load mime parents to determine text display

Cargo.lock

@@ -427,7 +427,7 @@ dependencies = [
 
 [[package]]
 name = "datatrash"
-version = "1.1.8"
+version = "1.1.9"
 dependencies = [
  "actix-files",
  "actix-governor",
@@ -437,12 +437,14 @@ dependencies = [
  "futures-util",
  "governor",
  "htmlescape",
+ "lazy_static",
  "log",
  "mime",
  "rand",
  "sqlx",
  "time",
  "tokio",
+ "tree_magic_db",
  "tree_magic_mini",
  "url",
  "urlencoding",
@@ -870,9 +872,9 @@ dependencies = [
 
 [[package]]
 name = "itoa"
-version = "1.0.3"
+version = "1.0.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6c8af84674fe1f223a982c933a0ee1086ac4d4052aa0fb8060c12c6ad838e754"
+checksum = "4217ad341ebadf8d8e724e264f13e593e0648f5b3e94b3896a5df283be015ecc"
 
 [[package]]
 name = "jobserver"
@@ -1607,9 +1609,9 @@ dependencies = [
 
 [[package]]
 name = "time"
-version = "0.3.14"
+version = "0.3.15"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3c3f9a28b618c3a6b9251b6908e9c99e04b9e5c02e6581ccbb67d59c34ef7f9b"
+checksum = "d634a985c4d4238ec39cacaed2e7ae552fbd3c476b552c1deac3021b7d7eaf0c"
 dependencies = [
  "itoa",
  "libc",

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "datatrash"
-version = "1.1.8"
+version = "1.1.9"
 authors = ["neri"]
 edition = "2021"
 
@@ -24,11 +24,13 @@ tokio = { version = "1.21.2", features = ["rt", "macros", "sync"] }
 actix-multipart = "0.4.0"
 futures-util = "0.3.24"
 rand = "0.8.5"
-time = "0.3.14"
+time = "0.3.15"
 htmlescape = "0.3.1"
 urlencoding = "2.1.2"
 tree_magic_mini = { version = "3.0.3", features = ["with-gpl-data"] }
+tree_magic_db = "*"
 mime = "0.3.16"
 url = "2.3.1"
 actix-governor = "0.3.2"
 governor = "0.4.2"
+lazy_static = "1.4.0"

src/download.rs

@@ -17,8 +17,7 @@ use time::OffsetDateTime;
 use tokio::fs;
 use url::Url;
 
-use crate::config::Config;
-use crate::deleter;
+use crate::{config::Config, deleter, mime_relations};
 
 const TEXT_VIEW_HTML: &str = include_str!("../template/text-view.html");
 const URL_VIEW_HTML: &str = include_str!("../template/url-view.html");
@@ -42,7 +41,8 @@ pub async fn download(
     path.push(&file_id);
 
     let mime = Mime::from_str(&content_type).unwrap_or(APPLICATION_OCTET_STREAM);
-    let mut response = match get_view_type(&req, &mime, &path, delete).await {
+    let mime = mime_relations::get_alias(&mime);
+    let mut response = match get_view_type(&req, mime, &path, delete).await {
         ViewType::Raw => build_file_response(false, &file_name, path, mime, &req).await,
         ViewType::Download => build_file_response(true, &file_name, path, mime, &req).await,
         ViewType::Html => build_text_response(&path).await,
@@ -91,7 +91,7 @@ async fn get_view_type(
     if req.query_string().contains("raw") {
         return ViewType::Raw;
     }
-    if mime.type_() != mime::TEXT {
+    if !mime_relations::matches_text(mime) {
         return ViewType::Raw;
     }
     if get_file_size(file_path).await >= TEXT_VIEW_SIZE_LIMIT {
@@ -146,7 +146,7 @@ async fn build_file_response(
     download: bool,
     file_name: &str,
     path: PathBuf,
-    mime: Mime,
+    mime: &Mime,
     req: &HttpRequest,
 ) -> Result<HttpResponse, Error> {
     let content_disposition = ContentDisposition {
@@ -162,7 +162,7 @@ async fn build_file_response(
             log::error!("file could not be read {:?}", file_err);
             error::ErrorInternalServerError("this file should be here but could not be found")
         })?
-        .set_content_type(mime)
+        .set_content_type(mime.clone())
         .set_content_disposition(content_disposition);
 
     let mut response = file.into_response(req);

src/main.rs

@@ -2,6 +2,7 @@ mod config;
 mod db;
 mod deleter;
 mod download;
+mod mime_relations;
 mod multipart;
 mod rate_limit;
 mod template;

src/mime_relations.rs

@@ -0,0 +1,48 @@
+use std::{collections::HashMap, str::FromStr};
+
+use lazy_static::lazy_static;
+use mime::Mime;
+
+lazy_static! {
+    static ref ALIASES: HashMap<Mime, Mime> = get_mime_aliases();
+    static ref PARENTS: Vec<(Mime, Mime)> = get_mime_parent_relations();
+}
+
+fn get_mime_aliases() -> HashMap<Mime, Mime> {
+    tree_magic_db::aliases()
+        .lines()
+        .flat_map(|line| line.split_once(' '))
+        .flat_map(|(a, b)| Some((Mime::from_str(a).ok()?, Mime::from_str(b).ok()?)))
+        .collect()
+}
+
+pub(crate) fn get_alias(mimetype: &Mime) -> &Mime {
+    match ALIASES.get(mimetype) {
+        Some(x) => x,
+        None => mimetype,
+    }
+}
+
+fn get_mime_parent_relations() -> Vec<(Mime, Mime)> {
+    tree_magic_db::subclasses()
+        .lines()
+        .flat_map(|line| line.split_once(' '))
+        .flat_map(|(child, parent)| {
+            Some((Mime::from_str(child).ok()?, Mime::from_str(parent).ok()?))
+        })
+        .collect()
+}
+
+fn get_mime_parents(mimetype: &Mime) -> Vec<&Mime> {
+    PARENTS
+        .iter()
+        .filter_map(|(child, parent)| (child == mimetype).then_some(parent))
+        .collect()
+}
+
+pub(crate) fn matches_text(mime: &Mime) -> bool {
+    if mime.type_() == mime::TEXT {
+        return true;
+    }
+    return get_mime_parents(mime).into_iter().any(matches_text);
+}