summaryrefslogtreecommitdiff
path: root/src/tparse/parse.go
diff options
context:
space:
mode:
authorKyle Gunger <corechg@gmail.com>2020-06-28 14:30:45 -0400
committerKyle Gunger <corechg@gmail.com>2020-06-28 14:30:45 -0400
commit8f9cf0d4856bb53009bb58b53a42e21e2cd1e947 (patch)
treeb022091a0c3105e2da54b9dc16e5f55852b788f3 /src/tparse/parse.go
[Initial parser] Upload existing
Diffstat (limited to 'src/tparse/parse.go')
-rw-r--r--src/tparse/parse.go203
1 files changed, 203 insertions, 0 deletions
diff --git a/src/tparse/parse.go b/src/tparse/parse.go
new file mode 100644
index 0000000..e9e1ee6
--- /dev/null
+++ b/src/tparse/parse.go
@@ -0,0 +1,203 @@
+package tparse
+
+import (
+ "bufio"
+ "io"
+ "os"
+ "strings"
+ "unicode"
+ "unicode/utf8"
+)
+
+// Read in a number (may be a float)
+func numericLiteral(r *bufio.Reader) Token {
+ decimal := false
+ run, _, err := r.ReadRune()
+
+ b := strings.Builder{}
+
+ for ; err == nil; run, _, err = r.ReadRune() {
+ if (run == '.' || run == ',') && !decimal {
+ decimal = true
+ } else if !unicode.IsNumber(run) {
+ break
+ }
+ b.WriteRune(run)
+ }
+
+ r.UnreadRune()
+
+ return Token{Type: LITERAL, Data: b.String()}
+}
+
+// Parse a string (will escape \" only in this stage)
+func stringLiteral(r *bufio.Reader) Token {
+ escape := false
+ run, _, err := r.ReadRune()
+
+ if run != '"' {
+ return Token{Type: LITERAL}
+ }
+
+ b := strings.Builder{}
+
+ for ; err == nil; run, _, err = r.ReadRune() {
+ b.WriteRune(run)
+ if run == '\\' && !escape {
+ escape = true
+ } else if run == '"' && !escape {
+ break
+ }
+
+ }
+
+ return Token{Type: LITERAL, Data: b.String()}
+}
+
+// Parse a character in (escape \\ or \')
+func charLiteral(r *bufio.Reader) Token {
+ escape := false
+ run, _, err := r.ReadRune()
+
+ if run != '\'' {
+ return Token{Type: LITERAL}
+ }
+
+ b := strings.Builder{}
+
+ for ; err == nil; run, _, err = r.ReadRune() {
+ b.WriteRune(run)
+ if run == '\\' && !escape {
+ escape = true
+ } else if run == '\'' && !escape {
+ break
+ }
+
+ }
+
+ return Token{Type: LITERAL, Data: b.String()}
+}
+
+// Split reserved runes into rune groups
+func splitResRunes(str string, max int) []Token {
+ out := []Token{}
+
+ rs := StringAsRunes(str)
+ s, e := 0, max
+
+ if max > len(rs) {
+ e = len(rs)
+ }
+
+ for e <= len(rs) && s < len(rs) {
+ if checkRuneGroup(RunesAsString(rs[s:e])) != -1 || e == s+1 {
+ tmp := RunesAsString(rs[s:e])
+ out = append(out, Token{Type: checkRuneGroup(tmp), Data: tmp})
+ s = e
+ if s+max < len(rs) {
+ e = s + max
+ } else {
+ e = len(rs)
+ }
+ } else if e != s+1 {
+ e--
+ }
+ }
+
+ return out
+}
+
+// ParseFile tries to read a file and turn it into a series of tokens
+func ParseFile(path string) []Token {
+ out := []Token{}
+
+ fd, err := os.Open(path)
+
+ if err != nil {
+ return out
+ }
+
+ read := bufio.NewReader(fd)
+
+ b := strings.Builder{}
+
+ max := maxResRunes()
+
+ for r := ' '; ; r, _, err = read.ReadRune() {
+ // If error in stream or EOF, break
+ if err != nil {
+ if err != io.EOF {
+ out = append(out, Token{Type: -1})
+ }
+ break
+ }
+
+ // Checking for a space
+ if unicode.IsSpace(r) {
+ if b.String() != "" {
+ out = append(out, Token{Type: checkToken(b.String()), Data: b.String()})
+ b.Reset()
+ }
+ continue
+ }
+
+ // Checking for a rune group
+ if checkResRune(r) != -1 {
+ if b.String() != "" {
+ out = append(out, Token{Type: checkToken(b.String()), Data: b.String()})
+ b.Reset()
+ }
+
+ for ; err == nil; r, _, err = read.ReadRune() {
+ if checkResRune(r) == -1 {
+ break
+ }
+ b.WriteRune(r)
+ }
+
+ read.UnreadRune()
+
+ out = append(out, splitResRunes(b.String(), max)...)
+
+ b.Reset()
+
+ continue
+ }
+
+ // Accumulate
+ b.WriteRune(r)
+ }
+
+ return out
+}
+
+// StringAsRunes returns a string as a rune slice
+func StringAsRunes(s string) []rune {
+ out := []rune{}
+ for i, j := 0, 0; i < len(s); i += j {
+ r, w := utf8.DecodeRuneInString(s[i:])
+ out = append(out, r)
+ j = w
+ }
+ return out
+}
+
+// BytesAsRunes returns a byte slice as a rune slice
+func BytesAsRunes(b []byte) []rune {
+ out := []rune{}
+ for i, j := 0, 0; i < len(b); i += j {
+ r, w := utf8.DecodeRune(b[i:])
+ out = append(out, r)
+ j = w
+ }
+ return out
+}
+
+// RunesAsString returns a string from a slice of runes
+func RunesAsString(rs []rune) string {
+ b := strings.Builder{}
+ for _, r := range rs {
+ b.WriteRune(r)
+ }
+ return b.String()
+}