Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improve Unicode handling when writing to an ostream on Windows #2994

Merged
merged 9 commits into from
Jul 23, 2022

Conversation

dimztimz
Copy link
Contributor

See also #2875. I split the whole patch into 4 commits so reviewing is easier.

dimztimz added 5 commits July 20, 2022 19:57
The part with SetConsoleW is a separate function, without fwrite().
Calls to print(ostream&) in the special Unicode case on Windows fallback
to writing via ostream::write instead of fwrite().
@dimztimz
Copy link
Contributor Author

Can you enable the CI again, I pushed a fix. It was a compiler warning.

Copy link
Contributor

@vitaut vitaut left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the PR.

include/fmt/format-inl.h Outdated Show resolved Hide resolved
Comment on lines +13 to +16
#if defined(_WIN32) && defined(__GLIBCXX__)
# include <ext/stdio_filebuf.h>
# include <ext/stdio_sync_filebuf.h>
#endif
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I suggest checking include presence with FMT_HAS_INCLUDE to make it more robust.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why? In case some future GCC removes these files? The checks for Windows and libstdc++ should be still there.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If the files are remove the check will be broken in either case but the removal is easier to support with FMT_HAS_INCLUDE. In general testing against features is more robust than testing against platforms.

Copy link
Contributor

@vitaut vitaut Jul 23, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So I suggest something like

#if defined(_WIN32) && FMT_HAS_INCLUDE(<ext/stdio_filebuf.h>)
#  include <ext/stdio_filebuf.h>
#  include <ext/stdio_sync_filebuf.h>
#  define FMT_USE_STDIO_FILEBUF
#endif

and later:

#ifdef FMT_USE_STDIO_FILEBUF
...
#endif

This will avoid duplicating checks and make the whole thing more robust (including to removal of headers).

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That will incur additional penalty during compilation when one does not use libstdc++ on Windows, e.g. in MSVC. There is additional filesystem access. The header ext/stdio_filebuf.h is very much platform specific thing so it is better to check for platform. Also think of the scenario when someone under MSVC somehow puts ext/stdio_filebuf.h. The check for the platform must be there, IMO.

See

  1. https://gcc.gnu.org/onlinedocs/libstdc++/manual/using_macros.html
  2. https://gcc.gnu.org/onlinedocs/libstdc++/manual/ext_io.html
  3. https://gcc.gnu.org/onlinedocs/gcc-12.1.0/libstdc++/api/a01178.html

Beginning with 3.1, the extra basic_filebuf constructor and the fd() function were removed from the standard filebuf. Instead, <ext/stdio_filebuf.h> contains a derived class template called __gnu_cxx::stdio_filebuf. This class can be constructed from a C FILE* or a file descriptor, and provides the fd() function.

Its there since GCC v3.1 and will remain there because it is stable extension (not experimental). The older supported GCC compiler of FMT is v4.8.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fair enough.

Comment on lines 102 to 103
inline bool write_ostream_gcc_mingw_unicode(std::ostream& os,
fmt::string_view data) {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please merge this function into write_ostream_msvc_unicode and keep the old name write. This should reduce the scope of conditional compilation a bit and remove redundant copy for wchar_t.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done, but i used the name write_ostream_unicode

include/fmt/ostream.h Outdated Show resolved Hide resolved
include/fmt/ostream.h Outdated Show resolved Hide resolved
include/fmt/ostream.h Outdated Show resolved Hide resolved
@vitaut vitaut merged commit 81f1cc7 into fmtlib:master Jul 23, 2022
@vitaut
Copy link
Contributor

vitaut commented Jul 23, 2022

Merged, thanks.

@dimztimz dimztimz deleted the ostream-unicode branch July 26, 2022 19:33
mtremer referenced this pull request in ipfire/ipfire-2.x Nov 28, 2022
- Update from version 9.0.0 to 9.1.0
- Update of rootfile
- Changelog
    9.1.0 - 2022-08-27
	* ``fmt::formatted_size`` now works at compile time
		  `#3026 <https://github.com/fmtlib/fmt/pull/3026>`_
			  For example (`godbolt <https://godbolt.org/z/1MW5rMdf8>`__):
			   .. code:: c++
			     #include <fmt/compile.h>
			     int main() {
			       using namespace fmt::literals;
			       constexpr size_t n = fmt::formatted_size("{}"_cf, 42);
			       fmt::print("{}\n", n); // prints 2
			     }
	* Fixed handling of invalid UTF-8
		  `#3038 <https://github.com/fmtlib/fmt/pull/3038>`_,
		  `#3044 <https://github.com/fmtlib/fmt/pull/3044>`_,
		  `#3056 <https://github.com/fmtlib/fmt/pull/3056>`_
	* Improved Unicode support in ``ostream`` overloads of ``print``
		  `#2994 <https://github.com/fmtlib/fmt/pull/2994>`_,
		  `#3001 <https://github.com/fmtlib/fmt/pull/3001>`_,
		  `#3025 <https://github.com/fmtlib/fmt/pull/3025>`_
	* Fixed handling of the sign specifier in localized formatting on systems with
	   32-bit ``wchar_t``
		  `#3041 <https://github.com/fmtlib/fmt/issues/3041>`_).
	* Added support for wide streams to ``fmt::streamed``
		  `#2994 <https://github.com/fmtlib/fmt/pull/2994>`_
	* Added the ``n`` specifier that disables the output of delimiters when
	   formatting ranges
		  `#2981 <https://github.com/fmtlib/fmt/pull/2981>`_,
		  `#2983 <https://github.com/fmtlib/fmt/pull/2983>`_
			  For example (`godbolt <https://godbolt.org/z/roKqGdj8c>`__):
			   .. code:: c++
			     #include <fmt/ranges.h>
			     #include <vector>
			     int main() {
			       auto v = std::vector{1, 2, 3};
			       fmt::print("{:n}\n", v); // prints 1, 2, 3
			     }
	* Worked around problematic ``std::string_view`` constructors introduced in C++23
		  `#3030 <https://github.com/fmtlib/fmt/issues/3030>`_,
		  `#3050 <https://github.com/fmtlib/fmt/issues/3050>`_
	* Improve handling (exclusion) of recursive ranges
		  `#2968 <https://github.com/fmtlib/fmt/issues/2968>`_,
		  `#2974 <https://github.com/fmtlib/fmt/pull/2974>`_
	* Improved error reporting in format string compilation
		  `#3055 <https://github.com/fmtlib/fmt/issues/3055>`_
	* Improved the implementation of
		  `Dragonbox <https://github.com/jk-jeon/dragonbox>`_, the algorithm used for
		   the default floating-point formatting
		  `#2984 <https://github.com/fmtlib/fmt/pull/2984>`_
	* Fixed issues with floating-point formatting on exotic platforms.
	* Improved the implementation of chrono formatting
		  `#3010 <https://github.com/fmtlib/fmt/pull/3010>`_
	* Improved documentation
		  `#2966 <https://github.com/fmtlib/fmt/pull/2966>`_,
		  `#3009 <https://github.com/fmtlib/fmt/pull/3009>`_,
		  `#3020 <https://github.com/fmtlib/fmt/issues/3020>`_,
		  `#3037 <https://github.com/fmtlib/fmt/pull/3037>`_
	* Improved build configuration
		  `#2991 <https://github.com/fmtlib/fmt/pull/2991>`_,
		  `#2995 <https://github.com/fmtlib/fmt/pull/2995>`_,
		  `#3004 <https://github.com/fmtlib/fmt/issues/3004>`_,
		  `#3007 <https://github.com/fmtlib/fmt/pull/3007>`_,
		  `#3040 <https://github.com/fmtlib/fmt/pull/3040>`_
	* Fixed various warnings and compilation issues
		  `#2969 <https://github.com/fmtlib/fmt/issues/2969>`_,
		  `#2971 <https://github.com/fmtlib/fmt/pull/2971>`_,
		  `#2975 <https://github.com/fmtlib/fmt/issues/2975>`_,
		  `#2982 <https://github.com/fmtlib/fmt/pull/2982>`_,
		  `#2985 <https://github.com/fmtlib/fmt/pull/2985>`_,
		  `#2988 <https://github.com/fmtlib/fmt/issues/2988>`_,
		  `#3000 <https://github.com/fmtlib/fmt/issues/3000>`_,
		  `#3006 <https://github.com/fmtlib/fmt/issues/3006>`_,
		  `#3014 <https://github.com/fmtlib/fmt/issues/3014>`_,
		  `#3015 <https://github.com/fmtlib/fmt/issues/3015>`_,
		  `#3021 <https://github.com/fmtlib/fmt/pull/3021>`_,
		  `#3023 <https://github.com/fmtlib/fmt/issues/3023>`_,
		  `#3024 <https://github.com/fmtlib/fmt/pull/3024>`_,
		  `#3029 <https://github.com/fmtlib/fmt/pull/3029>`_,
		  `#3043 <https://github.com/fmtlib/fmt/pull/3043>`_,
		  `#3052 <https://github.com/fmtlib/fmt/issues/3052>`_,
		  `#3053 <https://github.com/fmtlib/fmt/pull/3053>`_,
		  `#3054 <https://github.com/fmtlib/fmt/pull/3054>`_

Signed-off-by: Adolf Belka <[email protected]>
Reviewed-by: Michael Tremer <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants