From df30de261bfc15c9eec787d971ee5172647e8484 Mon Sep 17 00:00:00 2001
From: Ruihang Xia <waynestxia@gmail.com>
Date: Fri, 12 Jan 2024 22:58:57 +0800
Subject: [PATCH] feat: add datafusion to the list

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
---
 datafusion/core/Cargo.toml                    | 7 ++++---
 datafusion/core/src/datasource/listing/url.rs | 4 ++++
 datafusion/wasmtest/Cargo.toml                | 1 +
 datafusion/wasmtest/README.md                 | 1 +
 4 files changed, 10 insertions(+), 3 deletions(-)

diff --git a/datafusion/core/Cargo.toml b/datafusion/core/Cargo.toml
index 9de6a7f7d6a0..c2e8c2b44531 100644
--- a/datafusion/core/Cargo.toml
+++ b/datafusion/core/Cargo.toml
@@ -37,7 +37,7 @@ path = "src/lib.rs"
 # Used to enable the avro format
 avro = ["apache-avro", "num-traits", "datafusion-common/avro"]
 backtrace = ["datafusion-common/backtrace"]
-compression = ["xz2", "bzip2", "flate2", "zstd", "async-compression"]
+compression = ["xz2", "bzip2", "flate2", "zstd", "async-compression", "tokio-util"]
 crypto_expressions = ["datafusion-physical-expr/crypto_expressions", "datafusion-optimizer/crypto_expressions"]
 default = ["crypto_expressions", "encoding_expressions", "regex_expressions", "unicode_expressions", "compression", "parquet"]
 encoding_expressions = ["datafusion-physical-expr/encoding_expressions"]
@@ -87,8 +87,8 @@ pin-project-lite = "^0.2.7"
 rand = { workspace = true }
 sqlparser = { workspace = true }
 tempfile = { workspace = true }
-tokio = { version = "1.28", features = ["macros", "rt", "rt-multi-thread", "sync", "fs", "parking_lot"] }
-tokio-util = { version = "0.7.4", features = ["io"] }
+tokio = { version = "1.28", features = ["macros", "rt",  "sync"] }
+tokio-util = { version = "0.7.4", features = ["io"], optional = true }
 url = { workspace = true }
 uuid = { version = "1.0", features = ["v4"] }
 xz2 = { version = "0.1", optional = true }
@@ -113,6 +113,7 @@ rust_decimal = { version = "1.27.0", features = ["tokio-pg"] }
 serde_json = { workspace = true }
 test-utils = { path = "../../test-utils" }
 thiserror = { workspace = true }
+tokio = { version = "1.28", features = ["macros", "rt", "rt-multi-thread", "sync", "fs", "parking_lot"] }
 tokio-postgres = "0.7.7"
 [target.'cfg(not(target_os = "windows"))'.dev-dependencies]
 nix = { version = "0.27.1", features = ["fs"] }
diff --git a/datafusion/core/src/datasource/listing/url.rs b/datafusion/core/src/datasource/listing/url.rs
index 766dee7de901..6421edf77972 100644
--- a/datafusion/core/src/datasource/listing/url.rs
+++ b/datafusion/core/src/datasource/listing/url.rs
@@ -103,12 +103,14 @@ impl ListingTableUrl {
         let s = s.as_ref();
 
         // This is necessary to handle the case of a path starting with a drive letter
+        #[cfg(not(target_arch = "wasm32"))]
         if std::path::Path::new(s).is_absolute() {
             return Self::parse_path(s);
         }
 
         match Url::parse(s) {
             Ok(url) => Self::try_new(url, None),
+            #[cfg(not(target_arch = "wasm32"))]
             Err(url::ParseError::RelativeUrlWithoutBase) => Self::parse_path(s),
             Err(e) => Err(DataFusionError::External(Box::new(e))),
         }
@@ -146,6 +148,7 @@ impl ListingTableUrl {
     }
 
     /// Creates a new [`ListingTableUrl`] interpreting `s` as a filesystem path
+    #[cfg(not(target_arch = "wasm32"))]
     fn parse_path(s: &str) -> Result<Self> {
         let (path, glob) = match split_glob_expression(s) {
             Some((prefix, glob)) => {
@@ -282,6 +285,7 @@ impl ListingTableUrl {
 }
 
 /// Creates a file URL from a potentially relative filesystem path
+#[cfg(not(target_arch = "wasm32"))]
 fn url_from_filesystem_path(s: &str) -> Option<Url> {
     let path = std::path::Path::new(s);
     let is_dir = match path.exists() {
diff --git a/datafusion/wasmtest/Cargo.toml b/datafusion/wasmtest/Cargo.toml
index cc39cb031b54..e1664825f798 100644
--- a/datafusion/wasmtest/Cargo.toml
+++ b/datafusion/wasmtest/Cargo.toml
@@ -45,6 +45,7 @@ datafusion-physical-expr = { workspace = true }
 datafusion-sql = { workspace = true }
 datafusion-execution = { workspace = true }
 datafusion-physical-plan = { workspace = true }
+datafusion = { path = "../core", default-features = false }
 
 # getrandom must be compiled with js feature
 getrandom = { version = "0.2.8", features = ["js"] }
diff --git a/datafusion/wasmtest/README.md b/datafusion/wasmtest/README.md
index c1624fb60fee..4af0f94db9e9 100644
--- a/datafusion/wasmtest/README.md
+++ b/datafusion/wasmtest/README.md
@@ -59,6 +59,7 @@ Then open http://localhost:8080/ in a web browser and check the console to see t
 
 The following DataFusion crates are verified to work in a wasm-pack environment using the default `wasm32-unknown-unknown` target:
 
+- `datafusion` (datafusion-core) with default-features disabled to remove `bzip2-sys` from `async-compression`
 - `datafusion-common` with default-features disabled to remove the `parquet` dependency (see below)
 - `datafusion-expr`
 - `datafusion-execution`