Add line numbers to sup parsing errors

runtime/ztests/expr/function/parse-sup.yaml

@@ -12,4 +12,4 @@ output: |
   {a:1}
   null
   error({message:"parse_sup: string arg required",on:{}})
-  error({message:"parse_sup: SUP syntax error",on:"!"})
+  error({message:"parse_sup: line 1: parse error: syntax error",on:"!"})

service/ztests/curl-load-error.yaml

@@ -16,7 +16,7 @@ inputs:
 outputs:
   - name: stdout
     data: |
-      {"type":"Error","kind":"invalid operation","error":"format detection error\n\tarrows: schema message length exceeds 1 MiB\n\tbsup: malformed BSUP value\n\tcsup: auto-detection requires seekable input\n\tcsv: line 1: EOF\n\tjson: invalid character 'T' looking for beginning of value\n\tline: auto-detection not supported\n\tparquet: auto-detection requires seekable input\n\tsup: SUP syntax error\n\ttsv: line 1: EOF\n\tzeek: line 1: bad types/fields definition in zeek header\n\tjsup: line 1: malformed JSUP: bad type object: \"This is not a detectable format.\": unpacker error parsing JSON: invalid character 'T' looking for beginning of value"}
+      {"type":"Error","kind":"invalid operation","error":"format detection error\n\tarrows: schema message length exceeds 1 MiB\n\tbsup: malformed BSUP value\n\tcsup: auto-detection requires seekable input\n\tcsv: line 1: EOF\n\tjson: invalid character 'T' looking for beginning of value\n\tline: auto-detection not supported\n\tparquet: auto-detection requires seekable input\n\tsup: line 1: parse error: syntax error\n\ttsv: line 1: EOF\n\tzeek: line 1: bad types/fields definition in zeek header\n\tjsup: line 1: malformed JSUP: bad type object: \"This is not a detectable format.\": unpacker error parsing JSON: invalid character 'T' looking for beginning of value"}
       code 400
       {"type":"Error","kind":"invalid operation","error":"unsupported MIME type: unsupported"}
       code 400

service/ztests/load-garbage.yaml

@@ -20,7 +20,7 @@ outputs:
       	json: invalid character 'T' looking for beginning of value
       	line: auto-detection not supported
       	parquet: auto-detection requires seekable input
-      	sup: SUP syntax error
+      	sup: line 1: parse error: syntax error
       	tsv: line 1: delimiter '\t' not found
       	zeek: line 1: bad types/fields definition in zeek header
       	jsup: line 1: malformed JSUP: bad type object: "This file contains no records.": unpacker error parsing JSON: invalid character 'T' looking for beginning of value

sup/lexer.go

@@ -25,6 +25,7 @@ type Lexer struct {
 	primitive   *regexp.Regexp
 	indentation *regexp.Regexp
 	ip6follow   *regexp.Regexp
+	line        int
 }
 
 const (
@@ -43,6 +44,7 @@ func NewLexer(r io.Reader) *Lexer {
 		primitive:   primitive,
 		indentation: indentation,
 		ip6follow:   regexp.MustCompile(ip6followRE),
+		line:        1,
 	}
 }
 
@@ -90,6 +92,7 @@ func (l *Lexer) skip(n int) error {
 	if err := l.check(n); err != nil {
 		return err
 	}
+	l.line += bytes.Count(l.cursor[:n], []byte{'\n'})
 	l.cursor = l.cursor[n:]
 	return nil
 }
@@ -170,6 +173,9 @@ func (l *Lexer) readByte() (byte, error) {
 	}
 	b := l.cursor[0]
 	l.cursor = l.cursor[1:]
+	if b == '\n' {
+		l.line++
+	}
 	return b, nil
 }
 

sup/parser-values.go

@@ -2,7 +2,6 @@ package sup
 
 import (
 	"encoding/hex"
-	"errors"
 	"fmt"
 	"io"
 	"net/netip"
@@ -22,7 +21,7 @@ func (p *Parser) ParseValue() (ast.Value, error) {
 	}
 	if v == nil && err == nil {
 		if err := p.lexer.check(1); (err != nil && err != io.EOF) || len(p.lexer.cursor) > 0 {
-			return nil, errors.New("SUP syntax error")
+			return nil, p.error("syntax error")
 		}
 	}
 	return v, err

sup/parser.go

@@ -18,7 +18,5 @@ func (p *Parser) errorf(msg string, args ...any) error {
 }
 
 func (p *Parser) error(msg string) error {
-	// format a message based on the contents in the scanner buffer
-	// (could also track column and line number)
-	return fmt.Errorf("parse error: %s", msg)
+	return fmt.Errorf("line %d: parse error: %s", p.lexer.line, msg)
 }

sup/ztests/parser-errors.yaml

@@ -0,0 +1,36 @@
+# Test parse errors with unescaped quotes.
+spq: pass
+
+input: |
+  {id:1,text:"valid record before the bad one"}
+  {id:2,text:"this has "unescaped quotes" inside it"}
+  {id:3,text:"valid record after the bad one"}
+
+error: |
+  line 2: parse error: mismatched braces while parsing record type
+
+---
+
+# Unclosed object.
+spq: pass
+
+input: |
+  {x:1}
+  {x:2}
+  {x:3
+  {x:4}
+
+error: |
+  line 4: parse error: mismatched braces while parsing record type
+
+---
+
+spq: pass
+
+input-flags: -i sup
+
+input: |
+  {x}
+
+error: |
+  line 1: parse error: no type name found for field "x"