From 151ff191cc6821f016519f1a6b75d9d24b0cf5c6 Mon Sep 17 00:00:00 2001
From: Tyler Hallada <tyler@hallada.net>
Date: Fri, 23 Jul 2021 23:14:08 -0400
Subject: [PATCH] Parse strings in Windows-1252 encoding

According to UESP, this is what strings in plugin files are encoded in: https://en.uesp.net/wiki/Skyrim_Mod:File_Format_Conventions

Decoding from this encoding fixes parsing some files with zstrings that contain special characters like umlauts.
---
 Cargo.lock    | 10 ++++++++++
 Cargo.toml    |  1 +
 src/parser.rs | 15 ++++++++-------
 3 files changed, 19 insertions(+), 7 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index 874974e..68590a8 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -88,6 +88,15 @@ dependencies = [
  "cfg-if",
 ]
 
+[[package]]
+name = "encoding_rs"
+version = "0.8.28"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "80df024fbc5ac80f87dfef0d9f5209a252f2a497f7f42944cff24d8253cac065"
+dependencies = [
+ "cfg-if",
+]
+
 [[package]]
 name = "flate2"
 version = "1.0.20"
@@ -237,6 +246,7 @@ dependencies = [
  "anyhow",
  "argh",
  "bitflags",
+ "encoding_rs",
  "flate2",
  "nom",
  "serde",
diff --git a/Cargo.toml b/Cargo.toml
index 96fe616..33b8ee8 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -14,6 +14,7 @@ license = "MIT"
 anyhow = "1.0"
 argh = { version = "0.1", optional = true }
 bitflags = "1.2"
+encoding_rs = "0.8"
 flate2 = "1.0"
 nom = "6"
 serde = { version = "1.0", features = ["derive"] }
diff --git a/src/parser.rs b/src/parser.rs
index 17c62b4..ae199fb 100644
--- a/src/parser.rs
+++ b/src/parser.rs
@@ -1,7 +1,9 @@
+use std::borrow::Cow;
 use std::io::Read;
 use std::{convert::TryInto, str};
 
 use anyhow::{anyhow, Result};
+use encoding_rs::WINDOWS_1252;
 use flate2::read::ZlibDecoder;
 use nom::{
     branch::alt,
@@ -32,9 +34,9 @@ pub struct PluginHeader<'a> {
     pub version: f32,
     pub num_records_and_groups: i32,
     pub next_object_id: u32,
-    pub author: Option<&'a str>,
-    pub description: Option<&'a str>,
-    pub masters: Vec<&'a str>,
+    pub author: Option<Cow<'a, str>>,
+    pub description: Option<Cow<'a, str>>,
+    pub masters: Vec<Cow<'a, str>>,
 }
 
 /// Parsed [CELL records](https://en.uesp.net/wiki/Skyrim_Mod:Mod_File_Format/CELL)
@@ -526,10 +528,9 @@ fn parse_4char(input: &[u8]) -> IResult<&[u8], &str> {
     map_res(take(4usize), |bytes: &[u8]| str::from_utf8(bytes))(input)
 }
 
-fn parse_zstring(input: &[u8]) -> IResult<&[u8], &str> {
-    let (input, zstring) = map_res(take_while(|byte| byte != 0), |bytes: &[u8]| {
-        str::from_utf8(bytes)
-    })(input)?;
+fn parse_zstring(input: &[u8]) -> IResult<&[u8], Cow<str>> {
+    let (input, bytes) = take_while(|byte| byte != 0)(input)?;
+    let (zstring, _, _) = WINDOWS_1252.decode(bytes);
     let (input, _) = take(1usize)(input)?;
     Ok((input, zstring))
 }