1 files changed, 55 insertions, 30 deletions

diff --git a/src/pkg/unicode/maketables.go b/src/pkg/unicode/maketables.go
index 07b931d7e..b586bc655 100644
--- a/src/pkg/unicode/maketables.go
+++ b/src/pkg/unicode/maketables.go
@@ -14,6 +14,7 @@ import (
 	"http"
 	"log"
 	"os"
+	"path/filepath"
 	"sort"
 	"strconv"
 	"strings"
@@ -54,10 +55,47 @@ var cases = flag.Bool("cases",
 var test = flag.Bool("test",
 	false,
 	"test existing tables; can be used to compare web data with package data")
+var localFiles = flag.Bool("local",
+	false,
+	"data files have been copied to current directory; for debugging only")
 
 var scriptRe = regexp.MustCompile(`^([0-9A-F]+)(\.\.[0-9A-F]+)? *; ([A-Za-z_]+)$`)
 var logger = log.New(os.Stderr, "", log.Lshortfile)
 
+type reader struct {
+	*bufio.Reader
+	fd   *os.File
+	resp *http.Response
+}
+
+func open(url string) *reader {
+	file := filepath.Base(url)
+	if *localFiles {
+		fd, err := os.Open(file)
+		if err != nil {
+			logger.Fatal(err)
+		}
+		return &reader{bufio.NewReader(fd), fd, nil}
+	}
+	resp, err := http.Get(*dataURL)
+	if err != nil {
+		logger.Fatal(err)
+	}
+	if resp.StatusCode != 200 {
+		logger.Fatalf("bad GET status for %s: %d", file, resp.Status)
+	}
+	return &reader{bufio.NewReader(resp.Body), nil, resp}
+
+}
+
+func (r *reader) close() {
+	if r.fd != nil {
+		r.fd.Close()
+	} else {
+		r.resp.Body.Close()
+	}
+}
+
 var category = map[string]bool{
 	// Nd Lu etc.
 	// We use one-character names to identify merged categories
@@ -192,7 +230,7 @@ func parseCategory(line string) (state State) {
 		char.letter(field[FSimpleUppercaseMapping], field[FCodePoint], field[FSimpleTitlecaseMapping])
 	case "Lt":
 		char.letter(field[FSimpleUppercaseMapping], field[FSimpleLowercaseMapping], field[FCodePoint])
-	case "Lm", "Lo":
+	default:
 		char.letter(field[FSimpleUppercaseMapping], field[FSimpleLowercaseMapping], field[FSimpleTitlecaseMapping])
 	}
 	switch {
@@ -272,14 +310,7 @@ func loadChars() {
 	if *dataURL == "" {
 		flag.Set("data", *url+"UnicodeData.txt")
 	}
-	resp, err := http.Get(*dataURL)
-	if err != nil {
-		logger.Fatal(err)
-	}
-	if resp.StatusCode != 200 {
-		logger.Fatal("bad GET status for UnicodeData.txt", resp.Status)
-	}
-	input := bufio.NewReader(resp.Body)
+	input := open(*dataURL)
 	var first uint32 = 0
 	for {
 		line, err := input.ReadString('\n')
@@ -310,21 +341,14 @@ func loadChars() {
 			first = 0
 		}
 	}
-	resp.Body.Close()
+	input.close()
 }
 
 func loadCasefold() {
 	if *casefoldingURL == "" {
 		flag.Set("casefolding", *url+"CaseFolding.txt")
 	}
-	resp, err := http.Get(*casefoldingURL)
-	if err != nil {
-		logger.Fatal(err)
-	}
-	if resp.StatusCode != 200 {
-		logger.Fatal("bad GET status for CaseFolding.txt", resp.Status)
-	}
-	input := bufio.NewReader(resp.Body)
+	input := open(*casefoldingURL)
 	for {
 		line, err := input.ReadString('\n')
 		if err != nil {
@@ -355,7 +379,7 @@ func loadCasefold() {
 		}
 		chars[p1].foldCase = int(p2)
 	}
-	resp.Body.Close()
+	input.close()
 }
 
 const progHeader = `// Generated by running
@@ -366,7 +390,6 @@ package unicode
 
 `
 
-
 func printCategories() {
 	if *tablelist == "" {
 		return
@@ -663,15 +686,7 @@ func printScriptOrProperty(doProps bool) {
 	if flaglist == "" {
 		return
 	}
-	var err os.Error
-	resp, err := http.Get(*url + file)
-	if err != nil {
-		logger.Fatal(err)
-	}
-	if resp.StatusCode != 200 {
-		logger.Fatal("bad GET status for ", file, ":", resp.Status)
-	}
-	input := bufio.NewReader(resp.Body)
+	input := open(*url + file)
 	for {
 		line, err := input.ReadString('\n')
 		if err != nil {
@@ -682,7 +697,7 @@ func printScriptOrProperty(doProps bool) {
 		}
 		parseScript(line[0:len(line)-1], table)
 	}
-	resp.Body.Close()
+	input.close()
 
 	// Find out which scripts to dump
 	list := strings.Split(flaglist, ",")
@@ -865,6 +880,16 @@ func getCaseState(i int) (c *caseState) {
 	case ch.titleCase:
 		c._case = CaseTitle
 	}
+	// Some things such as roman numeral U+2161 don't describe themselves
+	// as upper case, but have a lower case.  Second-guess them.
+	if c._case == CaseNone && ch.lowerCase != 0 {
+		c._case = CaseUpper
+	}
+	// Same in the other direction.
+	if c._case == CaseNone && ch.upperCase != 0 {
+		c._case = CaseLower
+	}
+
 	if ch.upperCase != 0 {
 		c.deltaToUpper = ch.upperCase - i
 	}