From 7664ba3ec919d2dd50e4f67dea23891803957f29 Mon Sep 17 00:00:00 2001
From: neri <neri@ctdo.de>
Date: Tue, 25 Oct 2022 20:55:29 +0200
Subject: [PATCH] don't reread file to perform mime guess

---
 Cargo.lock       |  2 +-
 Cargo.toml       |  2 +-
 src/multipart.rs | 36 +++++++++++++++++++++---------------
 3 files changed, 23 insertions(+), 17 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index 601710f..992ff07 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -416,7 +416,7 @@ dependencies = [
 
 [[package]]
 name = "datatrash"
-version = "2.0.0"
+version = "2.0.1"
 dependencies = [
  "actix-files",
  "actix-governor",
diff --git a/Cargo.toml b/Cargo.toml
index 36dd045..93e1bee 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "datatrash"
-version = "2.0.0"
+version = "2.0.1"
 authors = ["neri"]
 edition = "2021"
 
diff --git a/src/multipart.rs b/src/multipart.rs
index d0d42d0..0fa4a3e 100644
--- a/src/multipart.rs
+++ b/src/multipart.rs
@@ -3,7 +3,10 @@ use actix_multipart::{Field, Multipart};
 use actix_web::{error, http::header::DispositionParam, Error};
 use futures_util::{StreamExt, TryStreamExt};
 use mime::{Mime, APPLICATION_OCTET_STREAM, TEXT_PLAIN};
-use std::path::Path;
+use std::{
+    cmp::{max, min},
+    path::Path,
+};
 use time::{Duration, OffsetDateTime};
 use tokio::{fs::File, io::AsyncWriteExt};
 
@@ -42,10 +45,10 @@ pub(crate) async fn parse_multipart(
                     continue;
                 }
                 original_name = uploaded_name;
-                size = create_file(file_path, field, config.max_file_size).await?;
-                println!("mime: {}", mime);
+                let first_bytes;
+                (size, first_bytes) = create_file(file_path, field, config.max_file_size).await?;
                 content_type = Some(if mime == APPLICATION_OCTET_STREAM {
-                    get_content_type(file_path)
+                    get_content_type(&first_bytes)
                 } else {
                     mime_relations::get_alias(mime)
                 });
@@ -54,8 +57,9 @@ pub(crate) async fn parse_multipart(
                 if original_name.is_some() {
                     continue;
                 }
-                size = create_file(file_path, field, config.max_file_size).await?;
-                content_type = Some(get_content_type(file_path));
+                let first_bytes;
+                (size, first_bytes) = create_file(file_path, field, config.max_file_size).await?;
+                content_type = Some(get_content_type(&first_bytes));
             }
             "delete_on_download" => {
                 delete_on_download = parse_string(name, field).await? != "false";
@@ -149,23 +153,25 @@ async fn create_file(
     filename: &Path,
     field: Field,
     max_file_size: Option<u64>,
-) -> Result<u64, Error> {
+) -> Result<(u64, Vec<u8>), Error> {
     let mut file = File::create(&filename).await.map_err(|file_err| {
         log::error!("could not create file {:?}", file_err);
         error::ErrorInternalServerError("could not create file")
     })?;
-    let written_bytes = write_to_file(&mut file, field, max_file_size).await?;
-    Ok(written_bytes)
+    write_to_file(&mut file, field, max_file_size).await
 }
 
 async fn write_to_file(
     file: &mut File,
     mut field: Field,
     max_size: Option<u64>,
-) -> Result<u64, error::Error> {
+) -> Result<(u64, Vec<u8>), error::Error> {
+    let mut first_bytes = Vec::with_capacity(2048);
     let mut written_bytes: u64 = 0;
     while let Some(chunk) = field.next().await {
         let chunk = chunk.map_err(error::ErrorBadRequest)?;
+        let remaining_first_bytes = min(max(0, 2048 - written_bytes) as usize, chunk.len());
+        first_bytes.extend_from_slice(&chunk[0..remaining_first_bytes]);
         written_bytes += chunk.len() as u64;
         if let Some(max_size) = max_size {
             if written_bytes > max_size {
@@ -179,7 +185,7 @@ async fn write_to_file(
             error::ErrorInternalServerError("could not write file")
         })?;
     }
-    Ok(written_bytes)
+    Ok((written_bytes, first_bytes))
 }
 
 fn get_file_metadata(field: &actix_multipart::Field) -> (Mime, Option<String>) {
@@ -195,9 +201,9 @@ fn get_file_metadata(field: &actix_multipart::Field) -> (Mime, Option<String>) {
     (mime, filename)
 }
 
-fn get_content_type(path: &Path) -> Mime {
-    let std_path = std::path::Path::new(path.as_os_str());
-    tree_magic_mini::from_filepath(std_path)
-        .and_then(|mime| mime.parse().ok())
+fn get_content_type(bytes: &[u8]) -> Mime {
+    tree_magic_mini::from_u8(bytes)
+        .parse()
+        .ok()
         .unwrap_or(TEXT_PLAIN)
 }