Skip to content
Open
Changes from 1 commit
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
208 changes: 131 additions & 77 deletions graph/src/ipfs/content_path.rs
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
use anyhow::anyhow;
use cid::Cid;
use url::Url;

use crate::ipfs::IpfsError;
use crate::ipfs::IpfsResult;
Expand All @@ -13,39 +14,79 @@ pub struct ContentPath {

impl ContentPath {
/// Creates a new [ContentPath] from the specified input.
///
/// Supports the following formats:
/// - <CID>[/<path>]
/// - /ipfs/<CID>[/<path>]
/// - ipfs://<CID>[/<path>]
/// - http[s]://.../ipfs/<CID>[/<path>]
/// - http[s]://.../api/v0/cat?arg=<CID>[/<path>]
pub fn new(input: impl AsRef<str>) -> IpfsResult<Self> {
let input = input.as_ref();
let input = input.as_ref().trim();

if input.is_empty() {
return Err(IpfsError::InvalidContentPath {
input: "".to_owned(),
source: anyhow!("path is empty"),
input: "".to_string(),
source: anyhow!("content path is empty"),
});
}

let (cid, path) = input
.strip_prefix("/ipfs/")
.unwrap_or(input)
.split_once('/')
.unwrap_or((input, ""));
if input.starts_with("http://") || input.starts_with("https://") {
return Self::parse_from_url(input);
}

Self::parse_from_cid_and_path(input)
}

fn parse_from_url(input: &str) -> IpfsResult<Self> {
let url = Url::parse(input).map_err(|_err| IpfsError::InvalidContentPath {
input: input.to_string(),
source: anyhow!("input is not a valid URL"),
})?;

if let Some((_, x)) = url.query_pairs().find(|(key, _)| key == "arg") {
return Self::parse_from_cid_and_path(&x);
}

if let Some((_, x)) = url.path().split_once("/ipfs/") {
return Self::parse_from_cid_and_path(x);
}

Self::parse_from_cid_and_path(url.path())
}

fn parse_from_cid_and_path(mut input: &str) -> IpfsResult<Self> {
input = input.trim_matches('/');

for prefix in ["ipfs/", "ipfs://"] {
if let Some(input_without_prefix) = input.strip_prefix(prefix) {
input = input_without_prefix
}
}

let (cid, path) = input.split_once('/').unwrap_or((input, ""));

let cid = cid
.parse::<Cid>()
.map_err(|err| IpfsError::InvalidContentPath {
input: input.to_owned(),
input: input.to_string(),
source: anyhow::Error::from(err).context("invalid CID"),
})?;

if path.contains('?') {
return Err(IpfsError::InvalidContentPath {
input: input.to_owned(),
input: input.to_string(),
source: anyhow!("query parameters not allowed"),
});
}

Ok(Self {
cid,
path: (!path.is_empty()).then_some(path.to_owned()),
path: if path.is_empty() {
None
} else {
Some(path.to_string())
},
})
}

Expand Down Expand Up @@ -97,13 +138,20 @@ mod tests {
const CID_V0: &str = "QmUNLLsPACCz1vLxQVkXqqLX5R1X345qqfHbsf67hvA3Nn";
const CID_V1: &str = "bafybeiczsscdsbs7ffqz55asqdf3smv6klcw3gofszvwlyarci47bgf354";

fn make_path(cid: &str, path: Option<&str>) -> ContentPath {
ContentPath {
cid: cid.parse().unwrap(),
path: path.map(ToOwned::to_owned),
}
}

#[test]
fn fails_on_empty_input() {
let err = ContentPath::new("").unwrap_err();

assert_eq!(
err.to_string(),
"'' is not a valid IPFS content path: path is empty",
"'' is not a valid IPFS content path: content path is empty",
);
}

Expand All @@ -119,75 +167,37 @@ mod tests {
#[test]
fn accepts_a_valid_cid_v0() {
let path = ContentPath::new(CID_V0).unwrap();

assert_eq!(
path,
ContentPath {
cid: CID_V0.parse().unwrap(),
path: None,
}
);
assert_eq!(path, make_path(CID_V0, None));
}

#[test]
fn accepts_a_valid_cid_v1() {
let path = ContentPath::new(CID_V1).unwrap();

assert_eq!(
path,
ContentPath {
cid: CID_V1.parse().unwrap(),
path: None,
}
);
assert_eq!(path, make_path(CID_V1, None));
}

#[test]
fn fails_on_a_leading_slash_followed_by_a_valid_cid() {
let err = ContentPath::new(format!("/{CID_V0}")).unwrap_err();
fn accepts_and_removes_leading_slashes() {
let path = ContentPath::new(format!("/{CID_V0}")).unwrap();
assert_eq!(path, make_path(CID_V0, None));

assert!(err.to_string().starts_with(&format!(
"'/{CID_V0}' is not a valid IPFS content path: invalid CID: "
)));
let path = ContentPath::new(format!("///////{CID_V0}")).unwrap();
assert_eq!(path, make_path(CID_V0, None));
}

#[test]
fn ignores_the_first_slash_after_the_cid() {
fn accepts_and_removes_trailing_slashes() {
let path = ContentPath::new(format!("{CID_V0}/")).unwrap();
assert_eq!(path, make_path(CID_V0, None));

assert_eq!(
path,
ContentPath {
cid: CID_V0.parse().unwrap(),
path: None,
}
);
let path = ContentPath::new(format!("{CID_V0}///////")).unwrap();
assert_eq!(path, make_path(CID_V0, None));
}

#[test]
fn accepts_a_path_after_the_cid() {
let path = ContentPath::new(format!("{CID_V0}/readme.md")).unwrap();

assert_eq!(
path,
ContentPath {
cid: CID_V0.parse().unwrap(),
path: Some("readme.md".to_owned()),
}
);
}

#[test]
fn accepts_multiple_consecutive_slashes_after_the_cid() {
let path = ContentPath::new(format!("{CID_V0}//")).unwrap();

assert_eq!(
path,
ContentPath {
cid: CID_V0.parse().unwrap(),
path: Some("/".to_owned()),
}
);
assert_eq!(path, make_path(CID_V0, Some("readme.md")));
}

#[test]
Expand All @@ -214,23 +224,67 @@ mod tests {
#[test]
fn accepts_and_removes_the_ipfs_prefix() {
let path = ContentPath::new(format!("/ipfs/{CID_V0}")).unwrap();

assert_eq!(
path,
ContentPath {
cid: CID_V0.parse().unwrap(),
path: None,
}
);
assert_eq!(path, make_path(CID_V0, None));

let path = ContentPath::new(format!("/ipfs/{CID_V0}/readme.md")).unwrap();
assert_eq!(path, make_path(CID_V0, Some("readme.md")));
}

assert_eq!(
path,
ContentPath {
cid: CID_V0.parse().unwrap(),
path: Some("readme.md".to_owned()),
}
);
#[test]
fn accepts_and_removes_the_ipfs_schema() {
let path = ContentPath::new(format!("ipfs://{CID_V0}")).unwrap();
assert_eq!(path, make_path(CID_V0, None));

let path = ContentPath::new(format!("ipfs://{CID_V0}/readme.md")).unwrap();
assert_eq!(path, make_path(CID_V0, Some("readme.md")));
}

#[test]
fn accepts_and_parses_ipfs_rpc_urls() {
let path = ContentPath::new(format!("http://ipfs.com/api/v0/cat?arg={CID_V0}")).unwrap();
assert_eq!(path, make_path(CID_V0, None));

let path =
ContentPath::new(format!("http://ipfs.com/api/v0/cat?arg={CID_V0}/readme.md")).unwrap();
assert_eq!(path, make_path(CID_V0, Some("readme.md")));

let path = ContentPath::new(format!("https://ipfs.com/api/v0/cat?arg={CID_V0}")).unwrap();
assert_eq!(path, make_path(CID_V0, None));

let path = ContentPath::new(format!(
"https://ipfs.com/api/v0/cat?arg={CID_V0}/readme.md"
))
.unwrap();
assert_eq!(path, make_path(CID_V0, Some("readme.md")));
}

#[test]
fn accepts_and_parses_ipfs_gateway_urls() {
let path = ContentPath::new(format!("http://ipfs.com/ipfs/{CID_V0}")).unwrap();
assert_eq!(path, make_path(CID_V0, None));

let path = ContentPath::new(format!("http://ipfs.com/ipfs/{CID_V0}/readme.md")).unwrap();
assert_eq!(path, make_path(CID_V0, Some("readme.md")));

let path = ContentPath::new(format!("https://ipfs.com/ipfs/{CID_V0}")).unwrap();
assert_eq!(path, make_path(CID_V0, None));

let path = ContentPath::new(format!("https://ipfs.com/ipfs/{CID_V0}/readme.md")).unwrap();
assert_eq!(path, make_path(CID_V0, Some("readme.md")));
}

#[test]
fn accepts_and_parses_paths_from_urls() {
let path = ContentPath::new(format!("http://ipfs.com/{CID_V0}")).unwrap();
assert_eq!(path, make_path(CID_V0, None));

let path = ContentPath::new(format!("http://ipfs.com/{CID_V0}/readme.md")).unwrap();
assert_eq!(path, make_path(CID_V0, Some("readme.md")));

let path = ContentPath::new(format!("https://ipfs.com/{CID_V0}")).unwrap();
assert_eq!(path, make_path(CID_V0, None));

let path = ContentPath::new(format!("https://ipfs.com/{CID_V0}/readme.md")).unwrap();
assert_eq!(path, make_path(CID_V0, Some("readme.md")));
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Everything in this file looks great, but since we are handling user input here, are there ways to abuse this in security-relevant ways? (I think the answer is 'no', but wanted to double-check)

}
}
Loading