Add harness for running Unicode.org Bidi tests.

Source/platform/text/BidiTestHarness.h

Index: Source/platform/text/BidiTestHarness.h
diff --git a/Source/platform/text/BidiTestHarness.h b/Source/platform/text/BidiTestHarness.h
new file mode 100644
index 0000000000000000000000000000000000000000..cc2a191f68ab9811f68ff495a81fbee18e1ad6d4
--- /dev/null
+++ b/Source/platform/text/BidiTestHarness.h
@@ -0,0 +1,261 @@
+/*
+ * Copyright (C) 2013 Google Inc. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are
+ * met:
+ *
+ *     * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ *     * Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following disclaimer
+ * in the documentation and/or other materials provided with the
+ * distribution.
+ *     * Neither the name of Google Inc. nor the names of its
+ * contributors may be used to endorse or promote products derived from
+ * this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef BidiTestHarness_h
+#define BidiTestHarness_h
+
+#include <fstream>
+#include <map>
+#include <stdio.h>
+#include <string>
+
+// FIXME: We don't have any business owning this code. We should try to
+// upstream this to unicode.org if possible (for other implementations to use).
+// Unicode.org provides a reference implmentation, including parser:
+// http://www.unicode.org/Public/PROGRAMS/BidiReferenceC/6.3.0/source/brtest.c
+// But it, like the other implementations I've found, is rather tied to
+// the algorithms it is testing. This file seeks to only implement the parser bits.
+
+// Other C/C++ implementations of this parser:
+// https://github.com/googlei18n/fribidi-vs-unicode/blob/master/test.c
+// http://source.icu-project.org/repos/icu/icu/trunk/source/test/intltest/bidiconf.cpp
+// Both of those are too tied to their respective projects to be use to Blink.
+
+// There are non-C implmentations to parse BidiTest.txt as well, including:
+// https://github.com/twitter/twitter-cldr-rb/blob/master/spec/bidi/bidi_spec.rb
+
+namespace bidi_test {
+
+enum ParagraphDirection {
+    DirectionAutoLTR = 1,
+    DirectionLTR = 2,
+    DirectionRTL = 4,
+};
+const int kMaxParagraphDirection = DirectionAutoLTR | DirectionLTR | DirectionRTL;
+
+// For error printing:
+std::string nameFromParagraphDirection(ParagraphDirection paragraphDirection)
+{
+    switch (paragraphDirection) {
+    case bidi_test::DirectionAutoLTR:
+        return "Auto-LTR";
+    case bidi_test::DirectionLTR:
+        return "LTR";
+    case bidi_test::DirectionRTL:
+        return "RTL";
+    }
+}
+
+template<class Runner>
+class Harness {
+public:
+    Harness(Runner& runner)
+        : m_runner(runner)
+    {
+    }
+    void parse(std::istream& bidiTestFile);
+
+private:
+    Runner& m_runner;
+};
+
+// This trim() example comes from:
+// http://stackoverflow.com/questions/216823/whats-the-best-way-to-trim-stdstring
+// We could use boost::trim, but no other part of Blink uses boost yet.
+
+inline void ltrim(std::string& s)
+{
+    s.erase(s.begin(), std::find_if(s.begin(), s.end(), std::not1(std::ptr_fun<int, int>(std::isspace))));
+}
+
+inline void rtrim(std::string& s)
+{
+    s.erase(std::find_if(s.rbegin(), s.rend(), std::not1(std::ptr_fun<int, int>(std::isspace))).base(), s.end());
+}
+
+inline void trim(std::string& s)
+{
+    rtrim(s);
+    ltrim(s);
+}
+
+static std::vector<std::string> parseStringList(const std::string& str)
+{
+    std::vector<std::string> strings;
+    std::string seperators(" \t");
+    size_t lastPos = str.find_first_not_of(seperators); // skip leading spaces

[Jeffrey Yasskin, 2013/10/29 23:16:03]

"seperators" still isn't spelled right. :)

[eseidel, 2013/10/30 23:04:11]

I checked a dictionary this time. separators :)

+    size_t pos = str.find_first_of(seperators, lastPos); // find next space
+
+    while (std::string::npos != pos || std::string::npos != lastPos) {
+        strings.push_back(str.substr(lastPos, pos - lastPos));
+        lastPos = str.find_first_not_of(seperators, pos);
+        pos = str.find_first_of(seperators, lastPos);
+    }
+    return strings;
+}
+
+static std::vector<int> parseIntList(const std::string& str)
+{
+    std::vector<int> ints;
+    std::vector<std::string> strings = parseStringList(str);
+    for (int x = 0; x < strings.size(); x++) {
+        int i = atoi(strings[x].c_str());
+        ints.push_back(i);
+    }
+    return ints;
+}
+
+static std::vector<int> parseLevels(const std::string& line)
+{
+    std::vector<int> levels;
+    std::vector<std::string> strings = parseStringList(line);
+    for (int x = 0; x < strings.size(); x++) {
+        const std::string& levelString = strings[x];
+        int i;
+        if (levelString == "x")
+            i = -1;
+        else
+            i = atoi(levelString.c_str());
+        levels.push_back(i);
+    }
+    return levels;
+}
+
+static std::basic_string<UChar> parseTestString(const std::string& line)
+{
+    std::basic_string<UChar> testString;
+    // This static is not thread-safe, but currently that's not an issue.

[Jeffrey Yasskin, 2013/10/29 23:16:03]

Oh, I meant you should comment the function as not thread-safe, since callers
have to be aware of that.

[eseidel, 2013/10/30 23:04:11]

Done.

+    static std::map<std::string, UChar> charClassExamples;
+    if (charClassExamples.empty()) {
+        charClassExamples.insert({"L", 0x6c}); // 'l' for L
+        charClassExamples.insert({"R", 0x05D0}); // HEBREW ALEF
+        charClassExamples.insert({"EN", 0x33}); // '3' for EN
+        charClassExamples.insert({"ES", 0x2d}); // '-' for ES
+        charClassExamples.insert({"ET", 0x25}); // '%' for ET
+        charClassExamples.insert({"AN", 0x0660}); // arabic 0
+        charClassExamples.insert({"CS", 0x2c}); // ',' for CS
+        charClassExamples.insert({"B", 0x0A}); // <control-000A>
+        charClassExamples.insert({"S", 0x09}); // <control-0009>
+        charClassExamples.insert({"WS", 0x20}); // ' ' for WS
+        charClassExamples.insert({"ON", 0x3d}); // '=' for ON
+        charClassExamples.insert({"NSM", 0x05BF}); // HEBREW POINT RAFE
+        charClassExamples.insert({"AL", 0x0608}); // ARABIC RAY
+        charClassExamples.insert({"BN", 0x00AD}); // SOFT HYPHEN
+        charClassExamples.insert({"LRE", 0x202A});
+        charClassExamples.insert({"RLE", 0x202B});
+        charClassExamples.insert({"PDF", 0x202C});
+        charClassExamples.insert({"LRO", 0x202D});
+        charClassExamples.insert({"RLO", 0x202E});
+        charClassExamples.insert({"LRI", 0x2066});
+        charClassExamples.insert({"RLI", 0x2067});
+        charClassExamples.insert({"FSI", 0x2068});
+        charClassExamples.insert({"PDI", 0x2069});
+    }
+
+    std::vector<std::string> charClasses = parseStringList(line);
+    for (int i = 0; i < charClasses.size(); i++) {
+        // FIXME: If the lookup failed we could return false for a parse error.
+        testString.push_back(charClassExamples.find(charClasses[i])->second);
+    }
+    return testString;
+}
+
+static bool parseParagraphDirectionMask(const std::string& line, int& modeMask)
+{
+    modeMask = atoi(line.c_str());
+    return modeMask >= 1 && modeMask <= kMaxParagraphDirection;
+}
+
+static void parseError(const std::string& line, size_t lineNumber)
+{
+    // Use printf to avoid the expense of std::cout.
+    printf("Parse error, line %d : %s\n", lineNumber, line.c_str());
+}
+
+template<class Runner>
+void Harness<Runner>::parse(std::istream& bidiTestFile)
+{
+    static const std::string levelsPrefix("@Levels");
+    static const std::string reorderPrefix("@Reorder");
+
+    // FIXME: UChar is an ICU type and cheating a bit to use here.
+    // uint16_t might be more portable.
+    std::basic_string<UChar> testString;
+    std::vector<int> levels;
+    std::vector<int> reorder;
+    int paragraphDirectionMask;
+
+    std::string originalLine;
+    std::string line;
+    size_t lineNumber = 0;
+    while (std::getline(bidiTestFile, originalLine)) {
+        lineNumber++;
+        line = originalLine;
+        size_t commentStart = originalLine.find_first_of('#');
+        if (commentStart != std::string::npos)
+            line = line.substr(0, commentStart);
+        trim(line);
+        if (line.empty())
+            continue;
+        if (line[0] == '@') {
+            if (!line.find(levelsPrefix)) {
+                levels = parseLevels(line.substr(levelsPrefix.length() + 1));

[Jeffrey Yasskin, 2013/10/29 23:16:03]

Note that line.substr makes a copy of the data. You can't do any better in just
the standard library, since we don't have a string_ref class, but if performance
turns out to matter, it may be worth writing one yourself.

[eseidel, 2013/10/30 23:04:11]

Performance only sorta matters for this code.  Ideally we'd make it fast if it
were to be used by many projects, but just being used for one unittest it
doesn't really matter that that unittest currently takes 20s to run on my
machine. :(

+                continue;
+            }
+            if (!line.find(reorderPrefix)) {
+                reorder = parseIntList(line.substr(reorderPrefix.length() + 1));
+                continue;
+            }
+        } else {
+            // Assume it's a data line.
+            size_t seperatorIndex = line.find_first_of(';');
+            if (seperatorIndex == std::string::npos) {
+                parseError(originalLine, lineNumber);
+                continue;
+            }
+            testString = parseTestString(line.substr(0, seperatorIndex));
+            if (!parseParagraphDirectionMask(line.substr(seperatorIndex + 1), paragraphDirectionMask)) {
+                parseError(originalLine, lineNumber);
+                continue;
+            }
+
+            if (paragraphDirectionMask & DirectionAutoLTR)
+                m_runner.runTest(testString, reorder, levels, DirectionAutoLTR, originalLine, lineNumber);
+            if (paragraphDirectionMask & DirectionLTR)
+                m_runner.runTest(testString, reorder, levels, DirectionLTR, originalLine, lineNumber);
+            if (paragraphDirectionMask & DirectionRTL)
+                m_runner.runTest(testString, reorder, levels, DirectionRTL, originalLine, lineNumber);
+        }
+    }
+}
+
+} // namespace bidi_test
+
+#endif // BidiTestHarness_h