[committed] libstdc++: Add valid range assertions to std::basic_regex [PR89927]

[Jonathan Wakely via Gcc-patches]

This adds some debug assertions to basic_regex. They don't actually
diagnose the error in the PR yet, but I have another patch to make them
more effective.

Also change the __glibcxx_assert(false) consistency checks to include a
string literal that tells the user a bit more about why the process
aborted. We could consider adding a __glibcxx_bug or
__glibcxx_internal_error macro for this purpose, but ideally we'll never
hit such bugs anyway so it shouldn't be needed.

libstdc++-v3/ChangeLog:

        PR libstdc++/89927
        * include/bits/regex.h (basic_regex(const _Ch_type*, size_t)):
        Add __glibcxx_requires_string_len assertion.
        (basic_regex::assign(InputIterator, InputIterator)): Add
        __glibcxx_requires_valid_range assertion.
        * include/bits/regex_scanner.tcc (_Scanner::_M_advance())
        (_Scanner::_M_scan_normal()): Use string literal in assertions.

Tested powerpc64le-linux. Committed to trunk.


commit 6b6788f8c2748060d922cc22173ff7f8500917e9
Author: Jonathan Wakely <jwak...@redhat.com>
Date:   Mon Oct 11 12:08:59 2021

    libstdc++: Add valid range assertions to std::basic_regex [PR89927]
    
    This adds some debug assertions to basic_regex. They don't actually
    diagnose the error in the PR yet, but I have another patch to make them
    more effective.
    
    Also change the __glibcxx_assert(false) consistency checks to include a
    string literal that tells the user a bit more about why the process
    aborted. We could consider adding a __glibcxx_bug or
    __glibcxx_internal_error macro for this purpose, but ideally we'll never
    hit such bugs anyway so it shouldn't be needed.
    
    libstdc++-v3/ChangeLog:
    
            PR libstdc++/89927
            * include/bits/regex.h (basic_regex(const _Ch_type*, size_t)):
            Add __glibcxx_requires_string_len assertion.
            (basic_regex::assign(InputIterator, InputIterator)): Add
            __glibcxx_requires_valid_range assertion.
            * include/bits/regex_scanner.tcc (_Scanner::_M_advance())
            (_Scanner::_M_scan_normal()): Use string literal in assertions.

diff --git a/libstdc++-v3/include/bits/regex.h 
b/libstdc++-v3/include/bits/regex.h
index bf02bff7c49..3c44bcd7e33 100644
--- a/libstdc++-v3/include/bits/regex.h
+++ b/libstdc++-v3/include/bits/regex.h
@@ -467,7 +467,10 @@ _GLIBCXX_BEGIN_NAMESPACE_CXX11
        */
       basic_regex(const _Ch_type* __p, std::size_t __len,
                  flag_type __f = ECMAScript)
-      { _M_compile(__p, __p + __len, __f); }
+      {
+       __glibcxx_requires_string_len(__p, __len);
+       _M_compile(__p, __p + __len, __f);
+      }
 
       /**
        * @brief Copy-constructs a basic regular expression.
@@ -684,6 +687,7 @@ _GLIBCXX_BEGIN_NAMESPACE_CXX11
          if constexpr (__detail::__is_contiguous_iter<_InputIterator>::value
                        && is_same_v<_ValT, value_type>)
            {
+             __glibcxx_requires_valid_range(__first, __last);
              const auto __len = __last - __first;
              const _Ch_type* __p = std::__to_address(__first);
              _M_compile(__p, __p + __len, __flags);
diff --git a/libstdc++-v3/include/bits/regex_scanner.tcc 
b/libstdc++-v3/include/bits/regex_scanner.tcc
index d81627dc3e9..2fa2303f27b 100644
--- a/libstdc++-v3/include/bits/regex_scanner.tcc
+++ b/libstdc++-v3/include/bits/regex_scanner.tcc
@@ -83,7 +83,7 @@ namespace __detail
        _M_scan_in_brace();
       else
        {
-         __glibcxx_assert(false);
+         __glibcxx_assert(!"unexpected state while processing regex");
        }
     }
 
@@ -195,7 +195,7 @@ namespace __detail
                _M_token = __it->second;
                return;
              }
-         __glibcxx_assert(false);
+         __glibcxx_assert(!"unexpected special character in regex");
        }
       else
        {