1894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//===- FileCheck.cpp - Check that File's Contents match what is expected --===//
2894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//
3894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//                     The LLVM Compiler Infrastructure
4894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//
5894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman// This file is distributed under the University of Illinois Open Source
6894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman// License. See LICENSE.TXT for details.
7894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//
8894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//===----------------------------------------------------------------------===//
9894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//
10894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman// FileCheck does a line-by line check of a file that validates whether it
11894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman// contains the expected content.  This is useful for regression tests etc.
12894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//
13894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman// This program exits with an error status of 2 on error, exit status of 0 if
14894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman// the file matched the expected contents, and exit status of 1 if it did not
15894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman// contain the expected contents.
16894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//
17894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//===----------------------------------------------------------------------===//
18894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
1919bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman#include "llvm/ADT/OwningPtr.h"
20894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman#include "llvm/Support/CommandLine.h"
21894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman#include "llvm/Support/MemoryBuffer.h"
22894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman#include "llvm/Support/PrettyStackTrace.h"
23894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman#include "llvm/Support/Regex.h"
24894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman#include "llvm/Support/SourceMgr.h"
25894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman#include "llvm/Support/raw_ostream.h"
2619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman#include "llvm/Support/Signals.h"
2719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman#include "llvm/Support/system_error.h"
28894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman#include "llvm/ADT/SmallString.h"
29894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman#include "llvm/ADT/StringMap.h"
30894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman#include <algorithm>
31894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanusing namespace llvm;
32894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
33894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanstatic cl::opt<std::string>
34894018228b0e0bdbd7aa7e8f47d4a9458789ca82John BaumanCheckFilename(cl::Positional, cl::desc("<check-file>"), cl::Required);
35894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
36894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanstatic cl::opt<std::string>
37894018228b0e0bdbd7aa7e8f47d4a9458789ca82John BaumanInputFilename("input-file", cl::desc("File to check (defaults to stdin)"),
38894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman              cl::init("-"), cl::value_desc("filename"));
39894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
40894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanstatic cl::opt<std::string>
41894018228b0e0bdbd7aa7e8f47d4a9458789ca82John BaumanCheckPrefix("check-prefix", cl::init("CHECK"),
42894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman            cl::desc("Prefix to use from check file (defaults to 'CHECK')"));
43894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
44894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanstatic cl::opt<bool>
45894018228b0e0bdbd7aa7e8f47d4a9458789ca82John BaumanNoCanonicalizeWhiteSpace("strict-whitespace",
46894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman              cl::desc("Do not treat all horizontal whitespace as equivalent"));
47894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
48894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//===----------------------------------------------------------------------===//
49894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman// Pattern Handling Code.
50894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//===----------------------------------------------------------------------===//
51894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
52894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanclass Pattern {
53894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  SMLoc PatternLoc;
5419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
5519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  /// MatchEOF - When set, this pattern only matches the end of file. This is
5619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  /// used for trailing CHECK-NOTs.
5719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  bool MatchEOF;
5819bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
59894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// FixedStr - If non-empty, this pattern is a fixed string match with the
60894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// specified fixed string.
61894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  StringRef FixedStr;
6219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
63894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// RegEx - If non-empty, this is a regex pattern.
64894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  std::string RegExStr;
6519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
66894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// VariableUses - Entries in this vector map to uses of a variable in the
67894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// pattern, e.g. "foo[[bar]]baz".  In this case, the RegExStr will contain
68894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// "foobaz" and we'll get an entry in this vector that tells us to insert the
69894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// value of bar at offset 3.
70894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  std::vector<std::pair<StringRef, unsigned> > VariableUses;
7119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
72894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// VariableDefs - Entries in this vector map to definitions of a variable in
73894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// the pattern, e.g. "foo[[bar:.*]]baz".  In this case, the RegExStr will
74894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// contain "foo(.*)baz" and VariableDefs will contain the pair "bar",1.  The
75894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// index indicates what parenthesized value captures the variable value.
76894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  std::vector<std::pair<StringRef, unsigned> > VariableDefs;
7719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
78894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanpublic:
7919bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
8019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  Pattern(bool matchEOF = false) : MatchEOF(matchEOF) { }
8119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
82894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  bool ParsePattern(StringRef PatternStr, SourceMgr &SM);
8319bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
84894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// Match - Match the pattern string against the input buffer Buffer.  This
85894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// returns the position that is matched or npos if there is no match.  If
86894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// there is a match, the size of the matched string is returned in MatchLen.
87894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  ///
88894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// The VariableTable StringMap provides the current values of filecheck
89894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// variables and is updated if this match defines new values.
90894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  size_t Match(StringRef Buffer, size_t &MatchLen,
91894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman               StringMap<StringRef> &VariableTable) const;
92894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
93894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// PrintFailureInfo - Print additional information about a failure to match
94894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// involving this pattern.
95894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  void PrintFailureInfo(const SourceMgr &SM, StringRef Buffer,
96894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        const StringMap<StringRef> &VariableTable) const;
97894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
98894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanprivate:
99894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  static void AddFixedStringToRegEx(StringRef FixedStr, std::string &TheStr);
100894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  bool AddRegExToRegEx(StringRef RegExStr, unsigned &CurParen, SourceMgr &SM);
101894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
102894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// ComputeMatchDistance - Compute an arbitrary estimate for the quality of
103894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// matching this pattern at the start of \arg Buffer; a distance of zero
104894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// should correspond to a perfect match.
105894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  unsigned ComputeMatchDistance(StringRef Buffer,
106894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                               const StringMap<StringRef> &VariableTable) const;
107894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman};
108894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
109894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
110894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanbool Pattern::ParsePattern(StringRef PatternStr, SourceMgr &SM) {
111894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  PatternLoc = SMLoc::getFromPointer(PatternStr.data());
11219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
113894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Ignore trailing whitespace.
114894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  while (!PatternStr.empty() &&
115894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman         (PatternStr.back() == ' ' || PatternStr.back() == '\t'))
116894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    PatternStr = PatternStr.substr(0, PatternStr.size()-1);
11719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
118894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Check that there is something on the line.
119894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (PatternStr.empty()) {
120894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    SM.PrintMessage(PatternLoc, "found empty check string with prefix '" +
121894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                    CheckPrefix+":'", "error");
122894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    return true;
123894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
12419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
125894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Check to see if this is a fixed string, or if it has regex pieces.
126894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (PatternStr.size() < 2 ||
127894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      (PatternStr.find("{{") == StringRef::npos &&
128894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman       PatternStr.find("[[") == StringRef::npos)) {
129894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    FixedStr = PatternStr;
130894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    return false;
131894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
13219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
133894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Paren value #0 is for the fully matched string.  Any new parenthesized
13419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  // values add from there.
135894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  unsigned CurParen = 1;
13619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
137894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Otherwise, there is at least one regex piece.  Build up the regex pattern
138894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // by escaping scary characters in fixed strings, building up one big regex.
139894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  while (!PatternStr.empty()) {
140894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // RegEx matches.
14119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    if (PatternStr.startswith("{{")) {
14219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
143894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // Otherwise, this is the start of a regex match.  Scan for the }}.
144894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      size_t End = PatternStr.find("}}");
145894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (End == StringRef::npos) {
146894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        SM.PrintMessage(SMLoc::getFromPointer(PatternStr.data()),
14719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman                        "found start of regex string with no end '}}'","error");
148894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        return true;
149894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      }
15019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
15119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman      // Enclose {{}} patterns in parens just like [[]] even though we're not
15219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman      // capturing the result for any purpose.  This is required in case the
15319bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman      // expression contains an alternation like: CHECK:  abc{{x|z}}def.  We
15419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman      // want this to turn into: "abc(x|z)def" not "abcx|zdef".
15519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman      RegExStr += '(';
15619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman      ++CurParen;
15719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
158894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (AddRegExToRegEx(PatternStr.substr(2, End-2), CurParen, SM))
159894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        return true;
16019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman      RegExStr += ')';
16119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
162894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      PatternStr = PatternStr.substr(End+2);
163894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      continue;
164894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
16519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
166894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Named RegEx matches.  These are of two forms: [[foo:.*]] which matches .*
167894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // (or some other regex) and assigns it to the FileCheck variable 'foo'. The
168894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // second form is [[foo]] which is a reference to foo.  The variable name
169894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // itself must be of the form "[a-zA-Z_][0-9a-zA-Z_]*", otherwise we reject
170894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // it.  This is to catch some common errors.
17119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    if (PatternStr.startswith("[[")) {
172894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // Verify that it is terminated properly.
173894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      size_t End = PatternStr.find("]]");
174894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (End == StringRef::npos) {
175894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        SM.PrintMessage(SMLoc::getFromPointer(PatternStr.data()),
176894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        "invalid named regex reference, no ]] found", "error");
177894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        return true;
178894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      }
17919bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
180894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      StringRef MatchStr = PatternStr.substr(2, End-2);
181894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      PatternStr = PatternStr.substr(End+2);
18219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
183894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // Get the regex name (e.g. "foo").
184894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      size_t NameEnd = MatchStr.find(':');
185894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      StringRef Name = MatchStr.substr(0, NameEnd);
18619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
187894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (Name.empty()) {
188894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        SM.PrintMessage(SMLoc::getFromPointer(Name.data()),
189894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        "invalid name in named regex: empty name", "error");
190894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        return true;
191894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      }
192894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
193894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // Verify that the name is well formed.
194894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      for (unsigned i = 0, e = Name.size(); i != e; ++i)
19519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman        if (Name[i] != '_' && !isalnum(Name[i])) {
196894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman          SM.PrintMessage(SMLoc::getFromPointer(Name.data()+i),
197894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                          "invalid name in named regex", "error");
198894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman          return true;
199894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        }
20019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
201894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // Name can't start with a digit.
202894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (isdigit(Name[0])) {
203894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        SM.PrintMessage(SMLoc::getFromPointer(Name.data()),
204894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        "invalid name in named regex", "error");
205894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        return true;
206894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      }
20719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
208894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // Handle [[foo]].
209894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (NameEnd == StringRef::npos) {
210894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        VariableUses.push_back(std::make_pair(Name, RegExStr.size()));
211894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        continue;
212894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      }
21319bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
214894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // Handle [[foo:.*]].
215894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      VariableDefs.push_back(std::make_pair(Name, CurParen));
216894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      RegExStr += '(';
217894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      ++CurParen;
21819bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
219894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (AddRegExToRegEx(MatchStr.substr(NameEnd+1), CurParen, SM))
220894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        return true;
221894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
222894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      RegExStr += ')';
223894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
22419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
225894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Handle fixed string matches.
226894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Find the end, which is the start of the next regex.
227894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    size_t FixedMatchEnd = PatternStr.find("{{");
228894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    FixedMatchEnd = std::min(FixedMatchEnd, PatternStr.find("[["));
229894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    AddFixedStringToRegEx(PatternStr.substr(0, FixedMatchEnd), RegExStr);
230894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    PatternStr = PatternStr.substr(FixedMatchEnd);
231894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    continue;
232894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
233894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
234894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  return false;
235894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman}
236894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
237894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanvoid Pattern::AddFixedStringToRegEx(StringRef FixedStr, std::string &TheStr) {
238894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Add the characters from FixedStr to the regex, escaping as needed.  This
239894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // avoids "leaning toothpicks" in common patterns.
240894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  for (unsigned i = 0, e = FixedStr.size(); i != e; ++i) {
241894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    switch (FixedStr[i]) {
242894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // These are the special characters matched in "p_ere_exp".
243894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case '(':
244894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case ')':
245894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case '^':
246894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case '$':
247894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case '|':
248894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case '*':
249894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case '+':
250894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case '?':
251894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case '.':
252894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case '[':
253894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case '\\':
254894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    case '{':
255894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      TheStr += '\\';
256894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // FALL THROUGH.
257894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    default:
258894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      TheStr += FixedStr[i];
259894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      break;
260894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
261894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
262894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman}
263894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
264894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanbool Pattern::AddRegExToRegEx(StringRef RegexStr, unsigned &CurParen,
265894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                              SourceMgr &SM) {
266894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  Regex R(RegexStr);
267894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  std::string Error;
268894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (!R.isValid(Error)) {
269894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    SM.PrintMessage(SMLoc::getFromPointer(RegexStr.data()),
270894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                    "invalid regex: " + Error, "error");
271894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    return true;
272894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
27319bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
274894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  RegExStr += RegexStr.str();
275894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  CurParen += R.getNumMatches();
276894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  return false;
277894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman}
278894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
279894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman/// Match - Match the pattern string against the input buffer Buffer.  This
280894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman/// returns the position that is matched or npos if there is no match.  If
281894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman/// there is a match, the size of the matched string is returned in MatchLen.
282894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumansize_t Pattern::Match(StringRef Buffer, size_t &MatchLen,
283894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                      StringMap<StringRef> &VariableTable) const {
28419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  // If this is the EOF pattern, match it immediately.
28519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  if (MatchEOF) {
28619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    MatchLen = 0;
28719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    return Buffer.size();
28819bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  }
28919bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
290894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // If this is a fixed string pattern, just match it now.
291894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (!FixedStr.empty()) {
292894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    MatchLen = FixedStr.size();
293894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    return Buffer.find(FixedStr);
294894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
295894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
296894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Regex match.
29719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
298894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // If there are variable uses, we need to create a temporary string with the
299894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // actual value.
300894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  StringRef RegExToMatch = RegExStr;
301894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  std::string TmpStr;
302894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (!VariableUses.empty()) {
303894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    TmpStr = RegExStr;
30419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
305894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    unsigned InsertOffset = 0;
306894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    for (unsigned i = 0, e = VariableUses.size(); i != e; ++i) {
307894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      StringMap<StringRef>::iterator it =
308894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        VariableTable.find(VariableUses[i].first);
309894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // If the variable is undefined, return an error.
310894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (it == VariableTable.end())
311894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        return StringRef::npos;
312894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
313894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // Look up the value and escape it so that we can plop it into the regex.
314894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      std::string Value;
315894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      AddFixedStringToRegEx(it->second, Value);
31619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
317894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // Plop it into the regex at the adjusted offset.
318894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      TmpStr.insert(TmpStr.begin()+VariableUses[i].second+InsertOffset,
319894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                    Value.begin(), Value.end());
320894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      InsertOffset += Value.size();
321894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
32219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
323894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Match the newly constructed regex.
324894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    RegExToMatch = TmpStr;
325894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
32619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
32719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
328894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  SmallVector<StringRef, 4> MatchInfo;
329894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (!Regex(RegExToMatch, Regex::Newline).match(Buffer, &MatchInfo))
330894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    return StringRef::npos;
33119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
332894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Successful regex match.
333894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  assert(!MatchInfo.empty() && "Didn't get any match");
334894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  StringRef FullMatch = MatchInfo[0];
33519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
336894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // If this defines any variables, remember their values.
337894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  for (unsigned i = 0, e = VariableDefs.size(); i != e; ++i) {
338894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    assert(VariableDefs[i].second < MatchInfo.size() &&
339894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman           "Internal paren error");
340894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    VariableTable[VariableDefs[i].first] = MatchInfo[VariableDefs[i].second];
341894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
34219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
343894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  MatchLen = FullMatch.size();
344894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  return FullMatch.data()-Buffer.data();
345894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman}
346894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
347894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanunsigned Pattern::ComputeMatchDistance(StringRef Buffer,
348894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                              const StringMap<StringRef> &VariableTable) const {
349894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Just compute the number of matching characters. For regular expressions, we
350894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // just compare against the regex itself and hope for the best.
351894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  //
352894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // FIXME: One easy improvement here is have the regex lib generate a single
353894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // example regular expression which matches, and use that as the example
354894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // string.
355894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  StringRef ExampleString(FixedStr);
356894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (ExampleString.empty())
357894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    ExampleString = RegExStr;
358894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
359894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Only compare up to the first line in the buffer, or the string size.
360894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  StringRef BufferPrefix = Buffer.substr(0, ExampleString.size());
361894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  BufferPrefix = BufferPrefix.split('\n').first;
362894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  return BufferPrefix.edit_distance(ExampleString);
363894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman}
364894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
365894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanvoid Pattern::PrintFailureInfo(const SourceMgr &SM, StringRef Buffer,
366894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                               const StringMap<StringRef> &VariableTable) const{
367894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // If this was a regular expression using variables, print the current
368894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // variable values.
369894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (!VariableUses.empty()) {
370894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    for (unsigned i = 0, e = VariableUses.size(); i != e; ++i) {
371894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      StringRef Var = VariableUses[i].first;
372894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      StringMap<StringRef>::const_iterator it = VariableTable.find(Var);
373894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      SmallString<256> Msg;
374894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      raw_svector_ostream OS(Msg);
375894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
376894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // Check for undefined variable references.
377894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (it == VariableTable.end()) {
378894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        OS << "uses undefined variable \"";
379894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        OS.write_escaped(Var) << "\"";;
380894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      } else {
381894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        OS << "with variable \"";
382894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        OS.write_escaped(Var) << "\" equal to \"";
383894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        OS.write_escaped(it->second) << "\"";
384894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      }
385894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
386894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), OS.str(), "note",
387894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                      /*ShowLine=*/false);
388894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
389894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
390894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
391894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Attempt to find the closest/best fuzzy match.  Usually an error happens
392894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // because some string in the output didn't exactly match. In these cases, we
393894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // would like to show the user a best guess at what "should have" matched, to
394894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // save them having to actually check the input manually.
395894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  size_t NumLinesForward = 0;
396894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  size_t Best = StringRef::npos;
397894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  double BestQuality = 0;
398894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
399894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Use an arbitrary 4k limit on how far we will search.
400894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  for (size_t i = 0, e = std::min(size_t(4096), Buffer.size()); i != e; ++i) {
401894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (Buffer[i] == '\n')
402894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      ++NumLinesForward;
403894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
404894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Patterns have leading whitespace stripped, so skip whitespace when
405894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // looking for something which looks like a pattern.
406894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (Buffer[i] == ' ' || Buffer[i] == '\t')
407894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      continue;
408894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
409894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Compute the "quality" of this match as an arbitrary combination of the
410894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // match distance and the number of lines skipped to get to this match.
411894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    unsigned Distance = ComputeMatchDistance(Buffer.substr(i), VariableTable);
412894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    double Quality = Distance + (NumLinesForward / 100.);
413894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
414894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (Quality < BestQuality || Best == StringRef::npos) {
415894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      Best = i;
416894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      BestQuality = Quality;
417894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
418894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
419894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
420894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Print the "possible intended match here" line if we found something
421894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // reasonable and not equal to what we showed in the "scanning from here"
422894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // line.
423894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (Best && Best != StringRef::npos && BestQuality < 50) {
424894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      SM.PrintMessage(SMLoc::getFromPointer(Buffer.data() + Best),
425894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                      "possible intended match here", "note");
426894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
427894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // FIXME: If we wanted to be really friendly we would show why the match
428894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // failed, as it can be hard to spot simple one character differences.
429894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
430894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman}
431894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
432894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//===----------------------------------------------------------------------===//
433894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman// Check Strings.
434894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman//===----------------------------------------------------------------------===//
435894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
436894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman/// CheckString - This is a check that we found in the input file.
437894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanstruct CheckString {
438894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// Pat - The pattern to match.
439894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  Pattern Pat;
44019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
441894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// Loc - The location in the match file that the check string was specified.
442894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  SMLoc Loc;
44319bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
444894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// IsCheckNext - This is true if this is a CHECK-NEXT: directive (as opposed
445894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// to a CHECK: directive.
446894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  bool IsCheckNext;
44719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
448894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// NotStrings - These are all of the strings that are disallowed from
449894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// occurring between this match string and the previous one (or start of
450894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// file).
451894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  std::vector<std::pair<SMLoc, Pattern> > NotStrings;
45219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
453894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  CheckString(const Pattern &P, SMLoc L, bool isCheckNext)
454894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    : Pat(P), Loc(L), IsCheckNext(isCheckNext) {}
455894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman};
456894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
457894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman/// CanonicalizeInputFile - Remove duplicate horizontal space from the specified
458894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman/// memory buffer, free it, and return a new one.
459894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanstatic MemoryBuffer *CanonicalizeInputFile(MemoryBuffer *MB) {
460894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  SmallString<128> NewFile;
461894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  NewFile.reserve(MB->getBufferSize());
46219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
463894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  for (const char *Ptr = MB->getBufferStart(), *End = MB->getBufferEnd();
464894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman       Ptr != End; ++Ptr) {
46519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    // Eliminate trailing dosish \r.
46619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    if (Ptr <= End - 2 && Ptr[0] == '\r' && Ptr[1] == '\n') {
46719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman      continue;
46819bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    }
46919bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
470894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // If C is not a horizontal whitespace, skip it.
471894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (*Ptr != ' ' && *Ptr != '\t') {
472894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      NewFile.push_back(*Ptr);
473894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      continue;
474894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
47519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
476894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Otherwise, add one space and advance over neighboring space.
477894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    NewFile.push_back(' ');
478894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    while (Ptr+1 != End &&
479894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman           (Ptr[1] == ' ' || Ptr[1] == '\t'))
480894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      ++Ptr;
481894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
48219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
483894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Free the old buffer and return a new one.
484894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  MemoryBuffer *MB2 =
485894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    MemoryBuffer::getMemBufferCopy(NewFile.str(), MB->getBufferIdentifier());
48619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
487894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  delete MB;
488894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  return MB2;
489894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman}
490894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
491894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
492894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman/// ReadCheckFile - Read the check file, which specifies the sequence of
493894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman/// expected strings.  The strings are added to the CheckStrings vector.
494894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanstatic bool ReadCheckFile(SourceMgr &SM,
495894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                          std::vector<CheckString> &CheckStrings) {
496894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Open the check file, and tell SourceMgr about it.
49719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  OwningPtr<MemoryBuffer> File;
49819bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  if (error_code ec =
49919bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman        MemoryBuffer::getFileOrSTDIN(CheckFilename.c_str(), File)) {
50019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    errs() << "Could not open check file '" << CheckFilename << "': "
50119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman           << ec.message() << '\n';
502894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    return true;
503894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
50419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  MemoryBuffer *F = File.take();
50519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
506894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // If we want to canonicalize whitespace, strip excess whitespace from the
507894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // buffer containing the CHECK lines.
508894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (!NoCanonicalizeWhiteSpace)
509894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    F = CanonicalizeInputFile(F);
51019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
511894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  SM.AddNewSourceBuffer(F, SMLoc());
512894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
513894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Find all instances of CheckPrefix followed by : in the file.
514894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  StringRef Buffer = F->getBuffer();
515894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
516894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  std::vector<std::pair<SMLoc, Pattern> > NotMatches;
51719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
518894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  while (1) {
519894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // See if Prefix occurs in the memory buffer.
520894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    Buffer = Buffer.substr(Buffer.find(CheckPrefix));
52119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
522894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // If we didn't find a match, we're done.
523894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (Buffer.empty())
524894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      break;
52519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
526894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    const char *CheckPrefixStart = Buffer.data();
52719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
528894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // When we find a check prefix, keep track of whether we find CHECK: or
529894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // CHECK-NEXT:
530894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    bool IsCheckNext = false, IsCheckNot = false;
53119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
532894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Verify that the : is present after the prefix.
533894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (Buffer[CheckPrefix.size()] == ':') {
534894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      Buffer = Buffer.substr(CheckPrefix.size()+1);
535894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    } else if (Buffer.size() > CheckPrefix.size()+6 &&
536894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman               memcmp(Buffer.data()+CheckPrefix.size(), "-NEXT:", 6) == 0) {
537894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      Buffer = Buffer.substr(CheckPrefix.size()+7);
538894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      IsCheckNext = true;
539894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    } else if (Buffer.size() > CheckPrefix.size()+5 &&
540894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman               memcmp(Buffer.data()+CheckPrefix.size(), "-NOT:", 5) == 0) {
541894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      Buffer = Buffer.substr(CheckPrefix.size()+6);
542894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      IsCheckNot = true;
543894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    } else {
544894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      Buffer = Buffer.substr(1);
545894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      continue;
546894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
54719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
548894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Okay, we found the prefix, yay.  Remember the rest of the line, but
549894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // ignore leading and trailing whitespace.
550894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    Buffer = Buffer.substr(Buffer.find_first_not_of(" \t"));
55119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
552894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Scan ahead to the end of line.
553894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    size_t EOL = Buffer.find_first_of("\n\r");
554894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
555894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Remember the location of the start of the pattern, for diagnostics.
556894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    SMLoc PatternLoc = SMLoc::getFromPointer(Buffer.data());
557894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
558894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Parse the pattern.
559894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    Pattern P;
560894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (P.ParsePattern(Buffer.substr(0, EOL), SM))
561894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      return true;
56219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
563894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    Buffer = Buffer.substr(EOL);
564894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
56519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
566894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Verify that CHECK-NEXT lines have at least one CHECK line before them.
567894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (IsCheckNext && CheckStrings.empty()) {
568894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      SM.PrintMessage(SMLoc::getFromPointer(CheckPrefixStart),
569894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                      "found '"+CheckPrefix+"-NEXT:' without previous '"+
570894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                      CheckPrefix+ ": line", "error");
571894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      return true;
572894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
57319bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
574894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Handle CHECK-NOT.
575894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (IsCheckNot) {
576894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      NotMatches.push_back(std::make_pair(SMLoc::getFromPointer(Buffer.data()),
577894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                                          P));
578894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      continue;
579894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
58019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
58119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
582894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Okay, add the string we captured to the output vector and move on.
583894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    CheckStrings.push_back(CheckString(P,
584894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                                       PatternLoc,
585894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                                       IsCheckNext));
586894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    std::swap(NotMatches, CheckStrings.back().NotStrings);
587894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
58819bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
58919bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  // Add an EOF pattern for any trailing CHECK-NOTs.
59019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  if (!NotMatches.empty()) {
59119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    CheckStrings.push_back(CheckString(Pattern(true),
59219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman                                       SMLoc::getFromPointer(Buffer.data()),
59319bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman                                       false));
59419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    std::swap(NotMatches, CheckStrings.back().NotStrings);
59519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  }
59619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
597894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (CheckStrings.empty()) {
598894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    errs() << "error: no check strings found with prefix '" << CheckPrefix
599894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman           << ":'\n";
600894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    return true;
601894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
60219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
603894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  return false;
604894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman}
605894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
606894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanstatic void PrintCheckFailed(const SourceMgr &SM, const CheckString &CheckStr,
607894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                             StringRef Buffer,
608894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                             StringMap<StringRef> &VariableTable) {
609894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Otherwise, we have an error, emit an error message.
610894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  SM.PrintMessage(CheckStr.Loc, "expected string not found in input",
611894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                  "error");
61219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
613894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Print the "scanning from here" line.  If the current position is at the
614894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // end of a line, advance to the start of the next line.
615894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  Buffer = Buffer.substr(Buffer.find_first_not_of(" \t\n\r"));
61619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
617894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), "scanning from here",
618894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                  "note");
619894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
620894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Allow the pattern to print additional information if desired.
621894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  CheckStr.Pat.PrintFailureInfo(SM, Buffer, VariableTable);
622894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman}
623894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
624894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman/// CountNumNewlinesBetween - Count the number of newlines in the specified
625894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman/// range.
626894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanstatic unsigned CountNumNewlinesBetween(StringRef Range) {
627894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  unsigned NumNewLines = 0;
628894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  while (1) {
629894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Scan for newline.
630894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    Range = Range.substr(Range.find_first_of("\n\r"));
631894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (Range.empty()) return NumNewLines;
63219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
633894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    ++NumNewLines;
63419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
635894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Handle \n\r and \r\n as a single newline.
636894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (Range.size() > 1 &&
637894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        (Range[1] == '\n' || Range[1] == '\r') &&
638894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        (Range[0] != Range[1]))
639894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      Range = Range.substr(1);
640894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    Range = Range.substr(1);
641894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
642894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman}
643894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
644894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Baumanint main(int argc, char **argv) {
645894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  sys::PrintStackTraceOnErrorSignal();
646894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  PrettyStackTraceProgram X(argc, argv);
647894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  cl::ParseCommandLineOptions(argc, argv);
648894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
649894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  SourceMgr SM;
65019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
651894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Read the expected strings from the check file.
652894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  std::vector<CheckString> CheckStrings;
653894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (ReadCheckFile(SM, CheckStrings))
654894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    return 2;
655894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
656894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Open the file to check and add it to SourceMgr.
65719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  OwningPtr<MemoryBuffer> File;
65819bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  if (error_code ec =
65919bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman        MemoryBuffer::getFileOrSTDIN(InputFilename.c_str(), File)) {
66019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    errs() << "Could not open input file '" << InputFilename << "': "
66119bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman           << ec.message() << '\n';
662894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    return true;
663894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
66419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  MemoryBuffer *F = File.take();
66519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
66619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  if (F->getBufferSize() == 0) {
66719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    errs() << "FileCheck error: '" << InputFilename << "' is empty.\n";
66819bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    return 1;
66919bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman  }
670894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
671894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Remove duplicate spaces in the input file if requested.
672894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  if (!NoCanonicalizeWhiteSpace)
673894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    F = CanonicalizeInputFile(F);
67419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
675894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  SM.AddNewSourceBuffer(F, SMLoc());
67619bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
677894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  /// VariableTable - This holds all the current filecheck variables.
678894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  StringMap<StringRef> VariableTable;
67919bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
680894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // Check that we have all of the expected strings, in order, in the input
681894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  // file.
682894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  StringRef Buffer = F->getBuffer();
68319bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
684894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  const char *LastMatch = Buffer.data();
68519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
686894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  for (unsigned StrNo = 0, e = CheckStrings.size(); StrNo != e; ++StrNo) {
687894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    const CheckString &CheckStr = CheckStrings[StrNo];
68819bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
689894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    StringRef SearchFrom = Buffer;
69019bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
691894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Find StrNo in the file.
692894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    size_t MatchLen = 0;
69319bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    size_t MatchPos = CheckStr.Pat.Match(Buffer, MatchLen, VariableTable);
69419bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    Buffer = Buffer.substr(MatchPos);
69519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
696894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // If we didn't find a match, reject the input.
69719bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman    if (MatchPos == StringRef::npos) {
698894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      PrintCheckFailed(SM, CheckStr, SearchFrom, VariableTable);
699894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      return 1;
700894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
701894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
702894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    StringRef SkippedRegion(LastMatch, Buffer.data()-LastMatch);
703894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
704894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // If this check is a "CHECK-NEXT", verify that the previous match was on
705894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // the previous line (i.e. that there is one newline between them).
706894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    if (CheckStr.IsCheckNext) {
707894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      // Count the number of newlines between the previous match and this one.
708894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      assert(LastMatch != F->getBufferStart() &&
709894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman             "CHECK-NEXT can't be the first check in a file");
710894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
711894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      unsigned NumNewLines = CountNumNewlinesBetween(SkippedRegion);
712894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (NumNewLines == 0) {
713894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        SM.PrintMessage(CheckStr.Loc,
714894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                    CheckPrefix+"-NEXT: is on the same line as previous match",
715894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        "error");
716894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()),
717894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        "'next' match was here", "note");
718894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        SM.PrintMessage(SMLoc::getFromPointer(LastMatch),
719894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        "previous match was here", "note");
720894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        return 1;
721894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      }
72219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
723894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (NumNewLines != 1) {
724894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        SM.PrintMessage(CheckStr.Loc,
725894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        CheckPrefix+
726894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        "-NEXT: is not on the line after the previous match",
727894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        "error");
728894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()),
729894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        "'next' match was here", "note");
730894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        SM.PrintMessage(SMLoc::getFromPointer(LastMatch),
731894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                        "previous match was here", "note");
732894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman        return 1;
733894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      }
734894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
73519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
736894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // If this match had "not strings", verify that they don't exist in the
737894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // skipped region.
738894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    for (unsigned ChunkNo = 0, e = CheckStr.NotStrings.size();
739894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman         ChunkNo != e; ++ChunkNo) {
740894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      size_t MatchLen = 0;
741894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      size_t Pos = CheckStr.NotStrings[ChunkNo].second.Match(SkippedRegion,
742894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                                                             MatchLen,
743894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                                                             VariableTable);
744894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      if (Pos == StringRef::npos) continue;
74519bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
746894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      SM.PrintMessage(SMLoc::getFromPointer(LastMatch+Pos),
747894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                      CheckPrefix+"-NOT: string occurred!", "error");
748894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      SM.PrintMessage(CheckStr.NotStrings[ChunkNo].first,
749894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman                      CheckPrefix+"-NOT: pattern specified here", "note");
750894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman      return 1;
751894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    }
75219bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
753894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman
754894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // Otherwise, everything is good.  Step over the matched text and remember
755894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    // the position after the match as the end of the last match.
756894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    Buffer = Buffer.substr(MatchLen);
757894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman    LastMatch = Buffer.data();
758894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  }
75919bac1e08be200c31efd26f0f5fd144c9b3eefd3John Bauman
760894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman  return 0;
761894018228b0e0bdbd7aa7e8f47d4a9458789ca82John Bauman}
762