diff options
| m--------- | dhall-lang | 0 | ||||
| -rw-r--r-- | dhall/build.rs | 8 | ||||
| -rw-r--r-- | dhall/src/dhall.pest.visibility | 4 | ||||
| -rw-r--r-- | dhall/src/syntax/text/parser.rs | 23 |
4 files changed, 22 insertions, 13 deletions
diff --git a/dhall-lang b/dhall-lang -Subproject 3b31dea1ba47f9de041a7288826f07a05a167ba +Subproject 464b7358545a0d5c10d0aa987fed94bedc7f516 diff --git a/dhall/build.rs b/dhall/build.rs index 70389bf..981e558 100644 --- a/dhall/build.rs +++ b/dhall/build.rs @@ -414,12 +414,20 @@ fn convert_abnf_to_pest() -> std::io::Result<()> { rules.remove("url_path"); writeln!(&mut file, "url_path = _{{ path }}")?; + // Work around some greediness issue in the grammar. rules.remove("missing"); writeln!( &mut file, r#"missing = {{ "missing" ~ !simple_label_next_char }}"# )?; + // Prefer my nice error message to illegible parse errors. + rules.remove("unicode_escape"); + writeln!( + &mut file, + r#"unicode_escape = _{{ HEXDIG{{4}} | "{{" ~ HEXDIG+ ~ "}}" }}"# + )?; + rules.remove("simple_label"); writeln!( &mut file, diff --git a/dhall/src/dhall.pest.visibility b/dhall/src/dhall.pest.visibility index 41aa833..5d3b4c8 100644 --- a/dhall/src/dhall.pest.visibility +++ b/dhall/src/dhall.pest.visibility @@ -24,6 +24,10 @@ quoted_label double_quote_chunk double_quote_escaped # unicode_escape +# unicode_suffix +# unbraced_escape +# braced_codepoint +# braced_escape double_quote_char double_quote_literal single_quote_continue diff --git a/dhall/src/syntax/text/parser.rs b/dhall/src/syntax/text/parser.rs index d9e5c3b..27e0382 100644 --- a/dhall/src/syntax/text/parser.rs +++ b/dhall/src/syntax/text/parser.rs @@ -217,19 +217,18 @@ impl DhallParser { "t" => "\t".to_owned(), // "uXXXX" or "u{XXXXX}" s => { - use std::convert::{TryFrom, TryInto}; + use std::convert::TryInto; let s = &s[1..]; let s = if &s[0..1] == "{" { &s[1..s.len() - 1] } else { - &s[0..s.len()] + s }; if s.len() > 8 { Err(input.error(format!( - "Escape sequences can't have more than 8 chars: \"{}\"", - s + "Escape sequences can't have more than 8 chars" )))? } @@ -242,12 +241,10 @@ impl DhallParser { // `s` has length 8, so `bytes` has length 4 let bytes: &[u8] = &hex::decode(s).unwrap(); let i = u32::from_be_bytes(bytes.try_into().unwrap()); - let c = char::try_from(i).unwrap(); match i { - 0xD800..=0xDFFF => { - let c_ecapsed = c.escape_unicode(); - Err(input.error(format!("Escape sequences can't contain surrogate pairs: \"{}\"", c_ecapsed)))? - } + 0xD800..=0xDFFF => Err(input.error(format!( + "Escape sequences can't contain surrogate pairs" + )))?, 0x0FFFE..=0x0FFFF | 0x1FFFE..=0x1FFFF | 0x2FFFE..=0x2FFFF @@ -264,12 +261,12 @@ impl DhallParser { | 0xDFFFE..=0xDFFFF | 0xEFFFE..=0xEFFFF | 0xFFFFE..=0xFFFFF - | 0x10_FFFE..=0x10_FFFF => { - let c_ecapsed = c.escape_unicode(); - Err(input.error(format!("Escape sequences can't contain non-characters: \"{}\"", c_ecapsed)))? - } + | 0x10_FFFE..=0x10_FFFF => Err(input.error(format!( + "Escape sequences can't contain non-characters" + )))?, _ => {} } + let c: char = i.try_into().unwrap(); std::iter::once(c).collect() } }) |