Some(allow_list) => allow_list.iter().any(|s| {
let s = s.strip_prefix("npm:").unwrap_or(s);
s == package_nv.name || s == package_nv.to_string()
}),
PackagesAllowedScripts::None => false,
}
}
pub fn has_run_scripts(&self, package: &NpmResolutionPackage) -> bool {
self.strategy.has_run(package)
}
/// Register a package for running lifecycle scripts, if applicable.
///
/// `package_path` is the path containing the package's code (its root dir).
/// `package_meta_path` is the path to serve as the base directory for lifecycle
/// script-related metadata (e.g. to store whether the scripts have been run already)
pub fn add(
&mut self,
package: &'a NpmResolutionPackage,
package_path: Cow<Path>,
) {
if has_lifecycle_scripts(package, &package_path) {
if self.can_run_scripts(&package.id.nv) {
if !self.has_run_scripts(package) {
self
.packages_with_scripts
.push((package, package_path.into_owned()));
}
} else if !self.has_run_scripts(package)
&& (self.config.explicit_install || !self.strategy.has_warned(package))
{
// Skip adding `esbuild` as it is known that it can work properly without lifecycle script
// being run, and it's also very popular - any project using Vite would raise warnings.
{
let nv = &package.id.nv;
if nv.name == "esbuild"
&& nv.version >= Version::parse_standard("0.18.0").unwrap()
{
return;
}
}
self
.packages_with_scripts_not_run
.push((package, package_path.into_owned()));
}
}
}
pub fn warn_not_run_scripts(&self) -> Result<(), AnyError> {
if !self.packages_with_scripts_not_run.is_empty() {
self
.strategy
.warn_on_scripts_not_run(&self.packages_with_scripts_not_run)?;
}
Ok(())
}
pub async fn finish(
self,
snapshot: &NpmResolutionSnapshot,
packages: &[NpmResolutionPackage],
root_node_modules_dir_path: &Path,
progress_bar: &ProgressBar,
) -> Result<(), AnyError> {
self.warn_not_run_scripts()?;
let get_package_path =
|p: &NpmResolutionPackage| self.strategy.package_path(p);
let mut failed_packages = Vec::new();
let mut bin_entries = BinEntries::new();
if !self.packages_with_scripts.is_empty() {
let package_ids = self
.packages_with_scripts
.iter()
.map(|(p, _)| &p.id)
.collect::<HashSet<_>>();
// get custom commands for each bin available in the node_modules dir (essentially
// the scripts that are in `node_modules/.bin`)
let base = resolve_baseline_custom_commands(
&mut bin_entries,
snapshot,
packages,
get_package_path,
)?;
let init_cwd = &self.config.initial_cwd;
let process_state = crate::npm::managed::npm_process_state(
snapshot.as_valid_serialized(),
Some(root_node_modules_dir_path),
);
let mut env_vars = crate::task_runner::real_env_vars();
// we want to pass the current state of npm resolution down to the deno subprocess
// (that may be running as part of the script). we do this with an inherited temp file
//
// SAFETY: we are sharing a single temp file across all of the scripts. the file position
// will be shared among these, which is okay since we run only one script at a time.
// However, if we concurrently run scripts in the future we will
// have to have multiple temp files.
let temp_file_fd =
deno_runtime::ops::process::npm_process_state_tempfile(
process_state.as_bytes(),
).context("failed to create npm process state tempfile for running lifecycle scripts")?;
// SAFETY: fd/handle is valid
let _temp_file =
unsafe { std::fs::File::from_raw_io_handle(temp_file_fd) }; // make sure the file gets closed
env_vars.insert(
deno_runtime::ops::process::NPM_RESOLUTION_STATE_FD_ENV_VAR_NAME
.to_string(),
(temp_file_fd as usize).to_string(),
);
for (package, package_path) in self.packages_with_scripts {
// add custom commands for binaries from the package's dependencies. this will take precedence over the
// baseline commands, so if the package relies on a bin that conflicts with one higher in the dependency tree, the
// correct bin will be used.
let custom_commands = resolve_custom_commands_from_deps(
base.clone(),
package,
snapshot,
get_package_path,
)?;
for script_name in ["preinstall", "install", "postinstall"] {
if let Some(script) = package.scripts.get(script_name) {
if script_name == "install"
&& is_broken_default_install_script(script, &package_path)
{
continue;
}
let _guard = progress_bar.update_with_prompt(
crate::util::progress_bar::ProgressMessagePrompt::Initialize,
&format!("{}: running '{script_name}' script", package.id.nv),
);
let crate::task_runner::TaskResult {
exit_code,
stderr,
stdout,
} = crate::task_runner::run_task(
crate::task_runner::RunTaskOptions {
task_name: script_name,
script,
cwd: &package_path,
env_vars: env_vars.clone(),
custom_commands: custom_commands.clone(),
init_cwd,
argv: &[],
root_node_modules_dir: Some(root_node_modules_dir_path),
stdio: Some(crate::task_runner::TaskIo {
stderr: TaskStdio::piped(),
stdout: TaskStdio::piped(),
}),
},
)
.await?;
let stdout = stdout.unwrap();
let stderr = stderr.unwrap();
if exit_code != 0 {
log::warn!(
"error: script '{}' in '{}' failed with exit code {}{}{}",
script_name,
package.id.nv,
exit_code,
if !stdout.trim_ascii().is_empty() {
format!(
"\nstdout:\n{}\n",
String::from_utf8_lossy(&stdout).trim()
)
} else {
String::new()
},
if !stderr.trim_ascii().is_empty() {
format!(
"\nstderr:\n{}\n",
String::from_utf8_lossy(&stderr).trim()
)
} else {
String::new()
},
);
failed_packages.push(&package.id.nv);
// assume if earlier script fails, later ones will fail too
break;
}
}
}
self.strategy.did_run_scripts(package)?;
}
// re-set up bin entries for the packages which we've run scripts for.
// lifecycle scripts can create files that are linked to by bin entries,
// and the only reliable way to handle this is to re-link bin entries
// (this is what PNPM does as well)
bin_entries.finish_only(
snapshot,
&root_node_modules_dir_path.join(".bin"),
|outcome| outcome.warn_if_failed(),
&package_ids,
)?;
}
if failed_packages.is_empty() {
Ok(())
} else {
Err(AnyError::msg(format!(
"failed to run scripts for packages: {}",
failed_packages
.iter()
.map(|p| p.to_string())
.collect::<Vec<_>>()
.join(", ")
)))
}
}
}
// take in all (non copy) packages from snapshot,
// and resolve the set of available binaries to create
// custom commands available to the task runner
fn resolve_baseline_custom_commands<'a>(
bin_entries: &mut BinEntries<'a>,
snapshot: &'a NpmResolutionSnapshot,
packages: &'a [NpmResolutionPackage],
get_package_path: impl Fn(&NpmResolutionPackage) -> PathBuf,
) -> Result<crate::task_runner::TaskCustomCommands, AnyError> {
let mut custom_commands = crate::task_runner::TaskCustomCommands::new();
custom_commands
.insert("npx".to_string(), Rc::new(crate::task_runner::NpxCommand));
custom_commands
.insert("npm".to_string(), Rc::new(crate::task_runner::NpmCommand));
custom_commands
.insert("node".to_string(), Rc::new(crate::task_runner::NodeCommand));
custom_commands.insert(
"node-gyp".to_string(),
Rc::new(crate::task_runner::NodeGypCommand),
);
// TODO: this recreates the bin entries which could be redoing some work, but the ones
// we compute earlier in `sync_resolution_with_fs` may not be exhaustive (because we skip
// doing it for packages that are set up already.
// realistically, scripts won't be run very often so it probably isn't too big of an issue.
resolve_custom_commands_from_packages(
bin_entries,
custom_commands,
snapshot,
packages,
get_package_path,
)
}
// resolves the custom commands from an iterator of packages
// and adds them to the existing custom commands.
// note that this will overwrite any existing custom commands
fn resolve_custom_commands_from_packages<
'a,
P: IntoIterator<Item = &'a NpmResolutionPackage>,
>(
bin_entries: &mut BinEntries<'a>,
mut commands: crate::task_runner::TaskCustomCommands,
snapshot: &'a NpmResolutionSnapshot,
packages: P,
get_package_path: impl Fn(&'a NpmResolutionPackage) -> PathBuf,
) -> Result<crate::task_runner::TaskCustomCommands, AnyError> {
for package in packages {
let package_path = get_package_path(package);
if package.bin.is_some() {
bin_entries.add(package, package_path);
}
}
let bins: Vec<(String, PathBuf)> = bin_entries.collect_bin_files(snapshot);
for (bin_name, script_path) in bins {
commands.insert(
bin_name.clone(),
Rc::new(crate::task_runner::NodeModulesFileRunCommand {
command_name: bin_name,
path: script_path,
}),
);
}
Ok(commands)
}
// resolves the custom commands from the dependencies of a package
// and adds them to the existing custom commands.
// note that this will overwrite any existing custom commands.
fn resolve_custom_commands_from_deps(
baseline: crate::task_runner::TaskCustomCommands,
package: &NpmResolutionPackage,
snapshot: &NpmResolutionSnapshot,
get_package_path: impl Fn(&NpmResolutionPackage) -> PathBuf,
) -> Result<crate::task_runner::TaskCustomCommands, AnyError> {
let mut bin_entries = BinEntries::new();
resolve_custom_commands_from_packages(
&mut bin_entries,
baseline,
snapshot,
package
.dependencies
.values()
.map(|id| snapshot.package_from_id(id).unwrap()),
get_package_path,
)
}