`map` is an ordinal structure with log(n) time searches.
`unordered_map` uses O(1) average-time searches and O(n) in the worst
case where a bucket has a to a colliding hash and has to start chaining.
The unordered version should speed up our general-case when looking up
constants.
I've added a trivial order-dependent(_(0,1) and (1,0) will return a
different hash_) hash to combine a 128-bit constant into a
64-bit hash that generally will not collide, using a bit-rotate to
preserve entropy.
In MSVC, having files with identical filenames will result into massive slowdowns when compiling.
The approach I have taken to resolve this is renaming the identically named files in frontend/(A32, A64) to (a32, a64)_filename.cpp/h
This makes dynarmic installable, and also adds a CMake package config
file, that allows projects to use `find_package(dynarmic)` to import the
library.
I know #636 adds the same thing, but while experimenting with the
different install options in
https://github.com/merryhime/dynarmic/pull/636#discussion_r725656034
I ended up with a working patch, so I'm proposing this as well. This
implements solution 2.
This adds versioning information to the built library.
When building the shared library on Linux systems, a new object will
be created: libdynarmic.so.5
This is really useful when talking about ABI compatibility.
The variables dynarmic_VERSION and dynarmic_VERSION_MAJOR
are implicitly created when calling project(dynarmic VERSION x.y.z)
Adds all elements of vector and puts the result into the lowest element.
Accelerates the `addv` instruction into a vectorized implementation
rather than a serial one.
The lane-splatting variant of `FMUL` and `FMLA` is very
common in instruction streams when implementing things like
matrix multiplication. When used, they are used very densely.
https://community.arm.com/developer/ip-products/processors/b/processors-ip-blog/posts/coding-for-neon---part-3-matrix-multiplication
The way this is currently implemented is by grabbing the particular lane
into a general purpose register and then broadcasting it into a simd
register through `VectorGetElement` and `VectorBroadcast`.
```cpp
const IR::U128 operand2 = v.ir.VectorBroadcast(esize, v.ir.VectorGetElement(esize, v.V(idxdsize, Vm), index));
```
What could be done instead is to keep it within
the vector-register and use a permute/shuffle to "splat" the particular
lane across all other lanes, removing the GPR-round-trip.
This is implemented as the new IR instruction `VectorBroadcastElement`:
```cpp
const IR::U128 operand2 = v.ir.VectorBroadcastElement(esize, v.V(idxdsize, Vm), index);
```
Recursive calls to `Replicate` beyond the first call might
cause an unintentional up-casting to an `int` type due
to `|` and `<<` operations on types such as `uint8_t` and `uint16_t`
This makes sure calls such as `Recursive<u8>` stay as the `u8` type
through-out.
Math operations such as Matrix multiplication utilize these particular
instructions enough that there should be some unit tests for thesein particular.
The lane-splatting form of FMUL and FMLA instructions are of particular
interest and I've found them to be very common in retail game binaries
such as Pokemon Sword.
https://community.arm.com/developer/ip-products/processors/b/processors-ip-blog/posts/coding-for-neon---part-3-matrix-multiplication
I'm primarily adding this unit test so that I can ensure compatibility
while I tune and optimize them.
xbyak is intended to be installed in /usr/local/include/xbyak.
Since we desire not to install xbyak before using it, we copy the headers
to the appropriate directory structure and use that instead
AVX512 introduces the _unsigned_ variant of float-to-integer conversion
functions via `vcvttp{sd}2u{dq}q`. In the case that a value is not
representable as an unsigned integer, it will result in `0xFFFFF...`
which can be utilized to get "free" saturation when the floating point
value exceeds the unsigned range, after masking away negative values.
https://www.felixcloutier.com/x86/vcvttps2udqhttps://www.felixcloutier.com/x86/vcvttpd2uqq
This PR also speeds up the _signed_ conversion function for fp64->int64
https://www.felixcloutier.com/x86/vcvttpd2qq
And(a, Not(b)) is a common enough operation that this can
be fused into a single `AndNot` operation. On x64 this is also
a single `pandn` instruction rather than two.
This implementation exists within the unsafe optimization paths and
utilize the 14-bit-precision `vrsqrt14*` and `vrcp14p*`
instructions provided by AVX512F+VL. These are _more_ accurate than
the fallback path and the current `rsqrt`-based unsafe code-path
but still falls in line with what is expected of the
`Unsafe_ReducedErrorFP` optimization flag.
Having AVX512 available will mean this function has 14 bits of precision.
Not having AVX512 available will mean these functions have 11 bits of precision.
9e8b86fd Update version
92fec0f0 Bump version
4749cc93 Update changelog
78a0ba0a Improve conversion of paragraphs
7a39837d Use a working breathe version
55b6e92d Fix docs
69dc3a85 Fix docs
27f4cdd5 Update changelog
70d61a0a Update changelog
427b5340 Add no_value state to value
e421d527 Simplify error handling in parse_nonnegative_int
a59678f3 Fix chrono_test.locale
c98254c3 Install locales into CI
c123a728 Fix set locale error in chrono formatter
3c8fad12 Optimize parse_nonnegative_int
f28cf330 adding a default format for std::chrono::time_point<std::chrono::syst… (#2345)
55010a9d Support non-`char` overloads (module)
0193e7c4 Support compile-time strings and compile-time format string compilation in module
3423d754 Remove the msvc workaround (#2351)
f6b5cc9f Fix chrono_test.weekday on legacy glibc
59a298f1 Enable `enforce-checks-test` for MSVC, too
36c29482 Update docs
c9fe1fa5 Remove unused flag
dccddc2b Apply clang-format
0e36681b Cleanup digit count
1de80f5b Workaround lack of static constexpr in constexpr functions
2039dce7 Detect consteval
d551b88a Move is_char specializations to xchar.h
16c3514d wchar-test -> xchar-test
206000a0 Workaround pathological conversion (#2343)
76ee4904 Move wchar/custom char overloads to xchar.h
e77b22d6 Deprecate memory buffer overload of format_to
07039f4b Update README.rst
4678192c Remove bsr2log10
7c3d3dfa Update thousands_sep_impl signature
ef826b86 Fix docs
5223f552 Remove FMT_ALWAYS_INLINE
cfde93af Add FMT_STATIC_CONSTEXPR
986a5a6c Fixed join_view formatter for wchar_t
7c8b35ff fix MSVC Win32 count_digits
3eeb084e Optimize count_digits
2ac0bfe5 Improve handling of thousands separator
024741b4 CI: set up multi-thread build for all platforms
f4c95f6d Improve handling of thousands separator
d4fbeacc Fix docs build
0eef389d Code style
e27b1ce5 Fix docs
9f8b6dac Fix wheel installation
6060bcfc Fix docs
ff967346 Fix docs
1085cc21 Fix docs
11addaa1 Update docs
760ca5cc Update docs
290d3f8b Cleanup ranges API
aa09e0f5 Update docs
d142579e Cleanup the format API
f286139d Fix "undefined reference to `fmt::v7::detail::basic_data<void>::digits'"
7b9d69b8 Add xchar.h to docs
cbd861f1 Update docs
faf972f0 Update docs
622d1c04 Update changelog
634c9487 Update changelog
a04e3a2d Comment
87876d54 Cleanup the printf implementation
d338d663 Cleanup the printf implementation
272660e7 Remove deprecated printf functions
5a95c5ae Update changelog
70e67ae0 Re-enable module testing Prepare for compilation with gcc (modules branch).
ad972589 Merge branch 'master' of github.com:fmtlib/fmt
ed2a6377 Workaround msvc constexpr issues
99768695 fix custom types formatting at compile-time, add test
8c1b22ba Workaround a gcc 9.1 bug (#2334)
2dba1cfa Update changelog
d7ba6c3e Use qualified name-lookup in module. (#2324)
bf9904ee Workaround msvc bugs
577bce90 Apply clang-format
ba4c7f19 Swap parameter order to match #2327 (#2329)
e9e89b35 Update ChangeLog.rst
9bb406d7 Update changelog
11a14db2 Update format_to taking a buffer and remove undocumented vformat_to overload
832ec098 Fix argument order in locale overload of vformat_to (#2327)
486a80e8 Move wchar_t overloads to xchar.h
19d45f4b Update changelog
5a2b88f6 Reduce binary size
00a39ad5 Enable `Char` types other than `char` (#2323)
ff37e416 wchar.h -> xchar.h because it handles other code unit types too
0901176f arg_join -> join_view
a9a90181 Move wmemory_buffer to wchar.h
4a7801c3 Update changelog
517578f8 Update changelog
85442ed0 Update changelog
6a12b13a Update changelog
1cfe3c73 Update ChangeLog.rst
c0601479 Update changelog
6fe04871 Update changelog
9d67988a FMT_DEPRECATED_WCHAR -> FMT_DEPRECATED_INCLUDE_WCHAR
765b451e Update changelog
17c993c7 Fixed compilation with CMake < 3.7 (#2321)
dde69373 Update changelog
272b0f36 More module tests (#2309)
126c8cb4 Export os.h API, too (#2318)
98b9ff47 Align hex floats right as default (#2317)
ece4b4b3 Update changelog
a70a4ae0 Ignore zero-padding for non-finite floating points (#2310)
7612f18d Update changelog
b9f2c276 Update changelog
4e21baff Simplify get_units
683ef11a Update changelog
ca466374 qualify make_format_args (#2315)
5a2a1856 Make buffers non-movable
ee52a6dc add `fmt::print()` overload to support compiled format (#2304)
82607efb Fixed int conversion warning (#2313)
35a2c2a7 Refactor chrono formatting
b955e7a6 Refactor chrono formatting
883d9595 Support alternative locale names in tests
1f308a3c Update integer presentation types documentation.
1cd9899c Add initial support for weekday formatting
069131dc Add unicode-test
dd8f38fc Cleanup printf API
a216f256 Remove undocumented and obsolete vprintf overload
0c092639 Add is_exotic_char trait
bc13c6de Update README.rst
8ec0b9e3 Do *not* export namespace `detail`
b99c2bd3 Remove deprecated `locale.h` from module interface unit
c04a2439 Update changelog
b099a56f Update changelog
703005c8 Deprecate locale.h
51f01786 Cleanup the format API
5d59dcf6 Remove deprecated aliases / undeprecate has_formatter
c242dd40 Move cerrno include to where it is used
2216e0b7 Update changelog
1c83a49b Simplify buffer extraction
2617384d Improve buffer extraction
34b8acae More wchar_t-specific API to wchar.h
6326c189 Improve code style consistency
5c4b0c86 Add missing Allocator template argument for basic_memory_buffer in format_to
00149c0b Move detail::null to chrono where it is used
c5c968cb Improve binary size
128cbdeb cmake: hide private symbols by default
18af1dc4 Fix binary size regression caused by b268f88
d1e6f0f8 Fix binary size regression caused by b268f88
5a0d99fa Add a test for the module
6e2e6b79 Restore support for `wchar_t` overloads in module
24b677d0 Improve symbol sizes
63271a51 Fix ADL issues
61b4c923 Reduce code bloat
2a2e4c58 addressing nits.
be48f4d6 Avoid unwanted sign extensions from MSVC in is_utf8.
13e65293 export missed symbols
71fb1138 fix compile error on msvc preview 4 (16.10) involving lookup clash /w STL
08d22503 Remove outdated comments
56f518a9 Update signatures
b7f29337 Update signatures
7483dfc6 Update signatures
95c358f7 Improve separation between code unit types
39c3c4ec Simplify the core API
e9c1c415 Improve compile-time checks
21d93bfd Move generic format functions to format.h
9a92eb41 Move more wchar overloads to wchar.h
0dd91e20 Add wchar.h for wide char overloads
ce14eafc Simplify format string checks
8d70c0ed Refactor the format API
813ac495 More API cleanups
4ab01fb1 Cleanup printf API
d5036b11 Remove deprecated APIs
25819462 Cleanup the core API
b35db4e0 Improve handling of 128-bit ints
d35f1ad5 Cleanup core
8f1902c0 Move format string checks to core.h
6469b903 Silence msvc warning about an unused named parameter
7d4c92fb Update ChangeLog.rst
0763d8ca Fix Visual Studio warning
5466373a Do *not* export namespace `detail`
588bdb54 Simplify get_arg_index_by_name
54f22a3e add support for statically named arguments with FMT_STRING
ea94d6d9 Prevent ambiguity in name lookup
57280762 Move specs checker to core.h
ced30375 Move dynamic specs to core.h
dd2bc998 Move specs to core.h
08da1adc Remove unused headers
3be0cc20 Fix handling of 128-bit ints
9648bdce add missing header
d1aebdbd Inline format_to
8f0fadfa Cleanup docs
02896dab Avoid use after move (#2278)
0036a1d1 Fix issue #2274.
2a9b3146 Replace fmt::error_code to std::error_code
2165bef4 Update README.rst
48629308 Optimize format string compilation
3207a8bb Get rid of unnecessary recursion to enable inlining
6214f15a Optimize standard formatter specialization
cd2c78fb Use write directly in formatter specializations
4211d865 Add a formatter specialization for std::error_code.
39f28424 Cleanup tests
84feeb0f Remove redundant comments and put common case check first
2665afb5 Cleanup add-subdirectory-test
d0abe7c2 Make chrono formatting locale-independent by default
50fb0b5e Fix formatting
16f2ef91 Replace fmt::system_error with std::system_error
4b885c86 Replace windows_error with system_error
5238055f Move esoteric char type support to format.h
9ac088f3 Add fmtlog to projects
849c9f61 Move is_name_start to core
23892caf Move more parsing to core
8e6390c3 Move FMT_STRING to core
51a33713 Move parsing to core
9c3af11a Cleanup tests
9d7b53cb Remove redundant formatter specialization for byte
f0095ccd Add support for ranges of types without formatters to join (#2262)
4f0eadfc Exclude fallback from is_formattable
400b953f Use [] instead of {} in ranges for consistency with Python format
38bcc04a Drop range limit and cleanup tests
c738c343 Cleanup tests
ed7c4320 Cleanup tests
9155e2de Cleanup tests
38127d9e Cleanup tests
c9c0e507 Cleanup tests
ccf4ccde Cleanup tests and format string compilation
e96a92f8 Cleanup tests and format string compilation
fd43e4dc gtest: fix std::is_trivially_copy_constructible for GCC 4.8 & 4.9 properly
3d51ccda gtest: remove obsolete `GTEST_LANG_CXX11` compile definition setting
833377ff gtest: add `.clang-format` file into `test/gtest` directory to prevent formatting there
53ca0cbe gtest: move GTest/GMock files to separate directory, update GTest/GMock usages
342973b3 Make wchar_t overloads usable in module Bring ''detail::find()' into scope.
355be4b1 Make FMT_COMPILE fallback on runtime without if constexpr (#2261)
0cd0fb91 C++17: std::char_traits<>::{compare,length} is constexpr - v2
d1a6e560 Keep defaulted destructors inline applies to exception classes in case of msvc only
84a36b99 Move data to functions
ab7c33ed Suppress checked iterator warnings
77258f60 fix FMT_CONSTEXPR_CHAR_TRAITS check for MSVC
d23e315e CI windows: add MSVC C++20 build
f085c3d7 use proper check for non-type template parameters
69bdc20a Workaround missing std::system on iOS, take 2
847aac43 Follow naming conventions in tests
39818e79 Cleanup core-test
0e6f989b __THROW warning fix for e2k (#2253)
1678ed62 simplify field::format() and spec_field::format(), fix typo
ca821982 use named arg with static name in compile-time API
ce6e7d86 use fixed_string to create named arg class with static name for _a literal
fc56af14 move fixed_string from compile.h to format.h
bb006f97 Replace TYPED_TEST_CASE with TYPED_TEST_SUITE
6956b10b Fix gcc 4.8 build
b4f9a058 Update gtest
8f9ddf45 Remove deprecated posix.h
dacd1356 Add module interface unit
d3c523e0 Export printf-related contexts from printf.h
2c25df08 Export replacement type_traits, too
553022dc Don't use std::system on iOS (#2248)
8a040d18 Cleanup core-test
064cac2b Bump version
5b2c740a Remove deprecated APIs
b9ab5c88 Remove printf.h dependency on ostream.h
c47f2112 Simplify data handling
54d3b171 Move more data out of basic_data
128f007b C++17: std::char_traits<>::{compare,length} is constexpr. (#2246)
841aad95 Move data out of basic_data
1d4199f4 fix udl_compiled_string with non-byte chars (e.g. wchar) (#2242)
c5d4fcb1 Appending a space to guarantee non-empty strftime() result. (#2244)
62714062 Fix a warning (#2233)
52bd62c7 Create separate dllexport marking points for clang and msvc. (#2229)
f4bbc54c Tag official API for module export (#2235)
d8910af8 Use qualified name lookup rather than ADL. (#2239)
92601141 Ranges wide strings support (#2236)
24c97515 Try to suppress MVSC warn of narrowing (#2230)
a1c6bfd7 Add a link to llvm diff
42eccac4 Fix clang warning about ignoring __declspec(dllexport) on basic_data<void> template instantitation definition (#2220)
aec50434 Update README.rst
0b411454 Update README.rst
00f3d16b Update docs
99c2f7a3 Allow including fmt/core.h in the header-only mode
b4415323 CI linux: add clang++-11 C++20 (with LLVM libc++) build
1dbadb65 CI linux: add clang++-11 C++20 build
09dbad47 CI linux: add missing build_type
e2facffe CI linux: remove excessive clang++-9 include
273d8865 Suppress redef warning of _CRT_SECURE_NO_WARNINGS if any. (#2218)
5a8bf1f6 Workaround hexfloat inconsistency on windows (#2205)
78776ee4 Fix a conditional expression is constant warning #2210 (#2211)
266107f5 constexpr uint128_wrapper (#2215)
2e0d64cf specify size for `prefixes` static data
95da4847 Fix a link
06b3a100 Add support for time points with arbitrary durations (#2208)
dac42f52 Inline fallback is_constant_evaluated
7c43f8b8 Don't use strlen at compile time (#2205)
c62e4c30 Make buffer_appender default-constructible when back_insert_iterator is
0d6b70d9 Install gcc 8
15c10b0c Add speech synthesis support
308510eb "Use" `fwrite` result (workaround for `warn_unused_result`)
afe23e7f Don't call fileno on NULL file in tests (#2196)
b49af043 Remove noexcept from file's move assignment
14848875 Fix: fmt::ostream cannot be moved while holding buffered data #2197 (#2198)
7d8c3401 Update pull_request_template.md
b966afcc Remove formattable
ec5315a9 Use strlen when possible in fallback basic_string_view
4f8778ba Inline basic_format_args's ctor
e2d87548 user-defined constructor
f7151d38 Extra flag to prevent Intel compiler with Clang front-end warning of 'unknown attribute no_sanitize'
0fb8ef8f Inline trivial argument handling functions
1b23e25f Simplify formattability check
35c71ff5 Only use -Og with optimizations disabled
243d8beb Enable minimal optimizations in debug mode
9b34681d Work around xl compiler bug when nvcc preprocesses this file (#2190)
4dc7170d Fix C++17 builds: (#2192)
9cb347b4 Simplify argument formatters
0f85a468 add default cases (#2186)
417e1cee Stop using deprecated UDL templates
f7e900e1 Simplify UDL definitions
d9661c8f Mark grouping as deprecated
14a2a64d Fix handling of formattable types with to_string_view (#2181)
6ae402fd Fix handling of types with to_string_view and formatter specialization (#2180)
a6408a3b Add args-test
1147782c Fix an ambiguous call to check caused by ADL (#2184)
2f3f3862 Fix harmless MSVS warning about using undefined _MANAGED symbol (#2183)
d0bded59 Fix MSVC /clr builds (#2179)
8308f52c Fix dynamic_format_arg_store::push_back comment
6151d0dc Fix the comment
5a1127b7 Don't wrap named arg in cref and clarify docs
b8ff3c18 optimize append (#2164)
c8d8b882 fix GCC 7,8,9 warning about unused but set parameter (#2177)
d2810187 Document ostream support limitation
bac14ef9 Simplify integer spec checking
8f9db3fc Make ubsan happy on empty format specs (#2175)
af567538 Bitpack integral prefixes
cdf877d4 Workaround missed optimization opportunity
eef4ba9c Optimize integer formatting without padding
a1ea8a82 Unbloat my heart
a457e163 Simplify integer formatter
05bc87a6 Optimize padding
605b6037 Optimize count_digits for powers of 2
85ba2716 Implement 128-bit count_digits in terms of count_digits_fallback
d9835737 spec -> specs
f9e0e904 Apply clang-format
60f5d244 Simplify arg_formatter
30e1302e Simplify on_format_specs
87c5cd46 Optimize parsing of argument ids
6a9016ea fix `formatted_size` with "compiled format" as argument (#2161)
6e1fc017 Move detail::truncating_iterator to fmt/compile.h
e718ec3e Make truncating_iterator an output_iterator (#2158)
772aeca3 Don't include <cassert>. (#2148) (#2152)
684b5b0e Fix fallback to runtime API from compile-time API (#2143)
d8b92543 use simplified `void_t` for all compilers other than gcc 4.x (#2160)
835b910e Add an is_formattable trait
57887403 Revert "Optimize handling of integer constants" (#2147)
640acba8 Print x.what() of FMT_THROW when exception is disabled (#2145)
d8e1c9f1 fix `fmt::get` for some GCC versions and legacy Clang (#2144)
2797588b Optimize handling of integer constants
e8eff3b8 Fix FMT_STATIC_THOUSANDS_SEPARATOR (#2142)
ab0f7d7f use const& for arguments
29cc8282 update chrono duration formatter (constness), use it in compile-test for specs checks
3f69af3a update wording in the error inside `arg_id_handler`, use `FMT_ASSERT` instead of `throw`
499047e1 fix incorrect indexing mode for named args, update tests
78c67157 prepare tests, fix incorrect handling of named args with simple `{}` replacement fields
b31bc2dc simplify `try_format_argument()`, make `manual_indexing_id()` a variable
95e1aa2d add support for manual indexing and named fields, add tests
7e72673d Improve width estimation (#2033)
13b117b5 Improve code point computation
ee0fed63 Fix handling of the + flag with locales (#2133)
c5979d56 Fix fmt::localtime formatting not working in wide-char string contexts
e6ef927e fmt::ptr: Support function pointers (#2131)
58aa0457 Fix ordering of install commands for CMake (#2122)
1980ca8c fix#2118: FMT_COMPILE did not work with tm formatter (#2119)
2a25e2bf Make ranges-test available with C++11 (#2114)
b0b56b43 fix#2116 (FMT_COMPILE requires exceptions enabled) (#2117)
373262f9 Update docs
ce519e93 Fix exception propagation from iterators (#2097)
acef0bb5 use gcc-10.2 instead of gcc-10.1 on CI, also fix one problem (#2110)
8bf28e6b Add support for s format specifier to bool (#2094) (#2109)
9c418bc4 Update README.rst
456efa46 add missing detail namespace (#2107)
80dc7cce Fixed format.h(1465): warning C4702: unreachable code (#2106)
7fd535c6 Cleanup 'L' handling
b4b8917c Update docs
e4f2cf45 Make 'L' a modifier
6972b5f3 Add build variable: FMT_MASTER_PROJECT (#2100)
ac352081 Install fmt/args.h (#2096)
532e846b Fix width computation in float formatter
f8c2f848 Fix handling of width when formatting int as char
0fe0b15e Fix handling of # in width computation
061e364b Document output_file
018688da Correct a typo on syntax.rst (documentation) (#2081)
9ec5592b Fix writing to stdout when redirected to NUL on Windows (#2080)
cdc5ef67 Remove fallback to inline specifier from FMT_CONSTEXPR(20) macro (#2075)
c9dd1eb9 Don't change charset
d09b5c14 Fix std::byte formatting with compile-time API (#2072)
bbd6ed5b Add support of most format_specs for formatting at compile-time (#2056)
a750bf3a Update api.rst
1256541d Fix formatting
4fa4c924 Add tests for FMT_ENFORCE_COMPILE_STRING, fix several errors (#2038)
aa89e380 add cwchar to format.h for std::fputws (#2073)
5a37e182 Disable warning about format string (#2067)
fa43fd14 Forward arguments to work with views (#2068)
3551f5d1 Workaround a gcc 10 -Warray-bounds bug (#2065)
e7376726 Remove an old mingw workaround (#2059)
25a41b80 Fix a link to Android.mk (#2057)
9293f707 Suppress gcc warning on privates-only class (#2053)
c20874c2 Reenable support for fallback formatter in join (#2040) (#2050)
5de0bc1d Add UDL as replacement for FMT_COMPILE (#2043)
a6fafe2f docs: use https for some links (#2051)
33f9a6d3 Fix handling of enums in to_string (#2036)
aabe0a84 simplify tests by reordering arguments of `EXPECT_EQ` (#2044)
1f4a76d2 Add a missing include (#2047)
4a6eadbd Make std::byte formattabe (#1981)
683a7450 fix formatting with empty compiled format string (#2042)
f43416e1 Add a link to contents from index
5a493560 Move some code from core.h to format.h where it is used
9ed0a981 Fix docs build
dac753b8 Basics of formatting at compile-time based on compile-time API (#2019)
119f7dc3 Truncate file by default
22a68d16 Don't emit trailing zeros by default
d0110b7e Update README.rst
3f4839ce Merge branch 'release' of github.com:fmtlib/fmt
7bdf0628 Update version
fc135511 Update changelog
926233bd Fix test
0683fa7d Bump version
6ce207b9 Fix formatting
07b1c1a1 Update changelog
58992761 Reintroduce ostream support to range formatters (#2014)
b8957f50 Fix an overflow in format_to_n (#2029)
df66516e Workaround an issue with mixing std versions in gcc (#2017)
a57baa69 Fix more linkage errors (#2011)
85534a13 Fix linkage errors when linking with a shared library (#2011)
a2fa5d62 Update changelog
cd300368 Fix more linkage errors (#2011)
d1ef29d6 Fix initialization of iterator_buffer (#1996)
5f41bb0f clang-format
a58a6b27 Add a newline
a036cc97 Reintroduce ostream support to range formatters (#2014)
38c7def4 Update clang version to 3.4 since there are ICEs on earlier ones
55336413 🆕 [CI] Test with C++14 in Windows 2019 (#2020)
55dfdd92 Update README.rst
2c734c9b Fix an overflow in format_to_n (#2029)
6cdd1be9 Update build.gradle for latest AGP (#2026)
bcc20b29 Implement compile-time checks by default
befd7d4a Always use FMT_STRING internally where possible [Issue #2002] (#2006)
f8640d40 Add more standards
f81c14aa Workaround an issue with mixing std versions in gcc (#2017)
5555651c Fix more linkage errors (#2011)
b268f881 detail::write in one more place relevant to printf with long argument… (#2016)
aa9b09a9 🐛 Cannot call non-constexpr function in constexpr context (#2010)
986fa004 Printf get container (#1982)
7abc3c01 Suppress a useless warning (#2004)
6d14f781 Fix linkage errors when linking with a shared library (#2011)
9534b9fe Refactor warning suppression
60dc2735 Simplify on_text
b5dac0f0 Reduce <algorithm> usage (#1998)
a07627b1 🐛 Implicit sign conversion warning in clang in c++17 and 20 modes (#2009)
1b8f499e 🔧 Silence useless cast warnings (#2008)
f428d286 Update README.rst
beb248b6 Optimize handling of large format strings
1936dddc fix gcc warning of missing override (#2001)
14f6bd0f Move one more headers to args.h
e01d26e1 Optimize includes
e528d919 Merge branch 'master' of github.com:fmtlib/fmt
48816772 Update signatures
3302fd10 use memchr for searching for '%' in printf format string (#1984)
4c2d6372 Update signatures
beaff396 Update signatures
ffa0a083 Use newer versions of Sphinx and Breathe
038057eb Document contexts
5bedcb66 Fix initialization of iterator_buffer (#1996)
2435ea41 Workaround MSVC mess
8c6215f5 Fix fmt/color.h
10ebe6cb Document color
1ac50fcb Suppress more bogus warnings
e098be8e Fix warning filtering
8cf0afaf Improve docs
e29f93e8 Suppress more bogus warnings
4e8d000f Suppress more bogus warnings
7787792e Fix re usage
6ee5e507 Fix imports
06ee32d1 Filter useless doxygen warnings
86bb7fe6 Add a missing import
959a9f5c Merge branch 'master' of github.com:fmtlib/fmt
4f7df299 Improve docs
b3ab0bc7 🎨 [CI] Specify the exact version of clang to use (#1991)
701ed6c8 Install deps in github actions instead of script
8f2131cf Document chrono
32c4af8f Document chrono
295a60ec Document chrono
a4fae96c Document chrono
263bb0e6 Document chrono
0506b328 Document chrono
4e426c19 Document chrono
9795d873 Update docs
2eb0be0b Remove debug code and fix bot contact
cd955798 Move less installation to actions
98639d0f Debug doc build
ab5e0632 Debug doc build
b123129f Dump the content of html dir
81d2b986 Print less command
7a0b1d57 Add key
9f0617cb Fix branch ref
75b07598 Chrono docs
dfbb6975 Remove travis config
5b3052f9 Switch doc build to github actions
506ff320 Fix build failure when not using fcntl with -Werror (#1990)
a30b279b Apply clang-format and tweak comments
6a2495c8 -Wattributes visibility warning with some GCC versions (#1975)
cba5970c Remove migrated build configs
689081d8 Merge branch 'release' of github.com:fmtlib/fmt
cc09f1a6 Update version
e4eb242c Update changelog and bump version
ce98e0c6 Fix fallback float formatter at assymetric bounds (#1976)
49544ea9 Fuzz fallback formatter
6b7bfed4 Fix fallback float formatter at assymetric bounds (#1976)
bcab36da Update CI config
1689e73e Move PR template
0103408a Update CI config
38a16ecb Move build config to github actions
205eb3a8 Update CI config
fe61b8c6 Update CI config
867b15d7 Update CI config
98cb9f99 Update CI config
95077d60 Update CI config
bc49f094 Update CI config
cef6dfb4 Update CI config
c8703ba4 Update CI config
ab4405be Update README.rst
78a55e28 Update CI config
d0a2494a Update cmake.yml
89d009ba Update cmake.yml
1f4ff47b Create cmake.yml
eb52ac7a 🆕 Enable -Wshadow in pedantic mode
e904e891 🎨🐛 Rename all shadowed types and variables
771292c3 Remove sizeof from unused variable silencer (#1974)
86bf6045 Merge branch 'release' of github.com:fmtlib/fmt
5f7f7b95 Update version
5d3f0741 Update changelog and bump version
563cbb6c Add a macro to workaround clang/gcc ABI incompatibility on ARM
425778aa Fix ABI compatibility (#1961)
69a84198 Remove accidental parenthesis (#1968)
5c045049 Removed [-Wsign-conversion] warning in GCC
556a1cfb Instantiate to_decimal to make gcc lto happy (#1955)
28a8eae8 Cleanup
236fea1f Workaround bugs in gcc 8
e50ced88 Add a macro to workaround clang/gcc ABI incompatibility on ARM
112755cf Remove FMT_SAFEBUFFERS (#1966)
4081b2fe Fix ABI compatibility (#1961)
2d9311e8 Remove accidental parenthesis (#1968)
b3a4f28a Fix implicit signedness conversion warning (#1963)
97c88732 Allocator::max_size support in basic_memory_buffer (#1960)
bb68f608 Removed [-Wsign-conversion] warning in GCC
f4ca065c Range support
cb224eca Instantiate to_decimal to make gcc lto happy (#1955)
7977c2b4 Cleanup
e54eb676 Workaround bugs in gcc 8
4fe0b111 Update version
df4bd60f Bump version
764fb35e Always install the required version of breathe
e1bdc0ec Use the correct version of sphinx
39bde329 Tweak markup
204d299a Tweak markup
e0995b1c Update readme
4af178bd Remove outdated build config
aa41dc02 Remove unused script
6a77ea3c Tweak markup
62c72059 Update changelog
c10e3f7f Update changelog
e542e695 Update changelog
530cf316 Point to the release, not dev documentation
740385d6 Update changelog
cd465111 Update changelog
46291be3 Update changelog
90071c1d Update ChangeLog.rst
25293d7a Update ChangeLog.rst
5024742f Update ChangeLog.rst
0452a4e7 Update changelog
8de96817 Woraround bugs in gcc 8
47e16767 Simplify arg formatter
f0a42346 Move parsing optimization one level up
86287b8d Optimize common case in parse_format_specs
8924211f Update README.rst
525e7649 Update CONTRIBUTING.md
0ecb3d18 Optimize alignment parsing
97553078 Optimize format_uint
7446818f Simplify vformat_to
280b5612 Add option to force usage of inline namespaces
e57ec7d5 Merge vformat_to overloads
2a3f4de3 Remove iterator_category
27fdb4ea Unshadow floaty
297e0bad Apply clang-format
e3b4c22e Simplify is_output_iterator
da8278e1 Update changelog and bump version
17fba753 added position independent documentation (#1939)
71e705a2 Update README.rst
74654c8c Fix compilation for systems without fcntl.h (#1942)
f468b203 Avoid conversion from long long to size_t (#1935)
20d4f2e8 Fix handling of weird character types when parsing sign (#1932)
08370c39 Update README.rst
bd3c7925 Fix float fuzzer
8d3fd86d Merge branch 'master' of github.com:fmtlib/fmt
40347157 Update README.rst
37d738fa Update README.rst
271eff14 Make classes derived from buffer<T> final to silence the virtual destructor warning. (#1937)
010efc31 Add float fuzzer and cleanup
811c8b58 Add float fuzzer and cleanup
82c4e223 Cleanup fuzzing
63e40c96 Fix naming of fuzzers
2f448ed5 Fix fuzzer timeouts
af283059 Cleanup
48ea8193 Explain why assert-test is a separate test
1d112bdd Remove old test
5eb292a6 Update README.rst
7e56b6b6 Fix coding style and remove duplicate fuzzer
41d97e1e Fix a UB on ridiculously large precision
01c37e0a Added check for `-mbig-obj` and ref qualifier check (#1929)
a5e7e7db Fix handling of thousand separator (#1927)
bf19051a Optimize floating point formatting
3c13a88b Optimize floating point formatting
f6d75c53 Refactor write_float
e9c0b2d6 Merge write_float overloads
7eddbfed Cleanup exponent handling in write_float
b347b302 Update dynamic_formatter comment (#1923)
3541880e Fix integer overflow when using max int precision
7b50dc0b Don't exclude all detail symbols from docs
28052431 Fix the doc config
34f22e88 Cleanup CMake config
a18b3fbb Fix fixed precision handling when rounding (#1917)
72770357 Fix long lines in usage.md
7612c1ea Add reference to lhelper package manager in usage
b91d39f2 Get rid of float_writer
b4b64b9c Refactor float formatting
712abe40 Workaround a bug in gcc 7.5 (#1912)
af8a180a Make GetCachedPower test more precise
a581e9e5 Fix warning C4018: '<=': signed/unsigned mismatch (#1908)
05a28312 Update docs
4d0aa4d8 Update link
575f4018 Simplify FP formatting and follow coding conventions
6f3536f9 Move zero-check to an earlier branch (#1906)
90ef46df Fix dragonbox integration
3ae88147 Fix declaration
64179525 Improve dragonbox integration
79694d42 Fix WriteConsole signature
51f2e2ca Move nan test to where it belongs
68555fdb Make format-test not depend on color.h
63e0c354 Make dragonbox::to_decimal available in format.h
2213a711 Update README.rst
79ba37f3 Update README.rst
a905d8f7 Merge grisu-test into format-test
762c33a9 Simplify windows handling (#1903)
253d6315 Remove dependency on windows.h (#1900)
c156093f Fix carry in fallback_format
34179b33 Update format.h (#1898)
0651e459 Minor tweaks to get_cached_power
6c025520 Test that max_k is correctly defined
51f8d0cc Reuse log10_2_significand constant
1305cbeb Fix MSVC2019 error C2049 when compiling with /clr (#1897)
2d4fde3a Don't emit trailing zero for consistency with std::format
5fd89d50 Minor simplifications
605ce5e4 Simplify divisible_by_power_of_2
085171e7 Remove grisu_count_digits
aa729bf2 Remove dead code
aa2ddf9b Simplify Dragonbox integration
c1654ce4 Simplify uint32_or_64_or_128_t definition
33712dc0 Combine pragmas
e5942ac9 Tweak comments
aae7a133 Remove unused pragmas
6bcde9aa https://github.com/fmtlib/fmt/pull/1882#issuecomment-696823912 (#1894)
bb0db5e5 clang-format
16410056 Optimize copy_str for counting_iterator
2591ab91 MSVC optimizations for count_digits. (#1890)
d5b8002d Update README.rst
821471e1 qkw: generalizing aliasing | using fmt library and it's features (#1888)
2e620ddb Small improvements that should have zero to negligible impact on the runtime (#1887)
2f7e0885 Disable range formatter if value type is not formattable (#1885)
c46a8de4 Simplify test
2696dc92 add forgotten template argument to make_format_args which made some u… (#1877)
0016da7a Don't generate zeros and fix UB on huge precision
ce3f7699 Merge intrinsic blocks
3b6248f6 Change formatting
2d9b1dd0 Fix sign mismatch
1f0600a2 Fix bug regarding FMT_SAFEBUFFERS
2ecdbb98 Fix a bug in ctzll
6f81ea15 Fix typo (and thus bug)
0c8ffe9b Implement Dragonbox (first version)
42699bf4 Fix msvc version of clz & clzll (#1880)
bc51a8df Disable fallthrough attributes for the Intel compilers on Linux and MacOS (#1879)
45da432d fix compiler warnings in public header files
d55e61f1 Improve FMT_ALWAYS_INLINE (#1878)
7e682752 Remove trailing zeros when using fallback formatter (#1873)
1d696dc2 Handle exotic character types in compilation
f674434a Add format_to_n overload that accepts FMT_COMPILE (from #1767) (#1869)
5b5a5971 Fix handling of wide alignment
f80ed64d Update README.rst
38139664 Simplify fallback format
dce8e49b Handle float in fallback formatter
78b59443 Spelling
f233b56c Don't generate insignificant digits
595902f8 Update test
4f2ee892 Use built-in FP formatter for any precision
58a044be Use built-in FP formatter for any precision
efe3694f Macro tweak and clang-format
9f312fe8 Implement fallback FP formatting with given precision (#1526)
fb289cf5 Fix coding conventions
86f0a704 Fix formatting
bff4d18e Add color format_to overloads * Fix variable size basic_memory_buffer colorization * Fix an unused arguments warning on GCC that blocks the CI otherwise * Ref #1842 * Ref #1593
f19b8885 Fixed a warning in mingw32/mingw64 (#1860)
f8e00a08 NOMINMAX not handled properly (#1855)
6cccdc24 Fix move constructor (#1844)
69902c17 Allow use of <fcntl.h> in Linux when __has_include is not available (#1848)
1edd38b9 Add append mode. (#1847)
e66ba169 Added build2 usage instructions. (#1838)
f39e6fb6 Add formatters for chrono::time_point<system_clock> (#1837)
77b627be Fix bogus MSVC warnings (#1825)
5dff01d3 Add complex tests
d16d585e Update signatures
c7e6d8af Fix usage of override (#1836)
92bff2fe Revert "Add missing includes"
a0dcfbc5 Add ptr to docs
1651b2d4 Fix detail::write with fallback formatter (#1829)
06895a76 Add missing includes
92a448a0 Apply clang-format
6be65446 Fixing buffer_appender's ++ slicing (#1822)
951e0d23 CMakeLists.txt: Added Wundef warning to clang and gcc. (#1823)
f9f02df7 CMakeLists.txt: Clang-warnings: removed -Wno-sign-conversion (#1817)
76e97dc4 Eliminate shadowed variable warnings on intel (#1816)
e204df0e nvcc compiler should be EDG-based, but fails test (#1818)
1c8bb547 small changes to reduce clang-9 warnings (#1808)
4b69c787 fix: warning C4100: unreferenced formal parameter (#1814)
fb0aeb82 fix: disabled UDL templates for PGI (#1811) (#1812)
54daa086 Add dynamic width support to FMT_COMPILE (#1809)
6fb7c6fb Workaround a bug in gcc10 (#1810)
16985fda Update README.rst
1378ddae Update README.rst
4fd95e4b Don't remove trailing zeros with #
e06ae322 Avoid warnings on functions with external linkage that don't have declarations
7fc3d1f5 Add override to grow
065889a5 Use correct capacity in iterator_buffer (#1807)
d0dd6786 Adding convenience append(range)
0e7cef06 Merge commit 'c13f79e0'
e2c8c455 Update README.rst
e4c954ff Update README.rst
c13f79e0 Merge release branch
d7921d64 Update README.rst
4a4fc225 Update changelog
61602a75 Remove -Wno-shadow
2f8fc29e Update README.rst
717b226b include/fmt/format.h: explicit cast to std::size_t for parameter to buffer.resize() in order to get rid of warning 'implicit conversion changes signedness:' in clang-8 (#1802)
2a69f567 Tweak buffer size
ea769338 Simplify ostream
5413713c Remove unused function
57f46242 Increase the default buffer size
0b6e7cc6 Update README.rst
e587adb4 Simplify count_digits
279d698e Fix handling of default alignmment with locale (#1801)
76cfb50b Test complex formatter
20829120 Optimize count_digits
8d9ab967 Cut a few cycles from count_digits
73434493 Simplify ostream_params
2a47a1e4 Update README.rst
7c4c5c79 Make buffer size configurable
f0b84da5 Don't use 128 bit integers with clang-cl (#1800)
a3dfd6f9 Workaround a bug in msvc
51d05521 Workaround broken numeric_limites, part 2 (#1787)
21c8b5c1 Report error on missing named argument (#1796)
d82fdcc9 Fix handling of iterators in locale-specific formatting (#1782)
633213d9 Merge release branch
e8f2580a Bump version
6cefe55a Update changelog
64e2da15 Update README.rst
1c8c810f Update README.rst
c2399ccf Update README.rst
a7c5db06 Update README.rst
a4c22acd Update README.rst
0c1f4b5a Update README.rst
63b422ee Update README.rst
26e81a67 Update README.rst
de5fc6af Update README.rst
9c2edfd1 Partially revert 638db5 because it breaks the doc build
810357c0 Document color
0a7032a4 Update README.rst
95d3abf9 Make format_to_n part of the core API
98626093 Correct the locale format specifier in api.rst (#1792)
47f8d7a3 Make formatted_size part of the core API
46a63b70 Update docs
430f393d Disabled __attribute__((deprecated)) usage for LCC (#1790)
febffa4e Make join() handle non-const-only begin/end ranges (#1786)
d69e2da2 Fix apidoc
ce73ea37 Reorder functions
d39d661b Workaround broken numeric_limits (#1725)
c228bfe8 Improve docs
38ce19f7 Update README.rst
d11849bc Add FMT_REDUCE_INT_INSTANTIATIONS flag (#1781)
c08518a2 Move make_args_checked to the public API
e2837084 Add a color section
9f0c0033 Simplify format string checks
d615137c Improve handling of buffer iterator
26b47b6f Bump tested CMake version to 3.18
7a01c9c5 Update README.rst
b17d5c4f Fix a regression in handling digit separators (#1782)
eb90da2e Type erase output iterators
9d3cd0af Type erase output iterators
18024853 Fix compatibility with CMake 3.4 (#1779)
f5d4215b Trying to clear ambiguous compile time claims (#1775)
c26349f4 Improve error reporting
f4b11ef6 Add a short anchor
0097cf11 Report unformattable type name more prominently
8fa20b47 dev -> latest
a03bd3dd Autodetect MSVC static runtime (#1770)
c108ee1d Clarify a comment
a8074a86 Update README.rst
5f629548 Update README.rst
bd903f96 Clarify precedence
16cac46a Improve handling of streamable and convertible to bool types (#1766)
415cd519 direct_buffered_file -> ostream
e1bfb596 Fix handling of code units in compile
ba8d98cb Cleanup direct_buffered_file
04a1f6e9 Improve handling of single code units in compile
e4f57bfd Add an overload of write for buffer_appender
d8704681 Make append work with fixed-size buffer
e8ec09ae Cleanup core-test
a2c4fed9 Double buffering no more
36406509 Add a fixed buffer
60c43e87 Apply clang-format
b998e0f3 Reduce symbol sizes and simplify iterator use
c5adfc51 Update README.rst
c4ad94ce Update README.rst
c1429651 Fix image source link
638db5ca Use Cmake to find Python and Sphinx-doc.
c0905697 Update readme
1efdb2dd Simplify readme
dc69afad Cleanup example
445f5d39 Break long lines
23063c34 Update readme
f57b6257 Move PR template to the top level
git-subtree-dir: externals/fmt
git-subtree-split: 9e8b86fd2d9806672cc73133d21780dd182bfd24
Removes dependency on the constants at the top of some files
such as `f16_negative_zero` and `f32_non_sign_mask` in favor
of the `FPInfo` trait-type.
Also removes bypass delays by selecting between instructions
such as `pand`, `andps`, or `andpd` depending on the type
and keeps them in their respective uop domain.
See https://www.agner.org/optimize/instruction_tables.pdf for
more info on bypass delays.
