refactor(tvix/nix-compat): support non-unicode Derivations r/6449

Derivations can have non-unicode strings in their env values, so the
ATerm representations are not necessarily String anymore, but Vec<u8>.

Change-Id: Ic23839471eb7f68d9c3c30667c878830946b6607
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8990
Tested-by: BuildkiteCI
Reviewed-by: raitobezarius <tvl@lahfa.xyz>
Autosubmit: flokli <flokli@flokli.de>

1 files changed, 87 insertions, 58 deletions

diff --git a/tvix/nix-compat/src/derivation/write.rs b/tvix/nix-compat/src/derivation/write.rs
index 52166294e078..cf62f850224f 100644
--- a/tvix/nix-compat/src/derivation/write.rs
+++ b/tvix/nix-compat/src/derivation/write.rs
@@ -3,10 +3,12 @@
 //!
 //! [ATerm]: http://program-transformation.org/Tools/ATermFormat.html
 
+use crate::derivation::escape::escape_bstr;
 use crate::derivation::output::Output;
-use crate::derivation::string_escape::escape_string;
+use bstr::BString;
 use std::collections::BTreeSet;
-use std::{collections::BTreeMap, fmt, fmt::Write};
+use std::io::Cursor;
+use std::{collections::BTreeMap, io, io::Error, io::Write};
 
 pub const DERIVATION_PREFIX: &str = "Derive";
 pub const PAREN_OPEN: char = '(';
@@ -16,32 +18,46 @@ pub const BRACKET_CLOSE: char = ']';
 pub const COMMA: char = ',';
 pub const QUOTE: char = '"';
 
+// Writes a character to the writer.
+pub(crate) fn write_char(writer: &mut impl Write, c: char) -> io::Result<()> {
+    let mut buf = [0; 4];
+    let b = c.encode_utf8(&mut buf).as_bytes();
+    io::copy(&mut Cursor::new(b), writer)?;
+    Ok(())
+}
+
+// Writes a string to the writer (as unicode)
+pub(crate) fn write_str(writer: &mut impl Write, s: &str) -> io::Result<()> {
+    io::copy(&mut Cursor::new(s.as_bytes()), writer)?;
+    Ok(())
+}
+
 fn write_array_elements(
     writer: &mut impl Write,
     quote: bool,
     open: &str,
     closing: &str,
-    elements: Vec<&str>,
-) -> Result<(), fmt::Error> {
-    writer.write_str(open)?;
+    elements: &[BString],
+) -> Result<(), io::Error> {
+    write_str(writer, open)?;
 
     for (index, element) in elements.iter().enumerate() {
         if index > 0 {
-            writer.write_char(COMMA)?;
+            write_char(writer, COMMA)?;
         }
 
         if quote {
-            writer.write_char(QUOTE)?;
+            write_char(writer, QUOTE)?;
         }
 
-        writer.write_str(element)?;
+        io::copy(&mut Cursor::new(element), writer)?;
 
         if quote {
-            writer.write_char(QUOTE)?;
+            write_char(writer, QUOTE)?;
         }
     }
 
-    writer.write_str(closing)?;
+    write_str(writer, closing)?;
 
     Ok(())
 }
@@ -49,41 +65,44 @@ fn write_array_elements(
 pub fn write_outputs(
     writer: &mut impl Write,
     outputs: &BTreeMap<String, Output>,
-) -> Result<(), fmt::Error> {
-    writer.write_char(BRACKET_OPEN)?;
+) -> Result<(), io::Error> {
+    write_char(writer, BRACKET_OPEN)?;
     for (ii, (output_name, output)) in outputs.iter().enumerate() {
         if ii > 0 {
-            writer.write_char(COMMA)?;
+            write_char(writer, COMMA)?;
         }
 
-        let mut elements: Vec<&str> = vec![output_name, &output.path];
+        let mut elements: Vec<BString> = vec![
+            output_name.as_bytes().to_vec().into(),
+            output.path.as_bytes().to_vec().into(),
+        ];
 
         let (e2, e3) = match &output.hash_with_mode {
             Some(hash) => match hash {
                 crate::nixhash::NixHashWithMode::Flat(h) => (
-                    h.algo.to_string(),
-                    data_encoding::HEXLOWER.encode(&h.digest),
+                    h.algo.to_string().as_bytes().to_vec(),
+                    data_encoding::HEXLOWER.encode(&h.digest).as_bytes().into(),
                 ),
                 crate::nixhash::NixHashWithMode::Recursive(h) => (
-                    format!("r:{}", h.algo),
-                    data_encoding::HEXLOWER.encode(&h.digest),
+                    format!("r:{}", h.algo).as_bytes().to_vec(),
+                    data_encoding::HEXLOWER.encode(&h.digest).as_bytes().into(),
                 ),
             },
-            None => ("".to_string(), "".to_string()),
+            None => (vec![], vec![]),
         };
 
-        elements.push(&e2);
-        elements.push(&e3);
+        elements.push(e2.into());
+        elements.push(e3.into());
 
         write_array_elements(
             writer,
             true,
             &PAREN_OPEN.to_string(),
             &PAREN_CLOSE.to_string(),
-            elements,
+            &elements,
         )?
     }
-    writer.write_char(BRACKET_CLOSE)?;
+    write_char(writer, BRACKET_CLOSE)?;
 
     Ok(())
 }
@@ -91,33 +110,37 @@ pub fn write_outputs(
 pub fn write_input_derivations(
     writer: &mut impl Write,
     input_derivations: &BTreeMap<String, BTreeSet<String>>,
-) -> Result<(), fmt::Error> {
-    writer.write_char(COMMA)?;
-    writer.write_char(BRACKET_OPEN)?;
+) -> Result<(), io::Error> {
+    write_char(writer, COMMA)?;
+    write_char(writer, BRACKET_OPEN)?;
 
-    for (ii, (input_derivation_path, input_derivation)) in input_derivations.iter().enumerate() {
+    for (ii, (input_derivation_path, input_derivation)) in input_derivations.into_iter().enumerate()
+    {
         if ii > 0 {
-            writer.write_char(COMMA)?;
+            write_char(writer, COMMA)?;
         }
 
-        writer.write_char(PAREN_OPEN)?;
-        writer.write_char(QUOTE)?;
-        writer.write_str(input_derivation_path.as_str())?;
-        writer.write_char(QUOTE)?;
-        writer.write_char(COMMA)?;
+        write_char(writer, PAREN_OPEN)?;
+        write_char(writer, QUOTE)?;
+        write_str(writer, input_derivation_path.as_str())?;
+        write_char(writer, QUOTE)?;
+        write_char(writer, COMMA)?;
 
         write_array_elements(
             writer,
             true,
             &BRACKET_OPEN.to_string(),
             &BRACKET_CLOSE.to_string(),
-            input_derivation.iter().map(|s| &**s).collect(),
+            &input_derivation
+                .iter()
+                .map(|s| s.as_bytes().to_vec().into())
+                .collect::<Vec<BString>>(),
         )?;
 
-        writer.write_char(PAREN_CLOSE)?;
+        write_char(writer, PAREN_CLOSE)?;
     }
 
-    writer.write_char(BRACKET_CLOSE)?;
+    write_char(writer, BRACKET_CLOSE)?;
 
     Ok(())
 }
@@ -125,39 +148,45 @@ pub fn write_input_derivations(
 pub fn write_input_sources(
     writer: &mut impl Write,
     input_sources: &BTreeSet<String>,
-) -> Result<(), fmt::Error> {
-    writer.write_char(COMMA)?;
+) -> Result<(), io::Error> {
+    write_char(writer, COMMA)?;
 
     write_array_elements(
         writer,
         true,
         &BRACKET_OPEN.to_string(),
         &BRACKET_CLOSE.to_string(),
-        input_sources.iter().map(|s| &**s).collect(),
+        &input_sources
+            .iter()
+            .map(|s| s.as_bytes().to_vec().into())
+            .collect::<Vec<BString>>(),
     )?;
 
     Ok(())
 }
 
-pub fn write_system(writer: &mut impl Write, platform: &str) -> Result<(), fmt::Error> {
-    writer.write_char(COMMA)?;
-    writer.write_str(escape_string(platform).as_str())?;
+pub fn write_system(writer: &mut impl Write, platform: &str) -> Result<(), Error> {
+    write_char(writer, COMMA)?;
+    io::copy(&mut Cursor::new(escape_bstr(platform.as_bytes())), writer)?;
     Ok(())
 }
 
-pub fn write_builder(writer: &mut impl Write, builder: &str) -> Result<(), fmt::Error> {
-    writer.write_char(COMMA)?;
-    writer.write_str(escape_string(builder).as_str())?;
+pub fn write_builder(writer: &mut impl Write, builder: &str) -> Result<(), Error> {
+    write_char(writer, COMMA)?;
+    io::copy(&mut Cursor::new(escape_bstr(builder.as_bytes())), writer)?;
     Ok(())
 }
-pub fn write_arguments(writer: &mut impl Write, arguments: &[String]) -> Result<(), fmt::Error> {
-    writer.write_char(COMMA)?;
+pub fn write_arguments(writer: &mut impl Write, arguments: &[String]) -> Result<(), io::Error> {
+    write_char(writer, COMMA)?;
     write_array_elements(
         writer,
         true,
         &BRACKET_OPEN.to_string(),
         &BRACKET_CLOSE.to_string(),
-        arguments.iter().map(|s| &**s).collect(),
+        &arguments
+            .iter()
+            .map(|s| s.as_bytes().to_vec().into())
+            .collect::<Vec<BString>>(),
     )?;
 
     Ok(())
@@ -165,14 +194,14 @@ pub fn write_arguments(writer: &mut impl Write, arguments: &[String]) -> Result<
 
 pub fn write_enviroment(
     writer: &mut impl Write,
-    environment: &BTreeMap<String, String>,
-) -> Result<(), fmt::Error> {
-    writer.write_char(COMMA)?;
-    writer.write_char(BRACKET_OPEN)?;
-
-    for (ii, (key, environment)) in environment.iter().enumerate() {
-        if ii > 0 {
-            writer.write_char(COMMA)?;
+    environment: &BTreeMap<String, BString>,
+) -> Result<(), io::Error> {
+    write_char(writer, COMMA)?;
+    write_char(writer, BRACKET_OPEN)?;
+
+    for (i, (k, v)) in environment.into_iter().enumerate() {
+        if i > 0 {
+            write_char(writer, COMMA)?;
         }
 
         write_array_elements(
@@ -180,11 +209,11 @@ pub fn write_enviroment(
             false,
             &PAREN_OPEN.to_string(),
             &PAREN_CLOSE.to_string(),
-            vec![&escape_string(key), &escape_string(environment)],
+            &[escape_bstr(k.as_bytes()), escape_bstr(v)],
         )?;
     }
 
-    writer.write_char(BRACKET_CLOSE)?;
+    write_char(writer, BRACKET_CLOSE)?;
 
     Ok(())
 }