From 6225b13e18d6ce5b42aa7112fa44a67b13b3bb41 Mon Sep 17 00:00:00 2001
From: Ben Pfaff <blp@cs.stanford.edu>
Date: Thu, 3 Aug 2023 08:44:54 -0700
Subject: [PATCH] work

---
 doc/dev/system-file-format.texi |  6 +++---
 rust/src/encoding.rs            | 27 +++++++++++++++++++++++++++
 rust/src/raw.rs                 | 14 ++++++++++++++
 3 files changed, 44 insertions(+), 3 deletions(-)

diff --git a/doc/dev/system-file-format.texi b/doc/dev/system-file-format.texi
index e7bda48d56..db8af7e374 100644
--- a/doc/dev/system-file-format.texi
+++ b/doc/dev/system-file-format.texi
@@ -714,7 +714,7 @@ been actually observed in system files:
 
 @table @asis
 @item 1
-EBCDIC.
+EBCDIC.  Only one example has been observed.
 
 @item 2
 7-bit ASCII.  Old versions of SPSS for Unix and Windows always wrote
@@ -732,10 +732,10 @@ ISO 8859-1 (IBM AIX code page number).
 The @code{windows-874} code page for Thai.
 
 @item 932
-The @code{windows-932} code page for Japanese.
+The @code{windows-932} code page for Japanese (aka @code{Shift_JIS}).
 
 @item 936
-The @code{windows-936} code page for simplified Chinese
+The @code{windows-936} code page for simplified Chinese (aka @code{GBK}).
 
 @item 949
 Probably @code{ks_c_5601-1987}, Unified Hangul Code.
diff --git a/rust/src/encoding.rs b/rust/src/encoding.rs
index a0e28af771..296e4e65a8 100644
--- a/rust/src/encoding.rs
+++ b/rust/src/encoding.rs
@@ -1 +1,28 @@
 include!(concat!(env!("OUT_DIR"), "/encodings.rs"));
+
+pub fn codepage_from_encoding(encoding: &str) -> Option<u32> {
+    CODEPAGE_NAME_TO_NUMBER
+        .get(encoding.to_ascii_lowercase().as_str())
+        .copied()
+}
+
+pub fn encoding_from_hints(encoding: Option<&str>, codepage: Option<u32>) -> Option<&str> {
+    if encoding.is_some() {
+        encoding
+    } else if let Some(codepage) = codepage {
+        match codepage {
+            1 => Some("EBCDIC-US"),
+            2 | 3 => {
+                // These ostensibly mean "7-bit ASCII" and "8-bit ASCII"[sic]
+                // respectively.  However, many files have character code 2 but
+                // data which are clearly not ASCII.  Therefore, ignore these
+                // values.
+                None
+            },
+            4 => Some("MS_KANJI"),
+            _ => CODEPAGE_NUMBER_TO_NAME.get(&codepage).copied()
+        }
+    } else {
+        None
+    }
+}
diff --git a/rust/src/raw.rs b/rust/src/raw.rs
index 3f7309c7ce..e017f74ac4 100644
--- a/rust/src/raw.rs
+++ b/rust/src/raw.rs
@@ -824,6 +824,20 @@ impl ExtensionRecord for FloatInfo {
     }
 }
 
+pub struct Encoding(pub String);
+
+impl ExtensionRecord for Encoding {
+    const SIZE: Option<u32> = Some(1);
+    const COUNT: Option<u32> = None;
+    const NAME: &'static str = "encoding record";
+
+    fn parse(ext: &Extension, endian: Endian) -> Result<Self, Error>{
+        ext.check_size::<Self>()?;
+
+        Ok(Encoding(String::from_utf8(ext.data)?))
+    }
+}
+
 pub struct Extension {
     /// Offset from the start of the file to the start of the record.
     pub offset: u64,
-- 
2.30.2