From abb1b5681ebd08f78d62420ce175ea852dcafa45 Mon Sep 17 00:00:00 2001
From: Olivia Lee <olivia@computer.surgery>
Date: Sun, 6 Apr 2025 17:43:12 -0700
Subject: [PATCH] add partial_canonicalize helper function

This is useful for checking for potential overlap between paths that
have not been fully created yet.
---
 Cargo.lock   |  14 +++++++
 Cargo.toml   |   1 +
 src/utils.rs | 111 ++++++++++++++++++++++++++++++++++++++++++++++++++-
 3 files changed, 125 insertions(+), 1 deletion(-)

diff --git a/Cargo.lock b/Cargo.lock
index 2641dca3..59f2ff41 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -962,6 +962,7 @@ dependencies = [
  "serde_yaml",
  "sha-1",
  "strum",
+ "tempfile",
  "thiserror 2.0.12",
  "thread_local",
  "tikv-jemallocator",
@@ -3195,6 +3196,19 @@ dependencies = [
  "syn",
 ]
 
+[[package]]
+name = "tempfile"
+version = "3.19.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7437ac7763b9b123ccf33c338a5cc1bac6f69b45a136c19bdd8a65e3916435bf"
+dependencies = [
+ "fastrand",
+ "getrandom 0.3.2",
+ "once_cell",
+ "rustix",
+ "windows-sys 0.59.0",
+]
+
 [[package]]
 name = "terminal_size"
 version = "0.4.2"
diff --git a/Cargo.toml b/Cargo.toml
index f0913909..00971197 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -153,6 +153,7 @@ nix = { version = "0.29", features = ["resource", "time"] }
 assert_cmd = "2.0.16"
 insta = { version = "1.42.2", features = ["filters", "json", "redactions"] }
 predicates = "3.1.3"
+tempfile = "3.19.1"
 
 [profile.dev.package.insta]
 opt-level = 3
diff --git a/src/utils.rs b/src/utils.rs
index 45211399..568bc5be 100644
--- a/src/utils.rs
+++ b/src/utils.rs
@@ -2,6 +2,8 @@ use std::{
     borrow::Cow,
     cmp, fmt,
     fmt::Write,
+    io,
+    path::{Component, Path, PathBuf},
     str::FromStr,
     time::{SystemTime, UNIX_EPOCH},
 };
@@ -14,6 +16,7 @@ use ruma::{
     api::client::error::ErrorKind, canonical_json::try_from_json_map,
     CanonicalJsonError, CanonicalJsonObject, MxcUri, MxcUriError, OwnedMxcUri,
 };
+use tokio::fs;
 
 use crate::{Error, Result};
 
@@ -380,9 +383,67 @@ pub(crate) fn u8_slice_to_hex(slice: &[u8]) -> String {
     })
 }
 
+/// Canonicalize a path where some components may not exist yet.
+///
+/// It's assumed that non-existent components will be created as
+/// directories. This should match the result of [`fs::canonicalize`]
+/// _after_ calling [`fs::create_dir_all`] on `path`.
+#[allow(dead_code)]
+pub(crate) async fn partial_canonicalize(path: &Path) -> io::Result<PathBuf> {
+    let mut ret = std::env::current_dir()?;
+
+    let mut base_path = Cow::Borrowed(path);
+    let mut components = base_path.components();
+
+    while let Some(component) = components.next() {
+        match component {
+            Component::Prefix(_) | Component::RootDir => {
+                let component_path: &Path = component.as_ref();
+                component_path.clone_into(&mut ret);
+            }
+            Component::CurDir => (),
+            Component::ParentDir => {
+                ret.pop();
+            }
+            Component::Normal(p) => {
+                let component_path = ret.join(p);
+                match fs::symlink_metadata(&component_path).await {
+                    // path is a symlink
+                    Ok(metadata) if metadata.is_symlink() => {
+                        let destination =
+                            fs::read_link(&component_path).await?;
+                        // iterate over the symlink destination components
+                        // before continuing with the original path
+                        base_path =
+                            Cow::Owned(destination.join(components.as_path()));
+                        components = base_path.components();
+                    }
+                    // path exists, not a symlink
+                    Ok(_) => {
+                        ret.push(p);
+                    }
+                    // path does not exist
+                    Err(error) if error.kind() == io::ErrorKind::NotFound => {
+                        // assume a directory will be created here
+                        ret.push(p);
+                    }
+                    Err(error) => return Err(error),
+                }
+            }
+        }
+    }
+
+    Ok(ret)
+}
+
 #[cfg(test)]
 mod tests {
-    use crate::utils::{dbg_truncate_str, u8_slice_to_hex};
+    use tempfile::TempDir;
+    use tokio::fs;
+
+    use crate::utils::{
+        dbg_truncate_str, partial_canonicalize, u8_slice_to_hex,
+    };
 
     #[test]
     fn test_truncate_str() {
@@ -412,4 +473,52 @@ mod tests {
              4242424242424242424242424242424242424242"
         );
     }
+
+    #[tokio::test]
+    async fn test_partial_canonicalize() {
+        let tmp_dir =
+            TempDir::with_prefix("test_partial_canonicalize").unwrap();
+        let path = tmp_dir.path();
+
+        fs::create_dir(&path.join("dir")).await.unwrap();
+        fs::symlink(path.join("dir"), path.join("absolute-link-to-dir"))
+            .await
+            .unwrap();
+        fs::symlink("./dir", path.join("relative-link-to-dir")).await.unwrap();
+
+        assert_eq!(partial_canonicalize(path).await.unwrap(), path);
+        assert_eq!(partial_canonicalize(&path.join("./")).await.unwrap(), path);
+        assert_eq!(
+            partial_canonicalize(&path.join("dir/..")).await.unwrap(),
+            path
+        );
+        assert_eq!(
+            partial_canonicalize(&path.join("absolute-link-to-dir"))
+                .await
+                .unwrap(),
+            path.join("dir")
+        );
+        assert_eq!(
+            partial_canonicalize(&path.join("relative-link-to-dir"))
+                .await
+                .unwrap(),
+            path.join("dir")
+        );
+        assert_eq!(
+            partial_canonicalize(&path.join("absolute-link-to-dir/new-dir"))
+                .await
+                .unwrap(),
+            path.join("dir/new-dir")
+        );
+        assert_eq!(
+            partial_canonicalize(
+                &path.join("absolute-link-to-dir/new-dir/../..")
+            )
+            .await
+            .unwrap(),
+            path,
+        );
+
+        tmp_dir.close().unwrap();
+    }
 }