问题
Is there a direct efficient way to convert std::sub_match to std::basic_string_view (without constructing an intermediate std::basic_string
and without intermediate heap allocation)? Or one abstraction level further, is there an alternative to std::regex_token_iterator for iterating regex submatches represented as std::basic_string_view instead of std::sub_match using the std
(C++17)?
The reasons why I rather like to use std::basic_string_view over std::sub_match are:
- std::basic_string_view refers to a constant contiguous sequence of char-like objects with the first element of the sequence at position zero. This enables the usage of charconv's std::from_chars (which surprisingly is not implemented using
ForwardIterator
s). This does not seem to be the case for std::sub_match, since it is represented as a pair ofBidirectionalIterator
s. - std::basic_string_view has a much richer string-like interface facilitating additional context-sensitive tokenization in some exceptional cases for some file formats.
回答1:
There's no general way to detect whether an iterator is contiguous. We can still handle known contiguous iterators - such as those of std::string
:
std::string_view as_sv(std::ssub_match m) {
if(!m.matched) return {};
return { &*m.first, m.second - m.first };
}
Handling the remaining named specializations of sub_match
is left as an exercise for the reader.
来源:https://stackoverflow.com/questions/50676058/iterating-regex-submatches-represented-as-stdbasic-string-view