[C++] Compilation errors with re2 2023.06.02 on windows #38263
Assignees
Labels
Milestone
Comments
|
@felipecrv @bkietz @zeroshade this seems related to string view, so not sure who's best to take a look. |
raulcd
changed the title
This is related to the C++ |
|
It looks like previously re2::StringPiece::begin() returned a pointer, whereas now it may return an iterator struct instead. The fix is for code in arrow which uses StringPiece::begin() to replace that with StringPiece::data(), which will always return a pointer. |
I checked the |
raulcd
pushed a commit
that referenced
this issue
Oct 16, 2023
… where a char pointer is expected (#38265) ### Rationale for this change The MSVC compiler doesn't seem to allow user code to assume `std::string_view::const_iterator` is `const char*`, so using only `re2::StringPiece` and preferring to call `.data()` instead of `.begin()` should make things more uniform across different compilers and STL implementations. ### What changes are included in this PR? - Using `re2::StringPiece` instead of `std::string_view` to interact with `re2` - Use `data()` instead of `begin()` where a `char*` is expected ### Are these changes tested? Yes, by existing tests. * Closes: #38263 Authored-by: Felipe Oliveira Carvalho <[email protected]> Signed-off-by: Raúl Cumplido <[email protected]>
raulcd
pushed a commit
that referenced
this issue
Oct 16, 2023
… where a char pointer is expected (#38265) ### Rationale for this change The MSVC compiler doesn't seem to allow user code to assume `std::string_view::const_iterator` is `const char*`, so using only `re2::StringPiece` and preferring to call `.data()` instead of `.begin()` should make things more uniform across different compilers and STL implementations. ### What changes are included in this PR? - Using `re2::StringPiece` instead of `std::string_view` to interact with `re2` - Use `data()` instead of `begin()` where a `char*` is expected ### Are these changes tested? Yes, by existing tests. * Closes: #38263 Authored-by: Felipe Oliveira Carvalho <[email protected]> Signed-off-by: Raúl Cumplido <[email protected]>
JerAguilon
pushed a commit
to JerAguilon/arrow
that referenced
this issue
Oct 23, 2023
…egin() where a char pointer is expected (apache#38265) ### Rationale for this change The MSVC compiler doesn't seem to allow user code to assume `std::string_view::const_iterator` is `const char*`, so using only `re2::StringPiece` and preferring to call `.data()` instead of `.begin()` should make things more uniform across different compilers and STL implementations. ### What changes are included in this PR? - Using `re2::StringPiece` instead of `std::string_view` to interact with `re2` - Use `data()` instead of `begin()` where a `char*` is expected ### Are these changes tested? Yes, by existing tests. * Closes: apache#38263 Authored-by: Felipe Oliveira Carvalho <[email protected]> Signed-off-by: Raúl Cumplido <[email protected]>
loicalleyne
pushed a commit
to loicalleyne/arrow
that referenced
this issue
Nov 13, 2023
…egin() where a char pointer is expected (apache#38265) ### Rationale for this change The MSVC compiler doesn't seem to allow user code to assume `std::string_view::const_iterator` is `const char*`, so using only `re2::StringPiece` and preferring to call `.data()` instead of `.begin()` should make things more uniform across different compilers and STL implementations. ### What changes are included in this PR? - Using `re2::StringPiece` instead of `std::string_view` to interact with `re2` - Use `data()` instead of `begin()` where a `char*` is expected ### Are these changes tested? Yes, by existing tests. * Closes: apache#38263 Authored-by: Felipe Oliveira Carvalho <[email protected]> Signed-off-by: Raúl Cumplido <[email protected]>
dgreiss
pushed a commit
to dgreiss/arrow
that referenced
this issue
Feb 19, 2024
…egin() where a char pointer is expected (apache#38265) ### Rationale for this change The MSVC compiler doesn't seem to allow user code to assume `std::string_view::const_iterator` is `const char*`, so using only `re2::StringPiece` and preferring to call `.data()` instead of `.begin()` should make things more uniform across different compilers and STL implementations. ### What changes are included in this PR? - Using `re2::StringPiece` instead of `std::string_view` to interact with `re2` - Use `data()` instead of `begin()` where a `char*` is expected ### Are these changes tested? Yes, by existing tests. * Closes: apache#38263 Authored-by: Felipe Oliveira Carvalho <[email protected]> Signed-off-by: Raúl Cumplido <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Describe the bug, including details regarding any error messages, version, and platform.
Due to some build changes, it took a while for conda-forge to ship re2 2023.06.02, but now that we're trying to compile arrow against that, we get some failures, though only on windows:
Normally there shouldn't be big API changes in re2 (though they did change the SOVERSION per that release). This happens both with/without the unity build.
Component(s)
C++, Packaging
The text was updated successfully, but these errors were encountered: