From e20bee27e959d2743c3d6929723d781b80a4354a Mon Sep 17 00:00:00 2001
From: Michael-F-Bryan <michael@wasmer.io>
Date: Mon, 29 May 2023 13:41:00 +0800
Subject: [PATCH] Added a polyfill for Url::to_file_path()

---
 .../runtime/package_loader/builtin_loader.rs  | 21 ++++++---
 lib/wasi/src/runtime/resolver/utils.rs        | 45 +++++++++++++++++--
 2 files changed, 56 insertions(+), 10 deletions(-)

diff --git a/lib/wasi/src/runtime/package_loader/builtin_loader.rs b/lib/wasi/src/runtime/package_loader/builtin_loader.rs
index c6f44d97922..627b2e2cfb7 100644
--- a/lib/wasi/src/runtime/package_loader/builtin_loader.rs
+++ b/lib/wasi/src/runtime/package_loader/builtin_loader.rs
@@ -89,13 +89,20 @@ impl BuiltinPackageLoader {
 
     async fn download(&self, dist: &DistributionInfo) -> Result<Bytes, Error> {
         if dist.webc.scheme() == "file" {
-            // Note: The Url::to_file_path() method is platform-specific
-            #[cfg(any(unix, windows, target_os = "redox", target_os = "wasi"))]
-            if let Ok(path) = dist.webc.to_file_path() {
-                // FIXME: This will block the thread
-                let bytes = std::fs::read(&path)
-                    .with_context(|| format!("Unable to read \"{}\"", path.display()))?;
-                return Ok(bytes.into());
+            match crate::runtime::resolver::utils::file_path_from_url(&dist.webc) {
+                Ok(path) => {
+                    // FIXME: This will block the thread
+                    let bytes = std::fs::read(&path)
+                        .with_context(|| format!("Unable to read \"{}\"", path.display()))?;
+                    return Ok(bytes.into());
+                }
+                Err(e) => {
+                    tracing::debug!(
+                        url=%dist.webc,
+                        error=&*e,
+                        "Unable to convert the file:// URL to a path",
+                    );
+                }
             }
         }
 
diff --git a/lib/wasi/src/runtime/resolver/utils.rs b/lib/wasi/src/runtime/resolver/utils.rs
index 521215b2cb2..152c93ead21 100644
--- a/lib/wasi/src/runtime/resolver/utils.rs
+++ b/lib/wasi/src/runtime/resolver/utils.rs
@@ -1,7 +1,7 @@
-use std::path::Path;
+use std::path::{Path, PathBuf};
 
-use anyhow::Error;
 use http::{HeaderMap, StatusCode};
+use anyhow::{Context, Error};
 use url::Url;
 
 use crate::http::{HttpResponse, USER_AGENT};
@@ -56,13 +56,39 @@ pub(crate) fn http_error(response: &HttpResponse) -> Error {
     Error::msg(status)
 }
 
+pub(crate) fn file_path_from_url(url: &Url) -> Result<PathBuf, Error> {
+    debug_assert_eq!(url.scheme(), "file");
+
+    // Note: The Url::to_file_path() method is platform-specific
+    cfg_if::cfg_if! {
+        if #[cfg(any(unix, windows, target_os = "redox", target_os = "wasi"))] {
+            if let Ok(path) = url.to_file_path() {
+                return Ok(path);
+            }
+
+            // Sometimes we'll get a UNC-like path (e.g.
+            // "file:///?\\C:/\\/path/to/file.txt") and Url::to_file_path()
+            // won't be able to handle the "\\?" so we try to "massage" the URL
+            // a bit.
+            // See <https://github.com/servo/rust-url/issues/450> for more.
+            let modified = url.as_str().replace(r"\\?", "").replace("//?", "").replace('\\', "/");
+            Url::parse(&modified)
+                .ok()
+                .and_then(|url| url.to_file_path().ok())
+                .context("Unable to extract the file path")
+        } else {
+            anyhow::bail!("Url::to_file_path() is not supported on this platform");
+        }
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;
 
     #[test]
     #[cfg(unix)]
-    fn behaviour_is_identical() {
+    fn from_file_path_behaviour_is_identical() {
         let inputs = [
             "/",
             "/path",
@@ -78,4 +104,17 @@ mod tests {
             assert_eq!(got, expected, "Mismatch for \"{path}\"");
         }
     }
+
+    #[test]
+    #[cfg(windows)]
+    fn to_file_path_can_handle_unc_paths() {
+        let path = Path::new(env!("CARGO_MANIFEST_DIR"))
+            .canonicalize()
+            .unwrap();
+        let url = Url::from_file_path(&path).unwrap();
+
+        let got = file_path_from_url(&url).unwrap();
+
+        assert_eq!(got.canonicalize().unwrap(), path);
+    }
 }