Commit 59d9612 made a change to the symbols list on Darwin in order to fix an issue preventing re2 from compiling on Mac OS X 10.9, which defaults to the libc++ C++ stdlib. Since the new symbol that was replaced only occurs in libc++, this has the effect of breaking re2 on older versions of OS X by default and on any non-clang compiler on 10.9.

This issue is reposted from https://code.google.com/p/re2/issues/detail?id=99.

The link in the markdown description to the brief docs still refers to code.google.com.
Which no longer exists. Where are the docs?

I'm wondering if it would be a problem to allow an untrusted string to be used as a regex? What is meant/covered by "safe"?

See:
http://www.perlmonks.org/?node_id=1126914
http://stackoverflow.com/questions/20357755/how-can-i-safely-validate-an-untrusted-regex-in-perl/20357964

I just followed the documentation link from http://golang.org/pkg/regexp/, which took me to https://github.com/google/re2/, which links to itself.

Could this be changed to something more useful?

The next most obvious link is this:

https://github.com/google/re2/tree/master/doc

Which isn't immediately useful.

This constructor is declared but never defined:
https://code.google.com/p/re2/source/browse/re2/re2.h#799

What steps will reproduce the problem?

1. Define bool ParseMyType(const char* str, int n, void* dest) { ... }
2. Call RE2::FullMatch(<string>, <regex>, RE2::Arg(&my_type, ParseMyType));
3. Get linker error

What is the expected output? What do you see instead?

Constructor should be defined somewhere, something like this:

template <typename T>
RE2::Arg::Arg(T* p, RE2::Arg::Parser parser) : arg_(p), parser_(parser) { }

Have a fix in https://github.com/dascandy/re2/commit/232c224d164fccce1f023ee7711c4d14bbdd4f6f but no idea how to get a pull request out of it. It makes the set work (and match nothing) if it's empty, and also allows you to ask if it's empty.

In util.h, vsnprintf is #defined as vsnprintf_s:

#define vsnprintf vsnprintf_s

....and it's used twice, both in re2::StringAppendV.

The first use:

  // First try with a small fixed size buffer
  char space[1024];

  // It's possible for methods that use a va_list to invalidate
  // the data in it upon use.  The fix is to make a copy
  // of the structure before using it and use that copy instead.
  va_list backup_ap;
  va_copy(backup_ap, ap);
  int result = vsnprintf(space, sizeof(space), format, backup_ap);
  va_end(backup_ap);

...uses a stack-allocated buffer, and thus template parameter inference works it's magic to infer the size of the buffer.

The second use:

  // Repeatedly increase buffer size until it fits
  int length = sizeof(space);
  while (true) {
    if (result < 0) {
      // Older behavior: just try doubling the buffer size
      length *= 2;
    } else {
      // We need exactly "result+1" characters
      length = result+1;
    }
    char* buf = new char[length];

    // Restore the va_list before we use it again
    va_copy(backup_ap, ap);
    result = vsnprintf(buf, length, format, backup_ap);
    va_end(backup_ap);

...uses a dynamically allocated buffer (char* buf = new char[length]), and thus template parameter inference fails. MSVC generates:

error C2660: 'vsnprintf_s' : function does not take 4 arguments

The correct code looks something like:

result = vsnprintf(buf, length, _TRUNCATE, format, backup_ap);

logging.h:59 generates:

Error   C4996: 'write': The POSIX name for this item is deprecated. Instead, use the ISO C++ conformant name: _write. See online help for details.

The problem is that the line:

if(write(2, s.data(), n) < 0) {}  // shut up gcc

...uses write, not _write.

This is very trivial, but would it be slightly more consistent to use "#ifndef _WIN32" instead of
"ifndef COMPILER_MSVC" in re2.cc line 36?

With REBUILD_FILES=1 set:

$ make
...
g++ -shared -Wl,-soname,libre2.so.0,--version-script=libre2.symbols -pthread -o obj/so/libre2.so.0 obj/so/util/hash.o obj/so/util/rune.o obj/so/util/stringprintf.o obj/so/util/strutil.o obj/so/util/valgrind.o obj/so/re2/bitstate.o obj/so/re2/compile.o obj/so/re2/dfa.o obj/so/re2/filtered_re2.o obj/so/re2/mimics_pcre.o obj/so/re2/nfa.o obj/so/re2/onepass.o obj/so/re2/parse.o obj/so/re2/perl_groups.o obj/so/re2/prefilter.o obj/so/re2/prefilter_tree.o obj/so/re2/prog.o obj/so/re2/re2.o obj/so/re2/regexp.o obj/so/re2/set.o obj/so/re2/simplify.o obj/so/re2/stringpiece.o obj/so/re2/tostring.o obj/so/re2/unicode_casefold.o obj/so/re2/unicode_groups.o
ld: fatal: unrecognized option '--version-script=libre2.symbols'
ld: fatal: use the -z help option for usage information
collect2: error: ld returned 1 exit status
make: *** [obj/so/libre2.so] Error 1
rm re2/unicode_groups.cc re2/unicode_casefold.cc

In dfa.cc, function RunWorkqOnByte, the parameter new_byte_loop is never referenced. What's up with that?

when i tried to use gbk encoding characters, it told me:
re2/re2.cc:201: Error parsing '好': invalid UTF-8
re2/re2.cc:795: Invalid RE2: invalid UTF-8

I want to replace url in a string and following is a basic scenario to reproduce bug.

urlToReplace is a string variable and it contains following data (ignore whitespace placed just before the a tag. Otherwise github does not show example):
< a href = "http://example.com/secure/ViewProfile.jspa?name=bug">

url variable is a string and contains following data:
http://example.com/secure/ViewProfile.jspa?name=bug

Failure occurs at this re2 function call.
bool result = re2::RE2::Replace(&urlToReplace, url, "Data to write...");

Result is false. It should be true. As far as i have seen problem is caused by ? character which exists at both urlToReplace variable and url variable.

Originally reported Sep 24, 2014.

What steps will reproduce the problem?

1. Observe the list of flags supported in Perl
2. Observe the list of flags documented for RE2

What is the expected output? What do you see instead?

I expected the Perl flags to be mentioned in RE2's documentation. In particular, I was interested in /x because it is so so useful.

Hi, we've been using re2 at double negative for some time, previously we pulled tarballs down from googlecode into our build system for each version of re2. Now that you've switched to Github though we can't pull a tarball until a release is made.

Would it be possible to have a release made sometime soon? The previous versioning system like 20150428 would be fine.

Thanks, Shane.

I have requirement where I need to get pattern(in string format) from pre-compiled object. This is needed because I create pre-compiled object in main thread. And passes the object further where I use FullMatch function to do matching. But here I need to see the pattern which I am using. Is there a way to get it?

thanks
Nitin

README points to http://code.google.com/p/re2/ for "documentation about how to install and use RE2", but that page just links back to github now.

Util.h, around line 68, redefines sprintf to sprintf_s:

ifdef _WIN32

define snprintf _snprintf_s

define sprintf sprintf_s

However, sprintf_s has different parameters. The call to sprintf in strutil.cc at line 42 causes a compile error:
sprintf(dest + used, "%03o", c);

C2664 'int sprintf_s(char *const ,const size_t,const char *const ,...)': cannot convert argument 2 from 'const char [6]' to 'const size_t' re2 c:\src\re2\util\strutil.cc 41

Some possible ways to correct this are:

1. Do not redefine sprintf. Build MSVC project with macro _CRT_SECURE_NO_DEPRECATE defined.
2. Do not redefine sprintf. Use something like:
   #ifdef _WIN32
   sprintf_s(dest + used, 5, "%03o", c);
   #else
   sprintf(dest + used, "%03o", c);
   #endif

This doesn't seem to cause any problems (that I'm aware of) but I noticed a few warnings when running my code through valgrind on OS X

==20082== Conditional jump or move depends on uninitialised value(s)
==20082==    at 0x30D96AA: pthread_rwlock_init (in /usr/lib/system/libsystem_c.dylib)
==20082==    by 0xE45D6: re2::Mutex::Mutex() 
==20082==    by 0xE0164: re2::Mutex::Mutex() 
==20082==    by 0xD66A1: re2::Prog::Prog() 
==20082==    by 0xD6624: re2::Prog::Prog() 
==20082==    by 0xB3666: re2::Compiler::Compiler() 
==20082==    by 0xB35E4: re2::Compiler::Compiler() 
==20082==    by 0xB6016: re2::Compiler::Compile(re2::Regexp*, bool, long long) 
==20082==    by 0xB6B53: re2::Regexp::CompileToProg(long long) 
==20082==    by 0xD8506: re2::RE2::Init(re2::StringPiece const&, re2::RE2::Options const&) 
==20082==    by 0xD88E1: re2::RE2::RE2(std::string const&) 
==20082==    by 0xD882C: re2::RE2::RE2(std::string const&) 


==20085== Conditional jump or move depends on uninitialised value(s)
==20085==    at 0x30D96AA: pthread_rwlock_init (in /usr/lib/system/libsystem_c.dylib)
==20085==    by 0xE45D6: re2::Mutex::Mutex() 
==20085==    by 0xE0164: re2::Mutex::Mutex() 
==20085==    by 0xD7F39: re2::RE2::Init(re2::StringPiece const&, re2::RE2::Options const&) 
==20085==    by 0xD88E1: re2::RE2::RE2(std::string const&) 
==20085==    by 0xD882C: re2::RE2::RE2(std::string const&) 

bazel build :all
INFO: Reading 'startup' options from /usr/local/google/home/hanwen/.bazelrc: --watchfs
......
INFO: Found 26 targets...
INFO: From Compiling util/benchmark.cc:
util/benchmark.cc: In function 're2::int64 nsec()':
util/benchmark.cc:29:24: error: 'gettimeofday' was not declared in this scope
if(gettimeofday(&tv, 0) < 0)
^
ERROR: /usr/local/google/home/hanwen/vc/re2/BUILD:73:1: C++ compilation of rule '//:bench-main' failed: gcc failed: error executing command /usr/bin/gcc -iquote . -iquote bazel-out/local_linux-fastbuild/genfiles -isystem . -isystem bazel-out/local_linux-fastbuild/genfiles -isystem tools/cpp/gcc3 '-std=c++0x' ... (remaining 9 argument(s) skipped): com.google.devtools.build.lib.shell.BadExitStatusException: Process exited with status 1.
INFO: Elapsed time: 4.742s, Critical Path: 3.51s

In 376ee99 RE2_HAVE_LONGLONG was introduced for disabling methods that use C++11 long long when using a C++03 compiler.

This patch is faulty as at the top of re2/re2.h RE2_HAVE_LONGLONG is set to 0 if it is not already defined. This means that the #ifdef RE2_HAVE_LONGLONGs always evaluate to true.

It's a minor issue, since StringPiece is not the thrust of this library, but I ran into a place where I wanted to use it.

What steps will reproduce the problem?

1. Install re2 prerequisites on a Cygwin64 environment
2. Download re2 and run "make && make install"
3. Make fixes specified at http://stackoverflow.com/a/8393475/4027566 and run "make test".

What is the expected output? What do you see instead?
All tests are expected to pass; however, obj/dbg/test/re2_test fails.

What version of the product are you using? On what operating system?
The latest from mercurial as of 2014-12-03 04:00UTC.

Please provide any additional information below.
The output from the logfile is as follows:
$ cat obj/dbg/test/re2_test.log
re2/testing/re2_test.cc:688: Memory at 0x6fffffe0000
re2/testing/re2_test.cc:826: Check failed: (v) == (6700000000081920.1f)6.6999997e+15 != 6.7000003e+15

I encountered a crash when calling FullMatchN.
From the code below, the first and third block are OK, but the second block (below the comment) will crash. Is there a problem with my code or a bug in re2?

#include <re2/re2.h>
#include <iostream>

using namespace re2;

int main() {
  {
    std::string s;
    int i = 0;
    if (RE2::FullMatch("hello123436", "h(e)l+o([0-9]+)", &s, &i)) {
      std::cout << "1. match\n";
      std::cout << "s = " << s << "\n";
      std::cout << "i = " << i << "\n";
    }
  }

  {
    std::string s;
    int i = 0;
    RE2::Arg* args = new RE2::Arg[2];
    args[0] = &s;
    args[1] = &i;

    /* the line below will crash, why? */
    if (RE2::FullMatchN("hello123436", "h(e)l+o([0-9]+)", &args, 2)) {
      std::cout << "2. match\n";
      std::cout << "s = " << s << "\n";
      std::cout << "i = " << i << "\n";
    }
    delete[] args;
  }

  {

    std::string s;
    int i = 0;

    RE2::Arg arg1 = &s;
    RE2::Arg arg2 = &i;
    const RE2::Arg* args[2] = {&arg1, &arg2};

    if (RE2::FullMatchN("hello123436", "h(e)l+o([0-9]+)", args, 2)) {
      std::cout << "3. match\n";
      std::cout << "s = " << s << "\n";
      std::cout << "i = " << i << "\n";
    }
  }
}

The following regexp ^a(/a+(#a+)*)*$ fails to match a/a#a when it should. Adding parenthesis around the first part ^(a)(/a+(#a+)*)*$ apparently fixes the issue and gives a match, see http://play.golang.org/p/6v0VjVNFVN .

For the background, see http://stackoverflow.com/questions/31663430/issue-with-regexp-with-nested-groups-in-golang

Hi there,

I've discovered a problem when trying to build re2 for M-Dessert, where stlport is not used any more.
In particular, re2 doesn't build because of the following lines in util.h:

#if defined(ANDROID)

#if defined(_STLPORT_VERSION)
#include <unordered_set>      // using stlport
#else
#include <tr1/unordered_set>  // using gnustl
#endif
using std::tr1::unordered_set;

#elif defined(__GNUC__) && !defined(USE_CXX0X)

#include <tr1/unordered_set>
using std::tr1::unordered_set;

#else

#include <unordered_set>
using std::unordered_set;

#endif

And because of this directive in Android.mk
LOCAL_SDK_VERSION := 14.

The fix "seems to be" the following:

#if defined(ANDROID) 

#if defined(_STLPORT_VERSION) || defined(_USING_LIBCXX)
#include <unordered_set>      // using stlport
using std::unordered_set;
#else
#include <tr1/unordered_set>  // using gnustl
using std::tr1::unordered_set;
#endif

#elif defined(__GNUC__) && !defined(USE_CXX0X)

#include <tr1/unordered_set>
using std::tr1::unordered_set;

#else

#include <unordered_set>
using std::unordered_set;

#endif

And there is the need to comment "LOCAL_SDK_VERSION := 14" in the makefile.

Please elaborate.
C

If I try to compile the test suite in MSVC 2013 without pcre, (removing pcre.cc from project test and project benchmark) I get link errors because entry points from pcre.cc are not found:

1>------ Build started: Project: benchmark, Configuration: Release x64 ------
2>------ Build started: Project: dfa_test, Configuration: Release x64 ------
2> dfa_test.cc
1> benchmark.cc
1>E:\src\re2\util\benchmark.cc(29): error C2079: 'tv' uses undefined struct 'nsec::timeval'
1>E:\src\re2\util\benchmark.cc(30): error C3861: 'gettimeofday': identifier not found
1>E:\src\re2\util\benchmark.cc(32): error C2228: left of '.tv_sec' must have class/struct/union
1> type is 'int'
1>E:\src\re2\util\benchmark.cc(32): error C2228: left of '.tv_usec' must have class/struct/union
1> type is 'int'
1>E:\src\re2\util\benchmark.cc(115): warning C4244: '=' : conversion from 'double' to 'int', possible loss of data
1>E:\src\re2\util\benchmark.cc(117): warning C4244: '=' : conversion from 'double' to 'int', possible loss of data
2>E:\src\re2\util/thread.h(25): error C2146: syntax error : missing ';' before identifier 'pid_'
2>E:\src\re2\util/thread.h(25): error C4430: missing type specifier - int assumed. Note: C++ does not support default-int
2>E:\src\re2\re2/prog.h(99): warning C4244: 'return' : conversion from '__int64' to 'int', possible loss of data
2>E:\src\re2\re2\testing\dfa_test.cc(111): warning C4244: '=' : conversion from '__int64' to 'int', possible loss of data
3>------ Build started: Project: exhaustive1_test, Configuration: Release x64 ------
4>------ Build started: Project: exhaustive2_test, Configuration: Release x64 ------
3>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::PCRE(class std::basic_string<char,struct std::char_traits,class std::allocator > const &,class re2::PCRE_Options const &)" (??0PCRE@re2@@qeaa@AEBV?$basic_string@DU?$char_traits@D@std@@v?$allocator@D@2@@std@@AEBVPCRE_Options@1@@z) referenced in function "public: __cdecl re2::TestInstance::TestInstance(class re2::StringPiece const &,enum re2::Prog::MatchKind,enum re2::Regexp::ParseFlags)" (??0TestInstance@re2@@qeaa@AEBVStringPiece@1@W4MatchKind@Prog@1@W4ParseFlags@Regexp@1@@z)
3>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::~PCRE(void)" (??1PCRE@re2@@qeaa@XZ) referenced in function "public: __cdecl re2::TestInstance::~TestInstance(void)" (??1TestInstance@re2@@qeaa@XZ)
3>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::HitLimit(void)" (?HitLimit@PCRE@re2@@QEAA_NXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
3>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: void __cdecl re2::PCRE::ClearHitLimit(void)" (?ClearHitLimit@PCRE@re2@@QEAAXXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
3>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::DoMatch(class re2::StringPiece const &,enum re2::PCRE::Anchor,int *,class re2::PCRE::Arg const * const *,int)const " (?DoMatch@PCRE@re2@@QEBA_NAEBVStringPiece@2@W4Anchor@12@PEAHPEBQEBVArg@12@H@Z) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
3>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_null(char const *,int,void *)" (?parse_null@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(void)" (??0Arg@PCRE@re2@@qeaa@XZ)
3>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_stringpiece(char const *,int,void *)" (?parse_stringpiece@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(class re2::StringPiece *)" (??0Arg@PCRE@re2@@qeaa@PEAVStringPiece@2@@z)
3>E:\src\re2\Release\exhaustive1_test.exe : fatal error LNK1120: 7 unresolved externals
4>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::PCRE(class std::basic_string<char,struct std::char_traits,class std::allocator > const &,class re2::PCRE_Options const &)" (??0PCRE@re2@@qeaa@AEBV?$basic_string@DU?$char_traits@D@std@@v?$allocator@D@2@@std@@AEBVPCRE_Options@1@@z) referenced in function "public: __cdecl re2::TestInstance::TestInstance(class re2::StringPiece const &,enum re2::Prog::MatchKind,enum re2::Regexp::ParseFlags)" (??0TestInstance@re2@@qeaa@AEBVStringPiece@1@W4MatchKind@Prog@1@W4ParseFlags@Regexp@1@@z)
4>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::~PCRE(void)" (??1PCRE@re2@@qeaa@XZ) referenced in function "public: __cdecl re2::TestInstance::~TestInstance(void)" (??1TestInstance@re2@@qeaa@XZ)
4>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::HitLimit(void)" (?HitLimit@PCRE@re2@@QEAA_NXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
4>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: void __cdecl re2::PCRE::ClearHitLimit(void)" (?ClearHitLimit@PCRE@re2@@QEAAXXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
4>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::DoMatch(class re2::StringPiece const &,enum re2::PCRE::Anchor,int *,class re2::PCRE::Arg const * const *,int)const " (?DoMatch@PCRE@re2@@QEBA_NAEBVStringPiece@2@W4Anchor@12@PEAHPEBQEBVArg@12@H@Z) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
4>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_null(char const *,int,void *)" (?parse_null@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(void)" (??0Arg@PCRE@re2@@qeaa@XZ)
4>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_stringpiece(char const *,int,void *)" (?parse_stringpiece@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(class re2::StringPiece *)" (??0Arg@PCRE@re2@@qeaa@PEAVStringPiece@2@@z)
4>E:\src\re2\Release\exhaustive2_test.exe : fatal error LNK1120: 7 unresolved externals
5>------ Build started: Project: exhaustive3_test, Configuration: Release x64 ------
6>------ Build started: Project: exhaustive_test, Configuration: Release x64 ------
5>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::PCRE(class std::basic_string<char,struct std::char_traits,class std::allocator > const &,class re2::PCRE_Options const &)" (??0PCRE@re2@@qeaa@AEBV?$basic_string@DU?$char_traits@D@std@@v?$allocator@D@2@@std@@AEBVPCRE_Options@1@@z) referenced in function "public: __cdecl re2::TestInstance::TestInstance(class re2::StringPiece const &,enum re2::Prog::MatchKind,enum re2::Regexp::ParseFlags)" (??0TestInstance@re2@@qeaa@AEBVStringPiece@1@W4MatchKind@Prog@1@W4ParseFlags@Regexp@1@@z)
5>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::~PCRE(void)" (??1PCRE@re2@@qeaa@XZ) referenced in function "public: __cdecl re2::TestInstance::~TestInstance(void)" (??1TestInstance@re2@@qeaa@XZ)
5>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::HitLimit(void)" (?HitLimit@PCRE@re2@@QEAA_NXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
5>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: void __cdecl re2::PCRE::ClearHitLimit(void)" (?ClearHitLimit@PCRE@re2@@QEAAXXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
5>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::DoMatch(class re2::StringPiece const &,enum re2::PCRE::Anchor,int *,class re2::PCRE::Arg const * const *,int)const " (?DoMatch@PCRE@re2@@QEBA_NAEBVStringPiece@2@W4Anchor@12@PEAHPEBQEBVArg@12@H@Z) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
5>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_null(char const *,int,void *)" (?parse_null@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(void)" (??0Arg@PCRE@re2@@qeaa@XZ)
5>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_stringpiece(char const *,int,void *)" (?parse_stringpiece@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(class re2::StringPiece *)" (??0Arg@PCRE@re2@@qeaa@PEAVStringPiece@2@@z)
5>E:\src\re2\Release\exhaustive3_test.exe : fatal error LNK1120: 7 unresolved externals
6>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::PCRE(class std::basic_string<char,struct std::char_traits,class std::allocator > const &,class re2::PCRE_Options const &)" (??0PCRE@re2@@qeaa@AEBV?$basic_string@DU?$char_traits@D@std@@v?$allocator@D@2@@std@@AEBVPCRE_Options@1@@z) referenced in function "public: __cdecl re2::TestInstance::TestInstance(class re2::StringPiece const &,enum re2::Prog::MatchKind,enum re2::Regexp::ParseFlags)" (??0TestInstance@re2@@qeaa@AEBVStringPiece@1@W4MatchKind@Prog@1@W4ParseFlags@Regexp@1@@z)
6>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::~PCRE(void)" (??1PCRE@re2@@qeaa@XZ) referenced in function "public: __cdecl re2::TestInstance::~TestInstance(void)" (??1TestInstance@re2@@qeaa@XZ)
6>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::HitLimit(void)" (?HitLimit@PCRE@re2@@QEAA_NXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
6>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: void __cdecl re2::PCRE::ClearHitLimit(void)" (?ClearHitLimit@PCRE@re2@@QEAAXXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
6>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::DoMatch(class re2::StringPiece const &,enum re2::PCRE::Anchor,int *,class re2::PCRE::Arg const * const *,int)const " (?DoMatch@PCRE@re2@@QEBA_NAEBVStringPiece@2@W4Anchor@12@PEAHPEBQEBVArg@12@H@Z) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
6>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_null(char const *,int,void *)" (?parse_null@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(void)" (??0Arg@PCRE@re2@@qeaa@XZ)
6>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_stringpiece(char const *,int,void *)" (?parse_stringpiece@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(class re2::StringPiece *)" (??0Arg@PCRE@re2@@qeaa@PEAVStringPiece@2@@z)
6>E:\src\re2\Release\exhaustive_test.exe : fatal error LNK1120: 7 unresolved externals
7>------ Build started: Project: random_test, Configuration: Release x64 ------
8>------ Build started: Project: re2_test, Configuration: Release x64 ------
8> re2_test.cc
8>E:\src\re2\re2\testing\re2_test.cc(9): fatal error C1083: Cannot open include file: 'sys/mman.h': No such file or directory
7>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::PCRE(class std::basic_string<char,struct std::char_traits,class std::allocator > const &,class re2::PCRE_Options const &)" (??0PCRE@re2@@qeaa@AEBV?$basic_string@DU?$char_traits@D@std@@v?$allocator@D@2@@std@@AEBVPCRE_Options@1@@z) referenced in function "public: __cdecl re2::TestInstance::TestInstance(class re2::StringPiece const &,enum re2::Prog::MatchKind,enum re2::Regexp::ParseFlags)" (??0TestInstance@re2@@qeaa@AEBVStringPiece@1@W4MatchKind@Prog@1@W4ParseFlags@Regexp@1@@z)
7>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::~PCRE(void)" (??1PCRE@re2@@qeaa@XZ) referenced in function "public: __cdecl re2::TestInstance::~TestInstance(void)" (??1TestInstance@re2@@qeaa@XZ)
7>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::HitLimit(void)" (?HitLimit@PCRE@re2@@QEAA_NXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
7>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: void __cdecl re2::PCRE::ClearHitLimit(void)" (?ClearHitLimit@PCRE@re2@@QEAAXXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
7>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::DoMatch(class re2::StringPiece const &,enum re2::PCRE::Anchor,int *,class re2::PCRE::Arg const * const *,int)const " (?DoMatch@PCRE@re2@@QEBA_NAEBVStringPiece@2@W4Anchor@12@PEAHPEBQEBVArg@12@H@Z) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
7>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_null(char const *,int,void *)" (?parse_null@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(void)" (??0Arg@PCRE@re2@@qeaa@XZ)
7>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_stringpiece(char const *,int,void *)" (?parse_stringpiece@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(class re2::StringPiece *)" (??0Arg@PCRE@re2@@qeaa@PEAVStringPiece@2@@z)
7>E:\src\re2\Release\random_test.exe : fatal error LNK1120: 7 unresolved externals
9>------ Build started: Project: regexp_benchmark, Configuration: Release x64 ------
10>------ Build started: Project: search_test, Configuration: Release x64 ------
9>LINK : fatal error LNK1181: cannot open input file 'Release\benchmark.lib'
10>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::PCRE(class std::basic_string<char,struct std::char_traits,class std::allocator > const &,class re2::PCRE_Options const &)" (??0PCRE@re2@@qeaa@AEBV?$basic_string@DU?$char_traits@D@std@@v?$allocator@D@2@@std@@AEBVPCRE_Options@1@@z) referenced in function "public: __cdecl re2::TestInstance::TestInstance(class re2::StringPiece const &,enum re2::Prog::MatchKind,enum re2::Regexp::ParseFlags)" (??0TestInstance@re2@@qeaa@AEBVStringPiece@1@W4MatchKind@Prog@1@W4ParseFlags@Regexp@1@@z)
10>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: __cdecl re2::PCRE::~PCRE(void)" (??1PCRE@re2@@qeaa@XZ) referenced in function "public: __cdecl re2::TestInstance::~TestInstance(void)" (??1TestInstance@re2@@qeaa@XZ)
10>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::HitLimit(void)" (?HitLimit@PCRE@re2@@QEAA_NXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
10>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: void __cdecl re2::PCRE::ClearHitLimit(void)" (?ClearHitLimit@PCRE@re2@@QEAAXXZ) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
10>test.lib(tester.obj) : error LNK2019: unresolved external symbol "public: bool __cdecl re2::PCRE::DoMatch(class re2::StringPiece const &,enum re2::PCRE::Anchor,int *,class re2::PCRE::Arg const * const *,int)const " (?DoMatch@PCRE@re2@@QEBA_NAEBVStringPiece@2@W4Anchor@12@PEAHPEBQEBVArg@12@H@Z) referenced in function "private: void __cdecl re2::TestInstance::RunSearch(enum re2::Engine,class re2::StringPiece const &,class re2::StringPiece const &,enum re2::Prog::Anchor,struct re2::TestInstance::Result *)" (?RunSearch@TestInstance@re2@@AEAAXW4Engine@2@AEBVStringPiece@2@1W4Anchor@Prog@2@PEAUResult@12@@z)
10>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_null(char const *,int,void *)" (?parse_null@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(void)" (??0Arg@PCRE@re2@@qeaa@XZ)
10>test.lib(tester.obj) : error LNK2019: unresolved external symbol "private: static bool __cdecl re2::PCRE::Arg::parse_stringpiece(char const *,int,void *)" (?parse_stringpiece@Arg@PCRE@re2@@CA_NPEBDHPEAX@Z) referenced in function "public: __cdecl re2::PCRE::Arg::Arg(class re2::StringPiece *)" (??0Arg@PCRE@re2@@qeaa@PEAVStringPiece@2@@z)
10>E:\src\re2\Release\search_test.exe : fatal error LNK1120: 7 unresolved externals
11>------ Skipped Build: Project: ALL_BUILD, Configuration: Release x64 ------
11>Project not selected to build for this solution configuration
========== Build: 0 succeeded, 10 failed, 15 up-to-date, 1 skipped ==========

Not sure if this is expected, but I'm seeing significantly slower performance (~200x) when compiling unicode regex. For example:

\pL\pL\pL\pN\pN -- Single regex compile Time: 8.09727ms.

versus:

[A-Za-z][A-Za-z][A-Za-z][0-9][0-9] -- Single regex compile Time: 0.035597ms

If I compile threadwin.cc in MSVC2013 I get a compiler error in thread.h:
e:\src\re2\util/thread.h(8): fatal error C1083: Cannot open include file: 'pthread.h': No such file or directory

If I put ifdef's around the include of pthread.h, I get no errors in thread.h but I get errors in compiling threadwin.cc

in thread.h:

ifndef _WIN32

include <pthread.h>

endif

e:\src\re2\util/thread.h(22): error C2146: syntax error : missing ';' before identifier 'pid_'
1>e:\src\re2\util/thread.h(22): error C4430: missing type specifier - int assumed. Note: C++ does not support default-int
1>e:\src\re2\util\threadwin.cc(11): error C2065: 'pid_' : undeclared identifier
1>e:\src\re2\util\threadwin.cc(27): error C2065: 'pid_' : undeclared identifier
1>e:\src\re2\util\threadwin.cc(31): error C2065: 'pid_' : undeclared identifier
1>e:\src\re2\util\threadwin.cc(32): error C2065: 'pid_' : undeclared identifier
1>e:\src\re2\util\threadwin.cc(39): error C2065: 'pid_' : undeclared identifier
1>e:\src\re2\util\threadwin.cc(41): error C2065: 'pid_' : undeclared identifier

If I change this line in thread.h:
pthread_t pid_;
to:

ifdef _WIN32

HANDLE pid_;

else

pthread_t pid_;

endif

then threadwin.cc compiles with no errors.

This is all with source-code cloned from the code-review repository as of about 9:30 PM on Monday, September18, with project files build with CMAKE.

A feature that I would really like to see in re2 is the ability to extract the capturing group ranges. I can access the content of named capturing groups matches with the functions NamedCapturingGroups(), CapturingGroupNames() and NumberOfCapturingGroups() and match this with args returned by FullMatchN() but I would really like to be able to get the range of the matches. I.e. the start and end offsets of the match with the input string. This would be very useful for replacing capturing groups in a string. Currently, I'd have to do multiple re2 calls to achieve this.

One way to do this with out breaking the current API would be to extend RE2:Arg by adding int start and int end variables which indicated the starting and ending offset of the match or -1 for no match.

Hi all,

I've noticed some minor leaks because of RE2 initialisation mechanism :

// static empty things for use as const references.
// To avoid global constructors, initialized on demand.
GLOBAL_MUTEX(empty_mutex);
static const string *empty_string;
static const map<string, int> *empty_named_groups;
static const map<int, string> *empty_group_names;

static void InitEmpty() {
  GLOBAL_MUTEX_LOCK(empty_mutex);
  if (empty_string == NULL) {
    empty_string = new string;
    empty_named_groups = new map<string, int>;
    empty_group_names = new map<int, string>;
  }
  GLOBAL_MUTEX_UNLOCK(empty_mutex);
}

Those three allocated objects are never freed, or maybe i did not find a way to do it ...
Wouldn't it be better to have those three objects in the RE2 object ? So that it can be properly freed during destructor.
I know that it would increase RE2 objects' size a little bit but imo it is worthwhile.
It would also remove the need for a global mutex here.

I can provide a patch if needed :)

In compile.cc, function Compiler::RuneByteSuffix, at line 506, MSVC generates:

warning C4805: '|' : unsafe mix of type 're2::uint64' and type 'bool' in operation

where the offending line is:

uint64 key = ((uint64)next << 17) | (lo<<9) | (hi<<1) | foldcase;

...and lo is a uint8, hi is a uint8, next is an int, and foldcase is a bool. Is this really what you intended?

Hey.

I have not found any other place where to ask questions - so if this is not the right one, i'd be happy if you could point me into the right direction where to ask.

Currently I am working on replacing a RegExp extension within a scripting language ( https://github.com/unitpoint/objectscript ) that was formerly based off pcrelib with re2. So I downloaded, test-compiled and ran all the tests that re2 had - and I had no issues at all, OS X 10.10.

But now I want to write the actual C++ wrapper but got a little bit stuck:

  static bool FullMatchN(const StringPiece& text, const RE2& re,
                         const Arg* const args[], int argc);
  static const VariadicFunction2<
      bool, const StringPiece&, const RE2&, Arg, RE2::FullMatchN> FullMatch;

I see that there is const Arg* const args[]. But I see no conversion operator within the Arg class to the parsed type...or at least, no way to convert the matches.

What I would like to do would be something like this:

{type} matches = RE2::fullMtchN(string, RegExp, &args, &argc);
for(int i=0; i<argc; i++) {
    // Add value to a return array to the script language.
    // Get the type within args[i] and add the correct value type.
}

I am not familiar with Python or any other language for which a re2 wrapper exists, so I have a lot of difficulties understanding how they work.

All that I want to do for now is to place in a content string and a regular expression and return the results in an array.

How would I go about this?

Kind regards,
Ingwie

The automata theory does allow this without any overhead (the overall time complexity can stay linear).

The question is how to do it without much extra space. Any ideas?

Hi , the default linux application usage is like ./configure --prefix=xxxx && make && make install , and now i want to compute the --prefix location via some commands in outer wrapper Makefile, do you know how to configure it (use some configure file),thanks very much!

In function compare:

  int compare(const StringPiece& x) const {
    int r = memcmp(ptr_, x.ptr_, std::min(length_, x.length_));
    if (r == 0) {
      if (length_ < x.length_) r = -1;
      else if (length_ > x.length_) r = +1;
    }
    return r;
  }

...re2 uses std::min, but stringpiece.h never #included std::min.

In mutex.h, re2 #defines HAVE_PTHREAD and HAVE_RWLOCK:

/*
 * A simple mutex wrapper, supporting locks and read-write locks.
 * You should assume the locks are *not* re-entrant.
 */

#ifndef RE2_UTIL_MUTEX_H_
#define RE2_UTIL_MUTEX_H_

#include <stdlib.h>

namespace re2 {

#define HAVE_PTHREAD 1
#define HAVE_RWLOCK 1

...without actually checking if we have Pthreads or not.

I think that the POSIX standard guarantees that _POSIX_THREADS is defined when Pthreads is present.

On Windows (VS2013), Pthreads is not present by default, and thus this breaks compilation.

Maybe re2 should check for _POSIX_THREADS instead of HAVE_PTHREAD? (I don't have access to a unix build system at the moment)

I'm not sure about HAVE_RWLOCK.

Simply commenting out the two #defines allows compilation to proceed a bit further, until it hits the #include <pthread.h> in re2.cc.

If I change the HAVE_PTHREAD to:

#if defined(_POSIX_THREADS)
#define HAVE_PTHREAD 1
#endif

...and then guard the #include <pthread.h>:

#if defined(HAVE_PTHREAD)
#include <pthread.h>
#endif

...that eliminates most (maybe all) Pthread issues with compiling on Windows. There are, of course, other issues, but this deals with the Pthreads issues.

Compiling benchmark.cc in MSVC 2013 gets compile errors.
1>------ Build started: Project: benchmark, Configuration: Release x64 ------
1> benchmark.cc
1>E:\src\re2\util\benchmark.cc(29): error C2079: 'tv' uses undefined struct 'nsec::timeval'
1>E:\src\re2\util\benchmark.cc(30): error C3861: 'gettimeofday': identifier not found
1>E:\src\re2\util\benchmark.cc(32): error C2228: left of '.tv_sec' must have class/struct/union
1> type is 'int'
1>E:\src\re2\util\benchmark.cc(32): error C2228: left of '.tv_usec' must have class/struct/union
1> type is 'int'
1>E:\src\re2\util\benchmark.cc(115): warning C4244: '=' : conversion from 'double' to 'int', possible loss of data
1>E:\src\re2\util\benchmark.cc(117): warning C4244: '=' : conversion from 'double' to 'int', possible loss of data

struct timeval and gettimeofday are not defined in MSVC

The following code compiles cleanly and I think it is correct (untested):

static int64 nsec() {

if defined(_MSC_VER)

LARGE_INTEGER freq, val;
QueryPerformanceFrequency(&freq);
QueryPerformanceCounter(&val);
return (1000*(double)val.QuadPart) / (double)freq.QuadPart;

elif defined(APPLE) || defined(_WIN32)

struct timeval tv;
if(gettimeofday(&tv, 0) < 0)
    return -1;
return (int64)tv.tv_sec*1000*1000*1000 + tv.tv_usec*1000;

else

struct timespec tp;
if(clock_gettime(CLOCK_REALTIME, &tp) < 0)
    return -1;
return (int64)tp.tv_sec*1000*1000*1000 + tp.tv_nsec;

endif

}

What steps will reproduce the problem?

1. run install steps as per wiki
2. in last step, make testinstall, it fails.

What is the expected output? What do you see instead?

Expected: to work :)
Actual result:

$ sudo make testinstall
cp testinstall.cc obj
(cd obj && c++ -I/usr/local/include -L/usr/local/lib testinstall.cc -lre2 -pthread -o testinstall)
Undefined symbols for architecture x86_64:
"re2::FilteredRE2::FirstMatch(re2::StringPiece const&, std::__1::vector<int, std::__1::allocator > const&) const", referenced from:
_main in testinstall-025vi1.o
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make: *** [testinstall] Error 1

What version of the product are you using? On what operating system?
Last from source. OS is mavericks (10.9.2)

Thanks!

Since C++11, narrowing in {} is forbidden. Thus I got the following error:

$ make CXXFLAGS="-Wall -O3 -g -pthread -std=c++11" test
g++ -o obj/dbg/re2/testing/re2_test.o -fPIC  -Wall -O3 -g -pthread -std=c++11 -Wsign-compare -c -I.    re2/testing/re2_test.cc
re2/testing/re2_test.cc: In function ‘void re2::RE2Bug18391750()’:
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘252’ from ‘int’ to ‘char’ inside { }
               0x2a, 0x29, 0x0};
                              ^
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘252’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘194’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘155’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘197’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘197’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘212’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘143’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘143’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘231’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘174’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘243’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘174’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1438:30: error: narrowing conversion of ‘174’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc: In function ‘void re2::RE2Bug18458852()’:
re2/testing/re2_test.cc:1454:66: error: narrowing conversion of ‘245’ from ‘int’ to ‘char’ inside { }
               0xf5, 0x87, 0x87, 0x90, 0x29, 0x5d, 0x29, 0x29, 0x0};
                                                                  ^
re2/testing/re2_test.cc:1454:66: error: narrowing conversion of ‘135’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1454:66: error: narrowing conversion of ‘135’ from ‘int’ to ‘char’ inside { }
re2/testing/re2_test.cc:1454:66: error: narrowing conversion of ‘144’ from ‘int’ to ‘char’ inside { }
Makefile:163: recipe for target 'obj/dbg/re2/testing/re2_test.o' failed
make: *** [obj/dbg/re2/testing/re2_test.o] Error 1

-std=gnu++11 causes the same errors. I encountered such errors when compiling facebook/hhvm, which uses -std=gnu++11 for the whole project. See facebook/hhvm#5209

I am using this lib for getting DFA tables to be downloaded to FPGA. I've modified method BuildAllStates in dfa.cc and go through the states using RunStateOnByteUnlocked method. There is no transition to the first state in regexps like "h.*o" or ".*o" or "h.+o" and so on. In other cases everything is nice.

The top-level README says to go to https://github.com/google/re2/ for docs but that just shows the README again. Where are the docs?

In dfa.cc, at line 872, MSVC produces this warning:

warning C4389: '==' : signed/unsigned mismatch  dfa.cc  872

The offending line is:

if ((ip->empty() & flag) == ip->empty())

...where ip->empty() returns an EmptyOp enum (defaulted to int), and flag is an unsigned int.

test code：
bool isFind;
RE2::Options opt;
opt.set_one_line(false);
RE2 test("^abc.*", opt);
isFind = RE2::PartialMatch("xsdfs\r\nabc0000", test);

$ make clean
$ make distclean
$ make LDFLAGS='-z' test

g++ -o obj/dbg/re2/parse.o -fPIC -Wall -O3 -g -pthread -Wsign-compare -c -I. re2/parse.cc
make: *** No rule to make target obj/dbg/re2/perl_groups.o', needed byobj/dbg/libre2.a'. Stop.

First of all: it is not an issue. I just noticed that RE2 has moved to github, and you collect links to RE2 wrappers.

Some time ago I wrote a regular expression addon for node.js that uses RE2 as its engine. All gory details can be found here: https://github.com/uhop/node-re2 (npm: https://www.npmjs.com/package/re2).

On Ubuntu 14.10 64 or 32-bit, which has gcc version 4.9.1,
compiling testinstall.cc with --static results in a program that
hangs. I don't see this problem on other platforms.

git clone https://github.com/google/re2.git
cd re2
make clean
make
g++ --static testinstall.cc -L obj -lre2 -I . -pthread
./a.out

the a.out program hangs, but if you take away --static, it does not hang.

Compiling re2_test.cc in windows fails because #include <unistd.h>. unistd.h is not needed to compile in Windows with the most recent changes and should not be included if _WIN32 is defined. With this one exception, re2_test.cc now compiles cleanly in Windows.

Hi all,
I have encounter a problem using re2, regarding the usage of \x construct in a replacement string.
The re2 library has different behavior than their counterpart in perl. In perl one can use the \x construct in the replacement string. Is this possible for re2?

It would be really handy if escape sequences in the form of \x6D can be used within a rewrite string.
If something like that is not provided by re2, are you interesting for such an enhancement?

in util/strutil.cc, buffer size for sprintf must be large enough to hold the trailing null character. Otherwise it could cause buffer overflow.

--- a/util/strutil.cc
+++ b/util/strutil.cc
@@ -36,9 +36,13 @@ int CEscapeString(const char* src, int src_len, char* dest,
         // digit then that digit must be escaped too to prevent it being
         // interpreted as part of the character code by C.
         if (c < ' ' || c > '~') {
-          if (dest_len - used < 4) // need space for 4 letter escape
+          if (dest_len - used < 5) // need space for 4 letter escape and null
             return -1;
+#ifndef _WIN32
           sprintf(dest + used, "\\%03o", c);
+#else
+          sprintf_s(dest + used, 5, "\\%03o", c);
+#endif
           used += 4;
         } else {
           dest[used++] = c; break;