compiler: change s.line_nr in just one place in tandem with s.last_nl_pos

* compiler: change s.line_nr in just one place, so that s.last_nl_pos will be updated in tandem too.

* v test v: run repl tests again

* Show gcc version in both windows gcc cases in a) github actions, and b) travis .

* adding inline to is_name_char is_nl and contains_capital does not help actually, just increases slightly binary size.

* Cleanup spurious spaces.

.github/workflows/ci.yml

@@ -64,6 +64,7 @@ jobs:
     #    node-version: 12.x
     - name: Build
       run: |
+        gcc --version
         git clone --depth=1 https://github.com/ubawurinna/freetype-windows-binaries.git thirdparty/freetype/
         .\make.bat -gcc
     - name: Test

.travis.yml

@@ -43,6 +43,7 @@ script:
       git clone --depth=1 https://github.com/ubawurinna/freetype-windows-binaries  thirdparty/freetype/
 
       if [[ "${TRAVIS_JOB_NAME}" == "windows_gcc" ]]; then
+        gcc --version
         echo "Building V with GCC"
         export VFLAGS="-os windows"
         ./make.bat -gcc

compiler/main.v

@@ -1057,9 +1057,6 @@ fn (v &V) test_v() {
 	println('Testing...')
 	mut tmark := benchmark.new_benchmark()
 	for dot_relative_file in test_files {		
-		if dot_relative_file.contains('repl_test.v') {
-			continue
-		}
 		relative_file := dot_relative_file.replace('./', '')
 		file := os.realpath( relative_file )
 		tmpcfilepath := file.replace('_test.v', '_test.tmp.c')

compiler/scanner.v

@@ -74,20 +74,28 @@ fn new_scanner(text string) &Scanner {
 	}
 }
 
+
+// The goal of ScannerPos is to track the current scanning position,
+// so that if there is an error found later, v could show a more accurate 
+// position about where the error initially was.
+// NB: The fields of ScannerPos *should be kept synchronized* with the 
+// corresponding fields in Scanner.
 struct ScannerPos {
 mut:
    pos int
    line_nr int
+   last_nl_pos int
 }
 fn (s ScannerPos) str() string {
-	return 'ScannerPos{ ${s.pos:5d} , ${s.line_nr:5d} }'
+	return 'ScannerPos{ ${s.pos:5d} , ${s.line_nr:5d} , ${s.last_nl_pos:5d} }'
 }
 fn (s &Scanner) get_scanner_pos() ScannerPos {
-	return ScannerPos{ pos: s.pos line_nr: s.line_nr }
+	return ScannerPos{ pos: s.pos line_nr: s.line_nr last_nl_pos: s.last_nl_pos }
 }
 fn (s mut Scanner) goto_scanner_position(scp ScannerPos) {
 	s.pos = scp.pos
 	s.line_nr = scp.line_nr
+	s.last_nl_pos = scp.last_nl_pos
 }
 
 
@@ -231,9 +239,8 @@ fn (s mut Scanner) ident_number() string {
 fn (s mut Scanner) skip_whitespace() {
 	for s.pos < s.text.len && s.text[s.pos].is_white() {
 		// Count \r\n as one line
-		if is_nl(s.text[s.pos]) && !s.expect('\r\n', s.pos-1) {
+		if is_nl(s.text[s.pos]) && !s.expect('\r\n', s.pos-1) {  
-			s.last_nl_pos = s.pos
+			s.inc_line_number()
-			s.line_nr++
 		}
 		s.pos++
 	}
@@ -431,6 +438,7 @@ fn (s mut Scanner) scan() ScanRes {
 	case `\r`:
 		if nextc == `\n` {
 			s.pos++
+			s.last_nl_pos = s.pos
 			return scan_res(.nl, '')
 		}
 	case `\n`:
@@ -444,10 +452,7 @@ fn (s mut Scanner) scan() ScanRes {
 		return scan_res(.dot, '')
 	case `#`:
 		start := s.pos + 1
-		for s.pos < s.text.len && s.text[s.pos] != `\n` {
+		s.ignore_line()
-			s.pos++
-		}
-		s.line_nr++
 		if nextc == `!` {
 			// treat shebang line (#!) as a comment
 			s.line_comment = s.text.substr(start + 1, s.pos).trim_space()
@@ -543,10 +548,7 @@ fn (s mut Scanner) scan() ScanRes {
 		}
 		if nextc == `/` {
 			start := s.pos + 1
-			for s.pos < s.text.len && s.text[s.pos] != `\n`{
+			s.ignore_line()
-				s.pos++
-			}
-			s.line_nr++
 			s.line_comment = s.text.substr(start + 1, s.pos)
 			s.line_comment = s.line_comment.trim_space()
 			s.fgenln('// ${s.prev_tok.str()} "$s.line_comment"')
@@ -565,7 +567,7 @@ fn (s mut Scanner) scan() ScanRes {
 					s.error('comment not terminated')
 				}
 				if s.text[s.pos] == `\n` {
-					s.line_nr++
+					s.inc_line_number()
 					continue
 				}
 				if s.expect('/*', s.pos) {
@@ -711,7 +713,7 @@ fn (s mut Scanner) ident_string() string {
 			break
 		}
 		if c == `\n` {
-			s.line_nr++
+			s.inc_line_number()
 		}
 		// Don't allow \0
 		if c == `0` && s.pos > 2 && s.text[s.pos - 1] == `\\` {
@@ -827,6 +829,23 @@ fn (s mut Scanner) debug_tokens() {
 	}
 }
 
+
+fn (s mut Scanner) ignore_line() {
+	s.eat_to_end_of_line()
+	s.inc_line_number()
+}
+
+fn (s mut Scanner) eat_to_end_of_line(){
+	for s.pos < s.text.len && s.text[s.pos] != `\n` {
+		s.pos++
+	}
+}
+
+fn (s mut Scanner) inc_line_number() {
+	s.last_nl_pos = s.pos
+	s.line_nr++
+}
+
 fn is_name_char(c byte) bool {
 	return c.is_letter() || c == `_`
 }