Merge pull request #100 from Chia-Network/20240706-modern-quotes

Change how string output is generated for atoms and strings

resources/tests/test_string_repr.clsp

@@ -0,0 +1,20 @@
+;;
+;; This program is used by the test_quote_string_generation test
+;; which checks the representation of these outputs and ensures
+;; that they're properly accepted by brun, which uses the classic
+;; chialisp tokenizer/parser.
+;;
+;; the qs macro appends the given string onto "test"
+;; the atom macro gives a quoted atom starting with test and
+;; ending with the text given in the string.
+;;
+;; modern's macro system makes these different objects.
+;;
+(mod ()
+  (include *standard-cl-23*)
+
+  (defmac qs (X) (string-append "test" X))
+  (defmac atom (X) (c 1 (string->symbol (string-append "test" X))))
+
+  (list (qs '"') (qs "'") (qs " hi") (atom '"') (atom "'") (atom "_hi"))
+  )

src/compiler/sexp.rs

@@ -372,9 +372,11 @@ pub fn decode_string(v: &[u8]) -> String {
 
 pub fn printable(a: &[u8], quoted: bool) -> bool {
     !a.iter().any(|ch| {
-        (*ch as char).is_control()
-            || !(*ch as char).is_ascii()
-            || (!quoted && ch.is_ascii_whitespace())
+        *ch < 32
+            || *ch > 126
+            || (!quoted && ((*ch as char).is_ascii_whitespace() || *ch == b'\''))
+            || *ch == b'"'
+            || *ch == b'\\'
     })
 }
 

src/tests/classic/run.rs

@@ -2429,6 +2429,41 @@ fn test_assign_cse_tricky_2() {
     assert_eq!(program, wanted_repr);
 }
 
+#[test]
+fn test_quote_string_generation() {
+    // The program run here produces a list of strings and quoted atoms that have
+    // representations that must be flattened in various ways in this test.
+    let filename = "resources/tests/test_string_repr.clsp";
+    let program = do_basic_run(&vec!["run".to_string(), filename.to_string()])
+        .trim()
+        .to_string();
+    // The proram produces this list
+    //   (list (qs '"') (qs "'") (qs " hi") (atom '"') (atom "'") (atom "_hi"))
+    //
+    // where qs puts "test" in front of the given string and atom puts test in front of the
+    // given string, converts it to an atom and quotes it so that it won't be interpreted
+    // as an identifier.
+    //
+    // in other words
+    //   (list 'test"' "test'" "test hi"
+    //     (q . (string->symbol (string-append "test" '"')))
+    //     (q . (string->symbol (string-append "test" "'")))
+    //     (q . test_hi)
+    //     )
+    //
+    // The result below shows that the strings and atoms are reproduced as expected.
+    let wanted_repr = "(4 (1 . 0x7465737422) (4 (1 . \"test'\") (4 (1 . \"test hi\") (4 (1 . 499918271522) (4 (1 . 499918271527) (4 (1 . test_hi) ()))))))";
+    assert_eq!(program, wanted_repr);
+    let brun_result = do_basic_brun(&vec!["brun".to_string(), program])
+        .trim()
+        .to_string();
+    // This shows that brun interpreted and passed through the values successfully.
+    assert_eq!(
+        brun_result,
+        "(0x7465737422 \"test'\" \"test hi\" 0x7465737422 \"test'\" \"test_hi\")"
+    );
+}
+
 #[test]
 fn test_classic_modpow() {
     let result = do_basic_brun(&vec![

src/tests/compiler/compiler.rs

@@ -4,17 +4,20 @@ use std::rc::Rc;
 use clvm_rs::allocator::Allocator;
 
 use crate::classic::clvm::__type_compatibility__::bi_one;
+use crate::classic::clvm_tools::binutils::disassemble;
 use crate::classic::clvm_tools::stages::stage_0::DefaultProgramRunner;
 use crate::compiler::clvm::run;
 use crate::compiler::compiler::{compile_file, DefaultCompilerOpts};
 use crate::compiler::comptypes::{CompileErr, CompilerOpts};
-use crate::compiler::dialect::AcceptedDialect;
+use crate::compiler::dialect::{AcceptedDialect, KNOWN_DIALECTS};
 use crate::compiler::frontend::{collect_used_names_sexp, frontend};
 use crate::compiler::rename::rename_in_cons;
 use crate::compiler::runtypes::RunFailure;
 use crate::compiler::sexp::{decode_string, enlist, parse_sexp, SExp};
 use crate::compiler::srcloc::Srcloc;
 
+use crate::tests::classic::run::do_basic_brun;
+
 const TEST_TIMEOUT: usize = 1000000;
 
 fn compile_string(content: &String) -> Result<String, CompileErr> {
@@ -2449,6 +2452,82 @@ fn test_almost_empty_lambda_gives_error() {
     assert!(format!("{res:?}").contains("Must provide at least arguments and body to lambda"));
 }
 
+#[test]
+fn test_exhaustive_chars() {
+    // Verify that we can create a program that gives the expected output using
+    // every byte value in the first, mid and last position of a value.
+    let mut substitute = vec![b'x', b'x', b'x'];
+
+    let srcloc = Srcloc::start("*extest*");
+    let make_test_program = |sub: Rc<SExp>| {
+        // (mod () (include *standard-cl-23.1*) (q . <sub>))
+        Rc::new(SExp::Cons(
+            srcloc.clone(),
+            Rc::new(SExp::Atom(srcloc.clone(), b"mod".to_vec())),
+            Rc::new(SExp::Cons(
+                srcloc.clone(),
+                Rc::new(SExp::Nil(srcloc.clone())),
+                Rc::new(SExp::Cons(
+                    srcloc.clone(),
+                    Rc::new(SExp::Cons(
+                        srcloc.clone(),
+                        Rc::new(SExp::Atom(srcloc.clone(), b"include".to_vec())),
+                        Rc::new(SExp::Cons(
+                            srcloc.clone(),
+                            Rc::new(SExp::Atom(srcloc.clone(), b"*standard-cl-23.1*".to_vec())),
+                            Rc::new(SExp::Nil(srcloc.clone())),
+                        )),
+                    )),
+                    Rc::new(SExp::Cons(
+                        srcloc.clone(),
+                        Rc::new(SExp::Cons(
+                            srcloc.clone(),
+                            Rc::new(SExp::Integer(srcloc.clone(), bi_one())),
+                            sub,
+                        )),
+                        Rc::new(SExp::Nil(srcloc.clone())),
+                    )),
+                )),
+            )),
+        ))
+    };
+
+    let runner = Rc::new(DefaultProgramRunner::new());
+
+    for i in 0..=2 {
+        for j in 0..=255 {
+            substitute[i] = j;
+
+            let sub_qe = Rc::new(SExp::QuotedString(srcloc.clone(), b'"', substitute.clone()));
+
+            let mut allocator = Allocator::new();
+            let mut opts: Rc<dyn CompilerOpts> = Rc::new(DefaultCompilerOpts::new("*extest*"));
+            let dialect = KNOWN_DIALECTS["*standard-cl-23.1*"].accepted.clone();
+            opts = opts.set_dialect(dialect);
+
+            let compiled = opts
+                .compile_program(
+                    &mut allocator,
+                    runner.clone(),
+                    make_test_program(sub_qe),
+                    &mut HashMap::new(),
+                )
+                .expect("should compile");
+
+            let compiled_output = compiled.to_string();
+            let result = do_basic_brun(&vec!["brun".to_string(), compiled_output])
+                .trim()
+                .to_string();
+
+            let classic_atom = allocator.new_atom(&substitute).expect("should work");
+            let disassembled = disassemble(&mut allocator, classic_atom, None);
+            assert_eq!(result, disassembled);
+
+            substitute[i] = b'x';
+        }
+    }
+}
+
 #[test]
 fn test_odd_hex_works() {
     let res = run_string(