Skip to content

Commit a409afa

Browse files
committed
[support] GlobPattern: add support for \ and [!...], and allow ] in more places
Summary: Update GlobPattern in libSupport to handle a few more cases. It does not fully match the `fnmatch` used by GNU objcopy since named character classes (e.g. `[[:digit:]]`) are not supported, but this should support most existing use cases (mostly just `*` is what's used anyway). This will be used to implement the `--wildcard` flag in llvm-objcopy to be more compatible with GNU objcopy. This is split off of D66613 to land the libSupport changes separately. The llvm-objcopy part will land soon. Reviewers: jhenderson, MaskRay, evgeny777, espindola, alexshap Reviewed By: MaskRay Subscribers: nickdesaulniers, emaste, arichardson, hiraditya, jakehehrlich, abrachet, seiya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D66613 undo objcopy changes to make this libsupport only llvm-svn: 375051
1 parent 4eb1a57 commit a409afa

File tree

3 files changed

+113
-46
lines changed

3 files changed

+113
-46
lines changed

llvm/include/llvm/Support/GlobPattern.h

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -21,7 +21,7 @@
2121
#include <vector>
2222

2323
// This class represents a glob pattern. Supported metacharacters
24-
// are "*", "?", "[<chars>]" and "[^<chars>]".
24+
// are "*", "?", "\", "[<chars>]", "[^<chars>]", and "[!<chars>]".
2525
namespace llvm {
2626
class BitVector;
2727
template <typename T> class ArrayRef;

llvm/lib/Support/GlobPattern.cpp

Lines changed: 16 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -19,7 +19,7 @@
1919
using namespace llvm;
2020

2121
static bool hasWildcard(StringRef S) {
22-
return S.find_first_of("?*[") != StringRef::npos;
22+
return S.find_first_of("?*[\\") != StringRef::npos;
2323
}
2424

2525
// Expands character ranges and returns a bitmap.
@@ -60,8 +60,9 @@ static Expected<BitVector> expand(StringRef S, StringRef Original) {
6060
}
6161

6262
// This is a scanner for the glob pattern.
63-
// A glob pattern token is one of "*", "?", "[<chars>]", "[^<chars>]"
64-
// (which is a negative form of "[<chars>]"), or a non-meta character.
63+
// A glob pattern token is one of "*", "?", "\", "[<chars>]", "[^<chars>]"
64+
// (which is a negative form of "[<chars>]"), "[!<chars>]" (which is
65+
// equivalent to "[^<chars>]"), or a non-meta character.
6566
// This function returns the first token in S.
6667
static Expected<BitVector> scan(StringRef &S, StringRef Original) {
6768
switch (S[0]) {
@@ -74,21 +75,28 @@ static Expected<BitVector> scan(StringRef &S, StringRef Original) {
7475
S = S.substr(1);
7576
return BitVector(256, true);
7677
case '[': {
77-
size_t End = S.find(']', 1);
78+
// ']' is allowed as the first character of a character class. '[]' is
79+
// invalid. So, just skip the first character.
80+
size_t End = S.find(']', 2);
7881
if (End == StringRef::npos)
7982
return make_error<StringError>("invalid glob pattern: " + Original,
8083
errc::invalid_argument);
8184

8285
StringRef Chars = S.substr(1, End - 1);
8386
S = S.substr(End + 1);
84-
if (Chars.startswith("^")) {
87+
if (Chars.startswith("^") || Chars.startswith("!")) {
8588
Expected<BitVector> BV = expand(Chars.substr(1), Original);
8689
if (!BV)
8790
return BV.takeError();
8891
return BV->flip();
8992
}
9093
return expand(Chars, Original);
9194
}
95+
case '\\':
96+
// Eat this character and fall through below to treat it like a non-meta
97+
// character.
98+
S = S.substr(1);
99+
LLVM_FALLTHROUGH;
92100
default:
93101
BitVector BV(256, false);
94102
BV[(uint8_t)S[0]] = true;
@@ -107,8 +115,9 @@ Expected<GlobPattern> GlobPattern::create(StringRef S) {
107115
return Pat;
108116
}
109117

110-
// S is something like "foo*". We can use startswith().
111-
if (S.endswith("*") && !hasWildcard(S.drop_back())) {
118+
// S is something like "foo*", and the "* is not escaped. We can use
119+
// startswith().
120+
if (S.endswith("*") && !S.endswith("\\*") && !hasWildcard(S.drop_back())) {
112121
Pat.Prefix = S.drop_back();
113122
return Pat;
114123
}

llvm/unittests/Support/GlobPatternTest.cpp

Lines changed: 96 additions & 38 deletions
Original file line numberDiff line numberDiff line change
@@ -14,57 +14,115 @@ namespace {
1414

1515
class GlobPatternTest : public ::testing::Test {};
1616

17-
TEST_F(GlobPatternTest, Basics) {
17+
TEST_F(GlobPatternTest, Empty) {
1818
Expected<GlobPattern> Pat1 = GlobPattern::create("");
1919
EXPECT_TRUE((bool)Pat1);
2020
EXPECT_TRUE(Pat1->match(""));
2121
EXPECT_FALSE(Pat1->match("a"));
22+
}
23+
24+
TEST_F(GlobPatternTest, Glob) {
25+
Expected<GlobPattern> Pat1 = GlobPattern::create("ab*c*def");
26+
EXPECT_TRUE((bool)Pat1);
27+
EXPECT_TRUE(Pat1->match("abcdef"));
28+
EXPECT_TRUE(Pat1->match("abxcxdef"));
29+
EXPECT_FALSE(Pat1->match(""));
30+
EXPECT_FALSE(Pat1->match("xabcdef"));
31+
EXPECT_FALSE(Pat1->match("abcdefx"));
32+
}
33+
34+
TEST_F(GlobPatternTest, Wildcard) {
35+
Expected<GlobPattern> Pat1 = GlobPattern::create("a??c");
36+
EXPECT_TRUE((bool)Pat1);
37+
EXPECT_TRUE(Pat1->match("axxc"));
38+
EXPECT_FALSE(Pat1->match("axxx"));
39+
EXPECT_FALSE(Pat1->match(""));
40+
}
2241

23-
Expected<GlobPattern> Pat2 = GlobPattern::create("ab*c*def");
42+
TEST_F(GlobPatternTest, Escape) {
43+
Expected<GlobPattern> Pat1 = GlobPattern::create("\\*");
44+
EXPECT_TRUE((bool)Pat1);
45+
EXPECT_TRUE(Pat1->match("*"));
46+
EXPECT_FALSE(Pat1->match("\\*"));
47+
EXPECT_FALSE(Pat1->match("a"));
48+
49+
Expected<GlobPattern> Pat2 = GlobPattern::create("a?\\?c");
2450
EXPECT_TRUE((bool)Pat2);
25-
EXPECT_TRUE(Pat2->match("abcdef"));
26-
EXPECT_TRUE(Pat2->match("abxcxdef"));
51+
EXPECT_TRUE(Pat2->match("ax?c"));
52+
EXPECT_FALSE(Pat2->match("axxc"));
2753
EXPECT_FALSE(Pat2->match(""));
28-
EXPECT_FALSE(Pat2->match("xabcdef"));
29-
EXPECT_FALSE(Pat2->match("abcdefx"));
30-
31-
Expected<GlobPattern> Pat3 = GlobPattern::create("a??c");
32-
EXPECT_TRUE((bool)Pat3);
33-
EXPECT_TRUE(Pat3->match("axxc"));
34-
EXPECT_FALSE(Pat3->match("axxx"));
35-
EXPECT_FALSE(Pat3->match(""));
36-
37-
Expected<GlobPattern> Pat4 = GlobPattern::create("[abc-fy-z]");
38-
EXPECT_TRUE((bool)Pat4);
39-
EXPECT_TRUE(Pat4->match("a"));
40-
EXPECT_TRUE(Pat4->match("b"));
41-
EXPECT_TRUE(Pat4->match("c"));
42-
EXPECT_TRUE(Pat4->match("d"));
43-
EXPECT_TRUE(Pat4->match("e"));
44-
EXPECT_TRUE(Pat4->match("f"));
45-
EXPECT_TRUE(Pat4->match("y"));
46-
EXPECT_TRUE(Pat4->match("z"));
47-
EXPECT_FALSE(Pat4->match("g"));
48-
EXPECT_FALSE(Pat4->match(""));
49-
50-
Expected<GlobPattern> Pat5 = GlobPattern::create("[^abc-fy-z]");
51-
EXPECT_TRUE((bool)Pat5);
52-
EXPECT_TRUE(Pat5->match("g"));
53-
EXPECT_FALSE(Pat5->match("a"));
54-
EXPECT_FALSE(Pat5->match("b"));
55-
EXPECT_FALSE(Pat5->match("c"));
56-
EXPECT_FALSE(Pat5->match("d"));
57-
EXPECT_FALSE(Pat5->match("e"));
58-
EXPECT_FALSE(Pat5->match("f"));
59-
EXPECT_FALSE(Pat5->match("y"));
60-
EXPECT_FALSE(Pat5->match("z"));
61-
EXPECT_FALSE(Pat5->match(""));
54+
}
55+
56+
TEST_F(GlobPatternTest, BasicCharacterClass) {
57+
Expected<GlobPattern> Pat1 = GlobPattern::create("[abc-fy-z]");
58+
EXPECT_TRUE((bool)Pat1);
59+
EXPECT_TRUE(Pat1->match("a"));
60+
EXPECT_TRUE(Pat1->match("b"));
61+
EXPECT_TRUE(Pat1->match("c"));
62+
EXPECT_TRUE(Pat1->match("d"));
63+
EXPECT_TRUE(Pat1->match("e"));
64+
EXPECT_TRUE(Pat1->match("f"));
65+
EXPECT_TRUE(Pat1->match("y"));
66+
EXPECT_TRUE(Pat1->match("z"));
67+
EXPECT_FALSE(Pat1->match("g"));
68+
EXPECT_FALSE(Pat1->match(""));
69+
}
70+
71+
TEST_F(GlobPatternTest, NegatedCharacterClass) {
72+
Expected<GlobPattern> Pat1 = GlobPattern::create("[^abc-fy-z]");
73+
EXPECT_TRUE((bool)Pat1);
74+
EXPECT_TRUE(Pat1->match("g"));
75+
EXPECT_FALSE(Pat1->match("a"));
76+
EXPECT_FALSE(Pat1->match("b"));
77+
EXPECT_FALSE(Pat1->match("c"));
78+
EXPECT_FALSE(Pat1->match("d"));
79+
EXPECT_FALSE(Pat1->match("e"));
80+
EXPECT_FALSE(Pat1->match("f"));
81+
EXPECT_FALSE(Pat1->match("y"));
82+
EXPECT_FALSE(Pat1->match("z"));
83+
EXPECT_FALSE(Pat1->match(""));
84+
85+
Expected<GlobPattern> Pat2 = GlobPattern::create("[!abc-fy-z]");
86+
EXPECT_TRUE((bool)Pat2);
87+
EXPECT_TRUE(Pat2->match("g"));
88+
EXPECT_FALSE(Pat2->match("a"));
89+
EXPECT_FALSE(Pat2->match("b"));
90+
EXPECT_FALSE(Pat2->match("c"));
91+
EXPECT_FALSE(Pat2->match("d"));
92+
EXPECT_FALSE(Pat2->match("e"));
93+
EXPECT_FALSE(Pat2->match("f"));
94+
EXPECT_FALSE(Pat2->match("y"));
95+
EXPECT_FALSE(Pat2->match("z"));
96+
EXPECT_FALSE(Pat2->match(""));
97+
}
98+
99+
TEST_F(GlobPatternTest, BracketFrontOfCharacterClass) {
100+
Expected<GlobPattern> Pat1 = GlobPattern::create("[]a]x");
101+
EXPECT_TRUE((bool)Pat1);
102+
EXPECT_TRUE(Pat1->match("]x"));
103+
EXPECT_TRUE(Pat1->match("ax"));
104+
EXPECT_FALSE(Pat1->match("a]x"));
105+
EXPECT_FALSE(Pat1->match(""));
106+
}
107+
108+
TEST_F(GlobPatternTest, SpecialCharsInCharacterClass) {
109+
Expected<GlobPattern> Pat1 = GlobPattern::create("[*?^]");
110+
EXPECT_TRUE((bool)Pat1);
111+
EXPECT_TRUE(Pat1->match("*"));
112+
EXPECT_TRUE(Pat1->match("?"));
113+
EXPECT_TRUE(Pat1->match("^"));
114+
EXPECT_FALSE(Pat1->match("*?^"));
115+
EXPECT_FALSE(Pat1->match(""));
62116
}
63117

64118
TEST_F(GlobPatternTest, Invalid) {
65119
Expected<GlobPattern> Pat1 = GlobPattern::create("[");
66120
EXPECT_FALSE((bool)Pat1);
67121
handleAllErrors(Pat1.takeError(), [&](ErrorInfoBase &EIB) {});
122+
123+
Expected<GlobPattern> Pat2 = GlobPattern::create("[]");
124+
EXPECT_FALSE((bool)Pat2);
125+
handleAllErrors(Pat2.takeError(), [&](ErrorInfoBase &EIB) {});
68126
}
69127

70128
TEST_F(GlobPatternTest, ExtSym) {

0 commit comments

Comments
 (0)