where:
        type          | fileName
        "ascii-only"  | "input-file.txt"
        "chinese"     | "输入文件.txt"
        "hungarian"   | "Dezső.txt"
        "space"       | "input file.txt"
        "zwnj"        | "input\u200cfile.txt"
        "url-quoted"  | "input%<file>#2.txt"
    }

    def "can pack trees having #type characters in name"() {
        def sourceOutputDir = temporaryFolder.file("source").createDir()

class BuildCacheEntryPackingIntegrationTest extends DaemonIntegrationSpec implements DirectoryBuildCacheFixture {

    private static final NON_ASCII_NAME = [
        "ascii-only": "ascii",
        "space": " ",
        "zwnj": "\u200c",
        "chinese": "敏捷的棕色狐狸跳过了懒狗",
        "cyrillic": "здравствуйте",
        "hungarian": "Árvíztűrő tükörfúrógép",
    ].values().join("-")

    private static final DEFAULT_ENCODINGS = [

		return &labelError{s, "V5"}
	}
	// Quickly return in the absence of zero-width (non) joiners.
	if strings.Index(s, zwj) == -1 && strings.Index(s, zwnj) == -1 {
		return nil
	}
	st := stateStart
	for i := 0; ; {
		jt := x.joinType()
		if s[i:i+sz] == zwj {
			jt = joinZWJ
		} else if s[i:i+sz] == zwnj {
			jt = joinZWNJ
		}
		st = joinStates[st][jt]
		if x.isViramaModifier() {

		return &labelError{s, "V5"}
	}
	// Quickly return in the absence of zero-width (non) joiners.
	if strings.Index(s, zwj) == -1 && strings.Index(s, zwnj) == -1 {
		return nil
	}
	st := stateStart
	for i := 0; ; {
		jt := x.joinType()
		if s[i:i+sz] == zwj {
			jt = joinZWJ
		} else if s[i:i+sz] == zwnj {
			jt = joinZWNJ
		}
		st = joinStates[st][jt]
		if x.isViramaModifier() {

        entities.put("tilde", "\u02dc");
        entities.put("ensp", "\u2002");
        entities.put("emsp", "\u2003");
        entities.put("thinsp", "\u2009");
        entities.put("zwnj", "\u200c");
        entities.put("zwj", "\u200d");
        entities.put("lrm", "\u200e");
        entities.put("rlm", "\u200f");
        entities.put("ndash", "\u2013");
        entities.put("mdash", "\u2014");

	"⇝":                         "\u21dd",
	"𝕫":                            "\U0001d56b",
	"𝓏":                            "\U0001d4cf",
	"‍":                             "\u200d",
	"‌":                            "\u200c",

        entities.put("tilde", "\u02dc");
        entities.put("ensp", "\u2002");
        entities.put("emsp", "\u2003");
        entities.put("thinsp", "\u2009");
        entities.put("zwnj", "\u200c");
        entities.put("zwj", "\u200d");
        entities.put("lrm", "\u200e");
        entities.put("rlm", "\u200f");
        entities.put("ndash", "\u2013");
        entities.put("mdash", "\u2014");

		"zigrarr;":                         '\U000021DD',
		"zopf;":                            '\U0001D56B',
		"zscr;":                            '\U0001D4CF',
		"zwj;":                             '\U0000200D',
		"zwnj;":                            '\U0000200C',
		"AElig":                            '\U000000C6',
		"AMP":                              '\U00000026',
		"Aacute":                           '\U000000C1',

	"Scaron":   "\u0160",
	"scaron":   "\u0161",
	"Yuml":     "\u0178",
	"circ":     "\u02C6",
	"tilde":    "\u02DC",
	"ensp":     "\u2002",
	"emsp":     "\u2003",
	"thinsp":   "\u2009",
	"zwnj":     "\u200C",
	"zwj":      "\u200D",
	"lrm":      "\u200E",
	"rlm":      "\u200F",
	"ndash":    "\u2013",
	"mdash":    "\u2014",
	"lsquo":    "\u2018",
	"rsquo":    "\u2019",
	"sbquo":    "\u201A",

//
// For our approximation, we group the Word Break types into the following
// categories, with associated rules:
//
// 1) Letter:
//    ALetter, Hebrew_Letter, Numeric, ExtendNumLet, Extend, Format_FE, ZWJ.
//    Rule: Never break between consecutive runes of this category.
//
// 2) Mid:
//    MidLetter, MidNumLet, Single_Quote.
//    (Cf. case-ignorable: MidLetter, MidNumLet, Single_Quote or cat is Mn,