Makes bracing consistent, and variables const where applicable to be
consistent with the rest of the codebase.
In most bracing cases, they'd need to be added to conditionals that
would involve checking stack pointer alignment in the future anyways.
Generally hash functions shouldn't throw exceptions. It's also a
requirement for the standard library-provided hash functions to not
throw exceptions.
An exception to this rule is made for user-defined specializations,
however we can just be consistent with the standard library on this to
allow it to play nicer with it.
While we're at it, we can also make the std::less specializations
noexcpet as well, since they also can't throw.
To a reader this might look particularly strange, given the function
itself has a void return value, but this is actually valid, given the
function in the return statement also has a void return value.
This instead alters it to be a little easier to parse and potentially be
a little less confusing at a glance.
These are cases that are defined as reserved within the ARMv8 reference
manual, so we can handle them as such instead of as unallocated
encodings.
While this doesn't actually change emulated behavior, it does at least
allow the JIT to generate the more appropriate exception.
Strictly speaking, immh being zero is defined as reserved in the ARMv8
reference manual. This was just an error on my part when introducing the
SQSHL immediate scalar variant.
Due to promotion rules (types < int, even if unsigned, get promoted to
int when arithmetic is performed on them), this is a potential spot for
undefined behavior.
9e554999c Update version
b34d92b05 Bump version
d39ece187 Make rst2md runnable and update changelog
fe2d715ff Update changelog
27b306701 Update changelog
68837079a Update changelog
c98b202eb Update changelog
587a7f663 Update changelog and docs
84e5170c9 Update changelog and deprecate visit
130e412b6 Update changelog and docs
0bbdca5b8 Fix conversion warnings (#989)
77a724480 Implement fill/align/width for strftime-like formatting
3e01376e0 Implement fill/align/width parsing in chrono formatter
1f92f8a9d Remove noexcept
8668639ae Get rid of null_terminating_iterator in format
93fd473b8 Add support for builtin terminal colors. (#974)
61ad543c3 Windows .sln filename changed from FORMAT to FMT
7f7504b3f Clean up docs
37f599b1a Fix docs
8c2e15aed Make printf work in search (#164)
de71db6d4 Fix asan error (#977)
b180b3915 Fix default formatting
24594c747 Disable printing the reset escape code when no style modifiers where applied. (#973)
b0f222471 Implement default chrono formatting
749276072 Add file stream support for stylized text printing. (#967)
f54f3d0fb Move chrono-specific code to a separate header
bf1f1c73e Fix time test
b6bc6ec24 Add default ctor and fix use of constexpr macros in text_style
acfa95d4a Workaround a bug in MSVC's strftime (#965)
628f83058 More chrono formatting
aa3b5aba4 Implement locale-specific minute formatting
639de2175 Workaround more MSVC bugs
3242ddf7b Fix warnings
bd1104046 Workaround a bug in MSVC
81b5c4a5f Add experimental emphasis support (#961)
7c4eb0fbe Fix warnings in time.h
2d624218b Fix another warning
b31680990 Fix a warning
b10ccb83e Add rpclib to projects
0497875ff Stop the orgy of casts
37dc495b9 Simplify MSVC workaround
2ff4996d0 Fix ambiguous complier error C2666 in vs2017.The '+' opeator may cause ambiguity.Avoid implicit conversion.
77656c672 Fix sign-conversion warnings reported by Clang7
ea5e4790b Fix formatting
86681c4bb Update README.rst
e867768ee Do not override user provided compile flag
0c7f5c3ca Update README.rst
e7e2ab107 Make return type of basic_format_args::max_size() consistent.
29352af36 Update README.rst
68214bd90 More time formatting
bcf3fcd67 Clean up bit fiddling for argument packing
9dcf127fa Workaround a bogus MSVC warning
b8b06e3e1 Fix conversion warnings in Grisu
322b2594e Implement more time specifiers
0835f1ba3 Use full paths for fmt.pc.in
a084495d7 Add Ceph to projects
fa1d4dbcf Fix warnings
2b2cfdac1 Update docs
99744f8f8 Suppress unfixable warning
f5fe84923 Specialize formatter for chrono durations
a5a9805a9 First stub at the datetime format parser
645c76a9a Fix dummy warnings
fecb2d6f0 Eliminate msvc compiler warnings (#931)
64690d3a9 Add context_base::arg()
01640f44c Fully qualify dummy_int (#941)
e37d6a984 add make_printf_args and make_wprintf_args functions (#934)
982ee5c69 parse_context -> format_parse_context
b7b854855 thousands_sep -> thousands_sep_impl (#939)
00a8cc832 Fix formatting
33fbb3a7e Fix remaining linker errors.
bd6121596 Disable fmt-impl-test in windows + shared lib.
702b3d161 Fix link error in windows with shared library.
9d4ef9435 Install pdb files.
6c95fb356 Default Context to format_context
16b78ee62 fix incompatibilities with c++2a mode in clang
19e008876 More locale support
f2ee98810 Improve locale support
1385050e2 More formatter tests
03c1b110a Fix gcc 4.4 build
cc805c616 Test enabled formatters
e01579231 Disallow leading zeros in arg-id
34030deca Cleanup warning flags
6b26e3f2d Manifest & Gradle comment
d286c9775 Update for Gradle build
d951f6dfe Get latest Gradle (ver. 4.10.2)
a23d59247 Fix check_format_string (#925)
36161284e Update docs
38f355d87 Revert "find sphinx-build before calling build.py"
324eac1aa Make locales work with any character type
bdda4d603 Simplify compile-time strings
5ee1a4bc8 check for property 'mutable iterator' and SFINAE on it
2dea780fb change type naming and fix sfinae bug
b98e8301d add non-char support for compile-time format check
ccd3e8bbf Make is_constructible public (#918)
437315380 Update usage.rst
73cfd8f32 Fix colored print
ec384302d additional test for print with background color
0a96c032b Parameterize v*printf on string type (#920)
61e6d2e38 Fix core version of vformat_to
ea4010d70 Merge has_to_string_view into is_string
486fff597 Add sprintf_format instantiations and remove syntactic noise
1e3dcbba8 fix: 'format_to_n' compiles 'std::back_inserter' arguments
f0328f8e3 Use char_traits::length in string_view ctor (#914)
895fb9845 Disallow gcc 4.4 failures
20c708bf6 Fix build on gcc 4.4
9d0c9c4bb cmake: output share/fmt.pc
2d2326a76 Fix compilation with older gcc
1ec027230 Get rid of FMT_UNION
2c81c851b Adapt any string-like type to be used by {fmt} just like the standard string types already supported. The adaption is totally non-intrusive.
846c644e8 Workeround broken sprintf in MSVC
13d472bd8 Compute output size for grisu
b71d3fe7a Remove use_grisu
847abb6f8 Fix test
dda47c946 Merge min_digits and max_digits
292462215 Fix naming of basic_format_specs members
bda5f9a55 Replace grisu2_specs with core_format_specs
b1ca608ba Remove unused empty_spec
e8efdef8d Avoid extra copy
98f1c1fe8 Remove unused code
50b18a3c1 Integrate Grisu
699297520 Implement Grisu rounding
4bb76ef0c Remove redundant definition of print
ddd7caf38 Fix locale-dependent formatting (#905)
10e03e695 use found python executable for launching sphinx-build
07200f445 find sphinx-build before calling build.py
08a65c228 Workaround broken constexpr in MSVC2017
167f8fe32 Fix a typo in api.rst
57983423c Remove signbit workaround
7bebb3e12 Clarify overload resolution in docs
939fbe556 Remove basic_fixed_buffer.
61f81a071 minor documentation corrections
f27defc63 Parameterize printf functions on the type of the format string.
6a685571d Make 'std::*::basic_string_view' a valid argument type for 'format_str' parameters.
87a0408c6 Fix ostream.h build
2b5acad4a Remove redundant size argument to write_padded
655ce5338 is_format_string -> is_string
fea712abb Parameterize ostream functions on the type of the format string.
f16a118e8 Fix non-matching char types.
041bf83d9 Improve fmt::format readability
229903239 Document how to write a formatter for a type hierarchy
f5480635c visit -> visit_format_arg
cdf3fa08d Put related code together in fmt/core.h
38325248e Count width in code points (#628)
deb901b9e Parameterize core functions on the type of the format string.
0f98de301 Update docs
c797708fc Workaround strlen being non-constexpr in ARM toolchain
49b4c1e9d Update docs
63a87beba Add to_string_view
4e0c31465 checked_format_args -> checked_args
c3538a1ee Simplify variadic functions further
2d7d0835d Simplify variadic functions
3f4cfa6c6 Implement UTF-8 string support
f8027414f Impelement char8_t support
76a47d41c Cleanup the use of FMT_CHAR
267fdc7a1 Parameterize core functions on the type of the format string.
5bced1242 Parameterize more functions on string type
674999c52 fix vs2017 warning fmt::v5::localtime 'not all control paths return a value'.
e4fea22d1 Make char8_t a strongly-typed enum
66992e90d Clarify that writing to memory_buffer appends (#877)
e864acfdb Fix compilation with intel compilers (ICC/ICPC) v14.0
4cf21f58b constrain templated format_to on proper format string type.
d7f17613f Fix compilation on platforms with exotic double (#878)
e4ca37ccf Parameterize format_to on string type (#880)
d66fa2216 Reduce syntactic noise
48e6dcd0f Implement workarounds for gcc 4.4
0ea3221d3 Remove is_named_arg and add FMT_CHAR
73c53d783 Parameterize 'printf(rgb color, ...)' and 'vprint_rgb(rgb color, ...)' on the type of the format string.
d41be23ac Simplify string_view detection
2def9e4c8 Remove FMT_DTOR_NOEXCEPT
ff6e46ed9 More cleanup
715f2b4c0 Remove require_wchar and internalize no_formatter_error
ec0cdc46f Workaround Windows slowness
git-subtree-dir: externals/fmt
git-subtree-split: 9e554999ce02cf86fcdfe74fe740c4fe3f5a56d5
Like FCVTN, now that we have half-precision floating point conversion
functions available, we can go ahead and use those to eliminate the
interpreter fallback.
This is supposed to call FPUnpackBase instead of FPUnpack. This would
result in alternate half-precision representations being misinterpreted
when it comes to dealing with NaNs.
Avoids potentially clobbering the intended sign bit value during
conversions to double-precision values. The other conversion types are
already properly handled, so those don't need to be addressed.
This just transfers values between vector registers and general-purpose
registers with no conversions performed, so this is trivial to add
support for half-precision to.
This corrects one case where floating-point exceptions could be set when
they're not supposed to be.
This also corrects a case where values were being treated as NaNs when
they weren't supposed to be.
This function is defined as always disabling the AHP bit in the fpcr
before performing any operations.
At the same time, rename the original FPUnpack function to FPUnpackBase
to match the pseudocode in the ARM reference manual.
Corresponds to the equivalent pseudocode within the ARMv8 reference
manual. This will be necessary for supporting half-precision
floating-point.
This also makes use of it within FPConvert
Adds a template function that performs the same behavior as in the ARM
pseudocode, and utilizes it in FPConvert, which will be necessary for
half-float support.
LDNP and STNP indicate that a memory access is non-temporal/streaming
(i.e. unlikely to be repeated), allowing data caching to not be
performed. However, given this is only a hint, we can treat these two
instructions as regular LDP and STP instructions for the time being.
Provides a more "correct" move constructor/assignment operator, since
these relevant functions shouldn't throw exceptions.
Has the benefit of playing nicely with std::move_if_noexcept and other
noexcept library facilities.
The AddressRange structure isn't used anywhere within the codebase, so
this can be removed. Particularly because there's no real appeal/heavy
potential use of it in the future that isn't trivial to add back if
needed.
These optimizations can also apply to the 64-bit variants of the shift
opcodes; we just need to check if the instruction has an associated
pseudo-op before performing the 32-bit variant's specifics.
While we're at it, we can also relocate the code to its own function
like the rest of the cases to keep organization consistent.
We can fold division operations if:
1. The divisor is zero, then we can replace the result with zero (as this is how
ARM platforms expect it).
2. Both values are known, in which case we can just do the operation and
store the result
3. The divisor is 1, in which case just return the other operand.
It's common for an folding operation to apply to both the 32-bit and
64-bit variant of the same opcode, which leads to checking which kind of
result we need to store the value as. This moves it to its own function,
so that we don't need to duplicate it in various functions.
Multiplication operations can currently be folded if:
1. Both arguments are known constant values
2. Either operand is zero (in which case the result is also zero)
3. Either operand is one (in which case the result is the non-one
operand).
By far, one of the most common things to check for is whether or not a
value is zero, as it typically allows folding away unnecesary
operations (other close contenders that can help with eliding operations are 1 and -1).
So instead of requiring a check for an immediate and then actually
retrieving the integral value and checking it, we can wrap it within a
function to make it more convenient.
3e75ad98 Update version
4f043f8e Bump version
cc02cbc4 Fix formatting
73c0238e Update changelog
cb122a4d Fix format_to formatting to wmemory_buffer
dc69cc45 Clean tests
9d8021f0 Add checks for NVIDIA's CUDA compiler
9d2221b9 Improve error message when formatting unknown types
70a6a4bb prevent ""fmt/range.h"" from specializing fmt::basic_string_view (#865)
e4fc856c Disable android build due to gradle issues
3f4984fb Clean core-test and fix linkage errors on older gcc
d4366505 Workaround visit lookup issues in printf.h on gcc
894b6fac Changed to use scoped enum
59f555ad Workaround more visit lookup issues on gcc
a7e356cc Update README.rst
e758bfba Merge branch 'release' of github.com:fmtlib/fmt
66381e30 Minor cleanup
295a0d84 Update version
1fb1c4c9 Update docs
465a5935 Add table support to rst2md
d62f4c3b Formatting
a243490a Add more methods to benchmark results
9e12ca60 Update changelog
fbca830d Update changelog, readme and improve compat
6146248c Update changelog
bc26fbf1 Move experimental color API to fmt/color.h
97cc8893 Workaround a visit lookup issue in gcc 8 (#851)
7110b460 Optimize default formatting
c8a8464f Optimize buffer construction
8cbfb6e7 Get rid of conversion warning in gcc-4.8 (#854)
6ffc828a Phasing out null_terminating_iterator
aeb6add3 Skip strchr for the common case
5614289d Optimize and simplify format string parsing
10c7f893 Optimize format string processing on dumb compilers
59c268a5 Use strlen when possible since it's constexpr on gcc
918bb1ce Optimize argument capture
a3ba6b4f Disable the fmt(...) macro by default (#853)
86716894 Update docs and formatting
cc10b460 Make format_to faster on older gcc
981797f0 Get rid of implicit-fallthrough warn. in GCC 7 and 8
21177757 Micro-optimize parsing
be0e2684 Optimize processing of trailing '}'
fbc38b90 Pass heavy arguments by ref
8dc69b9d Workaround a bug in Intellisense
1489d3b7 Implement exponential notation
dd8c5ce4 Implement more FP formatting options
46484da7 Fix a warning
802ff886 Fix compilation of time.h when localtime_t is a macro (#843)
95a71899 Remove conversion compiler warnings (#844)
e483a01a Implement some formatting options in Grisu
f5108091 Revert "Implement some formatting options in Grisu"
2a952dd0 Implement some formatting options in Grisu
0de44a46 Implement exponent formatting
f0d0a1eb Implement Grisu2 digit generation
569ac91e Implement Grisu boundary computation
a11eb3a0 Workaround various icc bugs (#822)
62010520 Disable gnu-string-literal-operator-template warning
98751476 Make convert_to_int public (#818)
ba95e36a Clarify that '\0' cannot be used as fill (#832)
abde38b4 Add compilation support with Newlib nano for embedded targets
18400503 Fix C4127 warning in basic_writer<Range>::write_double
9de31211 Reformat and add a comment
8bbb0b48 Update README.rst
5c0101ab Use the correct function signature in the docs
fbe6410e Fix docs
8b9fb9fb Fix ambiguous instantiation with formatter in fmt/ostream.h (#830)
0f04ec68 Fix package upload (#828)
80907385 Update changelog
5d02041c Update changelog
4b868b89 Re-enable compile-time format-string checking
4061a0d3 Parameterize vformat to support custom char types
c68bab70 Remove broken fmt::internal::format_enum (#818)
0c63d15e Improve wording
ce19309d Workaround a bug in icc 15
c6843491 Move contiguous version of format_to to fmt/core.h
8db14efa util-test -> core-test and minor cleanup
ffe414ca Add compile-time format string checks to format_to (#783)
c178ab44 Remove FMT_USE_RVALUE_REFERENCES
5befe658 Remove fmt/folly.h and clean up core API
35538ca6 Merge more format overloads
4f164097 Merge format overloads using SFINAE
2a4e9488 Add UTF-8 types
d778bded Make line in tests fit within 80chars
7b4f170c Fix warning about using old-style cast
b1d10a28 Add support for dynamic arg sets
cf2719bd Add support for types explicitly convertible to wstring_view
50584f42 Test formatting of an object with templated conversion to string-like
73bed45b Add support for types explicitly convertible to fmt::string_view
6eaa5074 Fix global initialization issue (#807)
48dff9f3 Update docs
a9e26159 Minor cleanup
efd8ee8a Reduce warnings, support #809
8615ff2a Micro-optimize argument retrieval
916ed99d Micro-optimize argument retrieval
e7e9578e Optimize format string parsing
c99a2597 Mark new functions with FMT_API (#808)
e0f6a2f8 Add a formatter for folly::StringPiece
ae4a3945 Revert "Better support for newer CMake's"
a317448b Keep noexcept specifier when exceptions are disabled.
0eb01b83 Better support for newer CMake's
2a4cd6d0 Fix the returned value of `format_to_n` with user-defined types having operator<<.
9c32e73a Fixing return unreachable warning on NVCC
e5c93108 Added clear() to basic_buffer
60c662b3 Add an example of reusing formatters
f66ba650 Optimize format string parsing
f21268aa Revert "Optimize format string parsing" because of a bug in MSVC
07b690a6 Update README.rst
f9e9bf02 Optimize format string parsing
c2ce7e4f Update version
434eb916 Update README.rst
09d94162 Update changelog
e6362642 Fix pedantic conversion warning
f0110e81 Update changelog and CI
479ee2a8 Fix MSVC build, take 2
e928b672 Fix MSVC 2013 build
ec218a3a Fix redefinition warning for RESET_COLOR
c04fb91b Fix handling of user-defined types in format_to (#793)
323b92bf Force linking of inline functions into the library (#795)
c6d9730d Fix sign conversion warnings (#790)
2e95823e Move new color support to format.h and mark old as deprecated
ab2d88ca Make format_to work with basic_memory_buffer (#776)
3abd036c Fix compilation on gcc 4
c2f38054 Add vformat_to_n (#769)
ce500635 Renamed enum color to colors. Added enum colors conversion to rgb struct. Added colors_test.cpp.
0508bbc7 Add wchar_t overload of format_to_n (#764)
c2fbadb9 Fixed issue #779
47268ecd Fixed GCC version test
9ff3b6af Fix handling of compile-time strings when including ostream.h (#768)
e3707ef1 Document that file should be in wide-oriented mode for wide print
45fa4ee9 Merge branch 'master' of github.com:fmtlib/fmt
9c07b37f Using enum class now. Renamed from hex to color. Changed colr names to snake case.
5b5886a9 Fixed line length.
d2bfee13 Added quotes for strings in ranges and tuple likes.
aff6e45e Added support for rgb color output.
1b8a7f8f Fix postincrement in truncating and counting iterators
4bc26f0a Merge branch 'master' of github.com:fmtlib/fmt
fc6e0fe9 Fix FP formatting to a non-back_insert_iterator with sign & numeric alignment (#756)
cd5b5670 Make is_range and is_tuple_like public API, fix#751
6322b47e Minor cleanup
691a7a91 Add more compilers to CI and increase FMT_PEDANTIC warning levels (#736)
dd1a5ef7 Let requests close the file
d5c46259 Fix formatting of more than 15 named arguments (#754)
47d147b6 Simplify the nvcc warning fix
911a7511 Fix nvcc warnings (#752)
94b47628 Fix docs
252f11f8 Fix a bogus MSVC warning about unreachable code, take 2
81d56638 Fix more bogus MSVC warnings about unreachable code (#748)
68f0ac82 Fix a bogus MSVC warning about unreachable code
b60a5c5d Improve floating-point formatting
8dc2360b Fix a comment
4e4b8570 Implement simple version of Grisu
40275579 Fix tests on 64-bit MSVC
5c32aa41 Workaround a bug in MSVC
468c243c Add a function to get cached power of 10
2f257b72 Implement normalization and simplify power table
6a5bb6e2 Move Android.mk to support and update
e282d963 Bump version
e2cd521b Fix incorrect call to on_align in '{:}=' (#750)
fba352a9 Don't use UDL templates on Intel C++ compiler (#742)
6dcc526d Update release script
5386f1df Update version
ba6640b2 Fix formatting
507a50c3 Fix changelog
147807c9 Detect integer_sequence support on MSVC
8b246531 Update changelog
5ad54256 Fix a conflict between fmt::join and fmt/ostream.h (#744)
6ebc1a96 Merge locale.h into format-inl.h
6966db1d Update docs
2196025d Fix a warning
589f5f37 Update changelog
edd5f144 Fix compilation errors on gcc 4.4
936aba5f Fix compilation errors on gcc 4.4
3e3a2774 Update changelog
b76bb796 Improve naming consistency
fbd51534 Update changelog
69823bf8 Improve naming consistency
d940fa67 Disable unsafe implicit conversion to std::string (#729)
d2bf93fe Update changelog
550ef1d2 MSVC improvements and data truncation cleanup.
728e4f5a Fix docs
8c255771 Update docs and changelog
a68fd44e Add ranges.h to FMT_HEADERS in CMakeLists.txt (#738)
e3f7f3a2 Add support for ranges, containers and tuple-like types in fmt/ranges.h
984232db Remove duplicate ChangeLog entries
78677e3f Update ChangeLog and docs
ad23270e Document to_wstring
3c0f8c26 Update ChangeLog
98937893 Detect inline namespaces on gcc
dfb65469 Fix docs
3aa29115 Update ChangeLog.rst
d3f6c841 Update ChangeLog.rst
c1441ae4 Update ChangeLog.rst
dece85b3 Fix docs, take 2
6a1df3bd Fix docs
838400d2 Add inline namespace fmt::v5
b64b24eb Update ChangeLog.rst
fc908711 Update ChangeLog.rst
46c374a8 Fix compilation with new gcc and -std=c++11 (#734)
f0ae7257 Clarify the use of allocators
d72d0462 Update paths in fmt.pro
edbbf7ce Fix FreeBSD 12
a4e4f745 Fix a -Wundef when FMT_GCC_VERSION < 600
7d3de497 Implement double to fp conversion
a4c7d99f Add bit_cast
0adccaef Fix a -Wundef of _LIBCPP_VERSION
2570f1af Provide more overloads for the wide string flavour
ca31ca13 Fixed arg_formatter_base::write_pointer to not mutate the format specs.
6cd66610 remove trailing spaces.
fe19c266 Move format_string to fmt namespace for ADL
2768af23 Add cached powers of 10
dd296e1d Add a script to compute powers of 10
0efc8a18 Fix compiler warning about narrowing
df1ba52b Update example
221b08fd Merge branch 'master' of github.com:fmtlib/fmt
fa9066fe context_base::begin -> out
90ff31b3 Fix a -Wundef warning on clang
b1f68c43 Merge branch 'master' of github.com:fmtlib/fmt
cd90097c Implement handmade FP
822eccc3 Sync API with standards proposal
2ae41242 allow time formatting with wchar_t contexts
a1579b0f Update key
ded921f0 Fix documentation build, take 2
3284751f Fix documentation build
bb738c4c Remove section on Write API since it's being superceeded by compile-time Format API
d180c25c Update godbolt link
1ed842a3 Update godbolt link
e80aba1c Remove format_float stub
7b8cb313 Make context_base::args() public
48ae0506 fixes MSVC compiler warning bloat (Visual Studio 2017, latest updates)
096c4051 Simplify char_traits
7610c536 Remove unused macro
111fa581 Update README.rst
52fcef1e Update docs
7d28674d make_args -> make_format_args
9382b76f context_t -> format_context_t
fd0b07a7 (w)context -> (w)format_context
26aa34f3 basic_context -> basic_format_context
44cc0346 Relax string_view requirements
0829cab8 Remove from_checked
cb7bbc62 Improve checked iterator support
5079f924 Fix a narrowing warning
5859e58b Fix msvc warnings
1e747f60 Fix msvc warnings
9d4efd7a Iterator Wars VI: Return of the checked iterator
9764f558 Update docs
4ef97b9b Add a missing comma
23759b26 basic_arg -> basic_format_arg, arg_store -> format_arg_store
4975297e Simplify counting iterators
e8e006f4 Fix compile checks for mixing narrow and wide strings (#690)
c5ebecf7 Document format_to_n
3cf05263 Return output iterator to the end from format_to_n
174087bf Implement format_to_n
050f3f1f Remove parts of obsolete write API
e90b1da3 Fix linker errors using fmt as shared library in MSVC
8e10d404 Fix compile tests
7a41d61d Add make_printf_args
4fea018b Fix string_view detection
6957d28c Detect string_view on libc++ (#686)
0ea70def Update readme
9ce5e30c Update readme
8c29459e Fix handling of empty string_view (#689)
a24005d5 Fix a narrowing warning
3651b7fc Fix a narrowing warning
b64486da Add format.cc
3da71d51 Move source files to the src directory
7971ed3d Update readme
f61ca2ec Update readme
84e520b7 Update readme
e8aa0f33 Update docs
17258e9c Update docs
6d339e32 Improve comment
c3d05245 Fix a shadowing warning
b58c8dde Update docs
505b3ae6 Workaround GCC bug 67371 (#682)
70dffc63 Remove unnecessary check
df828f88 Don't define FMT_GCC_VERSION on clang
42f70c8b Avoid narrowing casts
10b939b0 Remove unneeded usage of anonymous struct on clang
3adfaae2 Remove extra semicolon in format_args constructor
40066785 Fix warnings under MSVC (#679)
9c5f54a7 Add format example for padded hex byte
7bab90e5 Remove extra comma
2e21e7d1 Fix util-test
acb469ae Fixed UTF8/16 converters to support empty string input
c37c4c43 Fix find-package-test
6d21fc43 add alias targets with fmt namespace
e02aacc6 Add CMake namespace (#511)
aee4512c Gradle (#649)
7db0e94b Fix handling of numeric alignment with no width (#675)
9facc119 Update docs
a1d18711 Merge branch 'master' of github.com:fmtlib/fmt
daf650c4 Disallow formatting of multibyte strings into a wide buffer (#606)
8fd7e30f Update README.rst
ca93be13 Use fmt(s) as an alias for FMT_STRING(s)
80e57c7a Update to new naming conventions
ae3cc844 Check format string at compile time in print
585512fc Remove unnecessary instantiations
7755cdc1 Make symbols readable
f867d082 Update docs
a103b9bc Workaround missed optimization in gcc (#668)
bb47109a Cleanup
f1ede638 Make inline_buffer_size public and update docs
995b63ad Update copyright
40232917 Update docs
86a9bc82 Cleanup
b7632e96 Make format_to return iterator and update docs
5281ea6a do_vformat_to -> vformat_to and update docs
d07ba498 Fix docs
418659ad Fix compilation errors on gcc 4.4
1d2adef2 Fix compilation errors on gcc 4.4
45518c3f Fix compilation errors on gcc 4.4
698d9097 Workaround a bug in gcc 5.1
81074c70 Fix more compilation errors on gcc 4.6
1b452538 Fix more compilation errors on gcc 4.6
6090e51b Fix compilation errors on gcc 4.6
0827ec5a Fix compilation errors on gcc 4.6
4d35f941 Always use fallback string_view to pass format string (#664)
34cf54c2 Update README.rst
0565d654 Fix gcc 7.2 issue
f5dc0ed3 Break long lines
ea06f021 test: comment out one FormatStringErrors constexpr test
5b491773 test: Initialize some local variables
f45f70af Use trailing return type instead of deduction
db86e8d5 Remove a couple of unused argument names
55f5c9f2 Use FMT_NULL instead of 0 is a few more places.
e92ba107 Fix Python str.format link to point to Python 3 docs
a7ae5666 Enable join on msvc
24d249b0 Fix formatting of objects convertible to string_view
e508e308 Don't define FMT_LOCALE on OpenBSD
0ee4273b Put is_enum check first not to instantiate convert_to_int unnecessarily
8ca3ab2c Revert problematic pragma
18ac9870 Fix formatting of objects convertible to std::string
ce4a65ff Add pointer support to basic_writer
91721caa Add detection of wostream operator<< (#650)
1efc15c1 Fix MSVC build
8ed264fc Rename type enum constants to prevent collision with poorly written C libs (#644)
4ba3f7db Update docs
7d2723d5 posix.cc: Fix compilation with -fno-exceptions
24d66c5d compilation fix & warnings
229887bd Make constexpr remove_prefix gcc version check tighter (#648)
f3f19e76 Update docs
e9fa42ac Fix docs and build issues on gcc-4.6
affb35cf Replace using with typedef for compatibility with gcc-4.6
9710c058 Update documentation building script
1a4e8927 Move output_range to format.h
522de7b5 Replace using with typedef for compatibility with gcc-4.6
0b508fd2 Fix c++0x detection
1849735f Fallback to c++11 if c++14 not available
3239c518 Get rid of generic lambdas
78166ccd Get rid of generic lambdas
d8ef8a9e Cleanup
82222218 Update README.rst
b0005324 Merge the std branch
a502decd Added a fmt.pro to support build using qmake (#641)
61065e1a Fix unreachable code warning when signbit returns bool
403ae0a2 Add debug postfix for libfmt (#636)
5096c0fe Fix string_view detection
5b3f9eab Update syntax.rst
e802cf14 Add note about errno to the documentation
c96d6465 CMakeLists: Use GNUInstallDirs to set install location
dbd84697 Update usage.rst
5013c157 Silence MSVC 2017 constant if expression warning
cdfcee27 Use allocator_traits if available
66b25ef0 Add examples
6cb68f94 Fix warnings
0b635c9d Fix handling of fixed enums in clang (#580)
66afd9b3 Fix compilation on gcc 6
67e070fe Make format work with C++17 std::string_view (#571)
867b3309 Remove ANDROID macro check per comment in #458
64599973 Enable stream exceptions (#581)
35f8f036 Use less version 2.6.1 and sudo to fix npm install issues on travis
92a250fd Suppress Clang's warning on zero as a null pointer
2f13d41e Add to_wstring
1e19ae83 Workaround a bug in MSVC
3810d7e4 Workaround a bug in MSVC
5c7474e1 Relax constexpr requirements
1f57243b Relax constexpr requirements
dc540361 Conditionally compile constexpr
5d8ba816 Fix a segfault in test on glibc 2.26 #551
a9f810c1 Update README.rst
2582f41e Fix ifdefs
1a7d0ba2 Adding OpenSpace to the list of projects
8921f613 Update build script
f62e225e Automatically update version in release script (#431)
94806747 remove 'FMT_CPPFORMAT' CMake option
bfce29ff Improve conversion
8cf30aa2 Fix segfault on complex pointer formatting (#642)
f164e4c7 Remove old bcc-related comments
c57029c1 Add Drake & Lyft Envoy to the list of projects
8fa9acb8 Workaround broken __builtin_clz in clang with MS codegen (#519)
3dae2582 Describe cmake use of header-only target
1c7b751d Fix handling of implicit conversion to integral types larger than int
08dff377 Allow compiling and using as DLL in windows #502
c753a2af Don't include the world with WIN32_LEAN_AND_MEAN (#503)
a5185ec8 add SOURCELINK_SUFFIX for compatibility with Sphinx 1.5
768061c8 Fix FormatBuf implementation (#491)
0c136381 Move back_insert_range to format.h
5060568f %.f should have zero precision, not default precision
a09f7488 Add Kodi (xbmc) to the list of projects using fmt
f9fa7c40 Add FMT_API and FMT_OVERRIDE where needed
a980d3b4 Add fmt::join to format ranges (#466)
87eab90e Fix missing intrinsic when included from C++/CLI (#457)
75005bbc Don't export the -std=c++11 flag from the fmt target
19f990a9 Use https to fetch dependencies from github
bca9de9e Return iterator from format_to
0555cea5 Added a fmt.pro to support build using qmake (#641)
a93270fd Replace a bunch of craft with type_traits, take 2
21429c86 Revert "Replace a bunch of craft with type_traits"
0473c48f Add std::basic_string allocator support (#441)
72d9fffd Fix test compilation for FreeBSD (#433)
e79588d6 Replace a bunch of craft with type_traits
3a6c7d0c Fix signbit detection (#423)
5e4c34b2 Add version macro FMT_VERSION (#411)
bd8a7e7e More iteratification
f78c3e41 Fix unreachable code warning when signbit returns bool
0a402056 Add CONTRIBUTING.rst
e35d41ff Add extern templates for format_float (#413)
d8c25a17 Use nullptr if available
e95e4659 Add syntax.rst to build
e5111950 argument index -> argument id
229ee34e Fix compiler warnings
7fe0f3da Update ChangeLog
38b603a4 Update README.rst
a1e7e4a7 Fix compilation with -fno-exceptions (#402, #405)
3f24a388 Thread-safe time formatting (#396)
f853d94a Remove unnecessary fmt/ prefix (#397)
9649919d Document use of format_arg for user-defined type #393
c8efe145 Add api.rst to build
da80005f Fix compilation on Cygwin (#388)
8ed16353 Fix a typo
1760c31b Workaround Doxygen mess
72606f23 Add missing types to counting_iterator
c1571003 Add debug postfix for libfmt (#636)
6822466a Handle nested braces in join (#638)
64b349ae More iterator support & fmt::count
e3b69efb Suppress msvc warnings in gmock
322736d3 Add support for arbitrary output iterators
10291194 Cleanup
c1d137ed Add support for nonconiguous iterators
f6fd38bb More iterator support
c2fecb9b Clean API
9a53a706 Add support for back_insert_iterator
91ee9c9a Return iterator from the format method
67928eae Don't inherit context from parse_context
217e7c76 Pass ranges by value
22994c62 Decouple arg_formatter_base from buffer
00f1450d Update tesmplate parameter names
3a2e89e1 Reduce dependency on buffer
c719d944 Fix experimental/string_view detection
cea3c207 Give a better error message for function pointers (#633)
232ceabb Workaround an internal compiler error in MSVC
c0954453 Replace buffer with range
c3d6c5fc Replace buffer with range
0f987731 add transition helper to format.h
d165d9c4 Decouple locale and buffer
36634140 Parameterize basic_writer on buffer type
6f2769d0 Revert "Added support for format string containing '\0' in _format udl (#619) (#620)"
5f1c73db Shorten a comment in locale.h
31934602 Update version
51a16f8c Update ChangeLog.rst
a0087460 Merge release branch
941663d0 Merge ostream.cc into ostream.h
955062da Merge printf.cc into printf.h
5705bf1c Added support for pre-c++17 experimental string_view (#607)
cabce31f Update syntax.rst
ccaae0c0 Refer to jeaiii project
e3715102 Add a integer formatter based on jeaiii
b3495f2e Update README.rst
61f296e3 Move FMT_HAS_BUILTIN to format.h
ce801c90 Remove dependency on <vector> and <array>
41fc2990 Merge branch 'std' of github.com:fmtlib/fmt into std
971fb584 Allow mixing named and automatic arguments
af0f21da add missing inline in header-only mode (#626)
7cea1638 numeric -> arithmetic
5328907f Get rid of <limits> dependency
faaafc7e Remove <utility> dependency and replace typedefs with using
94edb1a7 Add a lightweight header for the core API
3aaa25fa Added support for format string containing '\0' in _format udl (#619) (#620)
84bd2f19 Merge include/fmt/CMakeLists.txt into the main CMake file
7f351dec Decouple <locale> for better compile times
81bd9e8e args -> format_args
10e70a06 Improve handling of custom arguments
e0243000 arg_index -> arg_id
ac5f9520 Automatically add package to release
0e914372 Avoid conflict with the macro CHAR_WIDTH
f03a35a6 Check string specs at compile time
e9da5741 Check char specs at compile time
b25a0292 Check pointer type specs are compile time
c8a9d902 Check floating-point type specifiers
6570dc31 Disallow formatting of multibyte strings into a wide buffer (#606)
3851994a Fix yet another internal compiler error in MSVC
44e18651 Refactor parse context and fix warnings
e7e270f5 Test error on invalid type spec and remove unused alias
692b82d3 UdlArg -> udl_arg
c523dd58 Use error handler to report errors
5a32e64b More tests
093e2a47 Improve error handling
dc104cba Workaround internal compiler errors in MSVC
39411504 More tests
e3eb5ea0 Add parse_context::error_handler()
734e722d Fix warnings
62af25dc Workaround yet another MSVC internal error
594bd8fe More tests
f2b52bba More tests
dfdb1ade More tests
7967c2f8 Disable test that triggers an MSVC bug
18a0b94b Fix overflow check
686ff942 Fix compile-time parsing and add more tests
5b95b5d7 Test compile-time errors
246bdafc Add FMT_STRING macro for compile-time strings
e8055433 Remove FMT_USE_VARIADIC_TEMPLATES
dba1ccc4 Update readme
e613b3c7 Update readme
9fda7a36 Check integral type specs at compile time
92847a0d Add integral type handler
a03842b0 More compile-time checks
1c855a47 Integrate constexpr format specs parsing
780b44bf Add compile-time format string check
8ca6e76d Detect user-defined literal templates
a7e98616 Workaround another MSVC madness
db9ffa14 Make parse_format_string constexpr
e926ae78 Add parse_format_string
57e266ab Rename handlers
d29c7c3a Workaround a bug in MSVC
aadb38a5 Make specs_checker constexpr
dd0b72e1 Remove refactoring artefacts
e52b10e3 Merge branch 'vitaut-patch-1' of github.com:fmtlib/fmt into std
529d88ce Make dynamic_format_specs construction constexpr
d2f2a8b0 constexpr support of dynamic width and precision
6b3840b7 Make format_specs construction constexpr
a38bd9ca Fix formatting and naming
91014f01 Naming conventions
932ab2bf Report error from parse_nonnegative_int via handler
0ebdf41e Fix compile-test
170f5c67 Move headers to include/fmt
3d11eac7 Workaround another MSVC constexpr bug
c69e3086 Update README.rst
25aac0be Fix travis build on macOS
b83241ff Make format spec parsing constexpr
bd5188c8 Remove MinGW because it's not on appveyor image
62616b88 Workaround a bug in MSVC's constexpr handling
b8f85f67 Use Visual Studio 2017 image on appveyor
7174de0d Fix contexpr-ness of pointer_from
3785afc5 Pass errors to handler instead of throwing (#566)
1b5ccf6c Make parse_arg_id constexpr
17f93fe0 Make basic_string_view ctors constexpr
d5e918b6 Detect C++14 compiler support
be5b4552 Make null_terminating_iterator more iteratory
643fb066 Check for argument indexing switch
d45544d1 Fix width handling in dynamic formatting
8cbf5447 Add parse context
ec4f5175 Replace Range with ParseContext in parse()
83dd2ab9 Simplify dynamic_specs_handler
5a8ae0bb Fix a warning
39bc319b Update test results
534bff7d Fix handling of max packed arguments
0cda806d Fix compile tests
a3191a99 Get rid of FMT_MAKE_WSTR_VALUE macro
fced79b0 Get rid of old compat macros
be887d92 Replace internal::get with std::declval
53cf0735 Get rid of FMT_MAKE_VALUE macro
2972de4b Char -> char_type
9ee7c216 Type -> type
1a09194a Cleanup type handling
c18a4041 Remove conditional and to_iterator
1cade7ef Remove FMT_USE_RVALUE_REFERENCES
7413239f Remove unnecessary qualification
af00e4f9 Remove printf_arg_formatter from format.h and cleanup
44a26e5e CharPtr -> pointer_type and move to writer
0fbd8465 Replace fmt::internal::make_unsigned with std::make_unsigned
8a2bc0ab Add nullptr support
80505995 Allow delayed type checking
b0867f3f AlignSpec -> align_spec and fix a warning
f194a418 Replace fmt::is_same with std::is_same
47c84d79 Move part of write API (spec factories) to a separate header
20168147 Add ptr, a helper function for pointer formatting
77c892c8 Fix more warnings
be7d72ba Fix expansion-to-defined warning
d4c504ae Fix a warning
27ad6cee Use standard enable_if
64681739 Fix a warning
38806167 Remove FMT_HAS_GXX_CXX11
a7320bdc Fix a warning
016acebb Remove legacy code
07f8ffc4 Suppress shadowing warnings
466386d5 Suppress a warning in gmock
70ef82a8 Workaround a bug in MSVC
5e0562ab Separate parsing and formatting
1102d465 Make format spec parsing context-independent
45911770 Separate parsing and formatting in extension API
7bd776e7 Explain why null_terminating_iterator is used
873c8451 Remove system_header pragma
9f7957c0 Separate argument parsing and formatting
da439f28 Suppress warning about missing noreturn attribute (#549)
eefdb379 Fix an unused argument warning
2f4f49fd Switch from cstring_view to string_view
a8d6f309 Minor optimizations
d16582a0 Move printf-related code to printf.cc
361911dd Use preinstalled version of cmake on travis
9ea183aa Fix MSVC build
8f4b918c Check argument index
4193485b Remove test files
07123e8f Use Ubuntu Trusty on Travis for a new CMake
586d6363 Implement more efficient handling of large number of format arguments
12252152 CStringRef -> cstring_view
5aa8d6ea Return locale by value
32ec13f1 Switch to C++ locale
b4f4b7e2 Clean the buffer API (#477)
f423e468 Replace clear() with resize(0) and data_ -> store_
23b8c24d Add noexcept
7175bd8a Fix error on MinGW
7258d1b8 Fix tests
3610f34c Fix windows build
572491ad Document which header defines formatting functions
c333dca0 Follow standard naming conventions
6a2ff287 Follow standard naming conventions
eedfd07f internal::MemoryBuffer -> basic_memory_buffer
4ec88607 ArgFormatter -> arg_formatter
50e71673 StringRef -> string_view, LongLong -> long_long
e022c21d Fix windows build
87b691d8 Merge StringWriter into StringBuffer
c2f02169 Merge ArrayWriter into FixedBuffer
fefaf07b Pass buffer instead of writer to format_value
6e568f3a buffer -> basic_buffer
bb1c82ef Fix build
a13b96ed Simplify API
624c5868 Simplify API
7ae8bd70 basic_format_arg -> basic_arg, Buffer -> buffer
bf0f1075 Parameterize format_specs on character type
296e9cad FrmatSpec -> format_spec
b5fb8dd1 stream -> buffer
984a1029 Remove IntFormatSpec and StrFormatSpec
4863730e Remove pad
aaa0fc39 Improve compatibility with old compilers and fix test
aea5d3ab Improve compatibility with older gcc and update tests
84850277 Use named argument emulation instead of nested functions
ec15ef7b Replace operator<< with write function
b77c8190 FPUtil -> fputil
8428621d BasicWriter -> basic_writer
939aff29 Remove unnecessary template arg from basic_format_args
f69786a7 Remove Not
b2a0d891 Merge value and MakeValue
acd1811c Value -> value
42a31907 Parameterize Value on context
a4d6cb32 Clean up basic_format_arg
d705d516 Parameterize basic_format_arg on context (#442)
422236af Don't erase writer type
abb6996f MakeArg -> make_arg
ee1651ce Handle empty format_arg state
3bbc5799 Fix MinGW build
63fcfc57 Fix build on older gcc
d86e51e9 Don't inherit basic_format_arg from internal::Value
f0588869 Fix handling of unpacked args (#437)
11836218 Add support for exotic character types
763ca978 Parameterize Value on character type
6cba8fe9 Move stuff out of internal::Value
e1ee5bf0 Replace StringValue with StringRef
0854f8c3 Parameterize formatting argument on char type.
9cf6c8fd Get rid of fmt::internal::Arg
5f022ae0 Remove FMT_DISPATCH
41d4bcf0 Ingore Xcode files
28429701 Merge BasicArgFormatter and ArgFormatter
d4084ac5 Get rid of ArgVisitor
d58cc8a4 Merge BasicPrintfArgFormatter and PrintfArgFormatter
e2dfd39c Update arg visitors
751ff64b Update ArgConverter to the new visitor API
c9dc41ab Replace ArgVisitor::visit with a free visit function
caa60b9c Update comment
95a53e1f Refactor argument visitor API (#422)
6d241167 Improve visitor API
a1dd524b format_arg -> do_format_arg
55a1ac50 Fix test
85793a18 Simplify API
9998f66f Replace formatter with context
2bba4203 Pass writer directly to format_value (#400)
b656a1c1 Make value the second argument to format_value
edf98792 Pass writer to format_value
64ca334a CharType -> Char
be613204 Char -> char_type
f85d5f4d BasicFormatter -> basic_formatter
18dfa257 Pass correct formatters to make_format_args
dafbec75 Fix type safety when using custom formatters (#394)
506435bf Fix formatting
f2879940 Fix formatting
48fe9783 Add format_arg::operator bool
119a63ab internal::Arg -> format_arg
65a8c2c3 format_arg -> format_value
13b04044 Add format_args::size_type
8a77e792 Enable C++11 in tests.
1e8553d6 Enable C++11 in tests.
06bab3ed Workaround mingw bug https://sourceforge.net/p/mingw/bugs/1531/
6fd6ecc1 Enable C++11 for no-windows-h-test
c4212f9e format -> vformat
21c6700b Don't build std branch with -std=c++0=98
209a1d58 Get rid of macros
9a079732 Test types
ea28a637 Get rid of FMT_VARIADIC_CTOR
0d8aca8d Get rid of FMT_VARIADIC_VOID
4ece95a7 Make make_format_args public
0028ce57 Get rid of FMT_VARIADIC
ece7ae5f Make format_arg_store convertible to format_args
621447fe Make initialization C++11-compatible
a0190e4b Add a missing include
b903f5c1 format -> vformat
43c0095a Refactor type mapping
4873685c ArgArray -> format_arg_store
fc73e106 ArgList -> format_args
92605eb4 Remove FMT_USE_VARIADIC_TEMPLATES
9bb213e9 FormatError -> format_error
REVERT: 135ab5cf Update version
REVERT: 93d95f17 Fix markup
REVERT: 4f15c72f Fix markup
REVERT: e9b19414 Automatically add package to release
REVERT: c3d1f604 Fix markup
REVERT: c96062bf Update changelog and version number
git-subtree-dir: externals/fmt
git-subtree-split: 3e75ad9822980e41bc591938f26548f24eb88907
This'll make it slightly nicer to do basic constant folding for 32-bit
and 64-bit variants of the same IR opcode type. By that, I mean it's
possible to inspect immediate values without a bunch of conditional
checks beforehand to verify that it's possible to call GetU32() or
GetU64, etc.
This is useful when we wish to know if a contained value is something
like 0xFFFFFFFF, as this helps perform constant folding. For example the
operation: x & 0xFFFFFFFF can be folded to just x in the 32-bit case.
It's possible to fold cases of exclusive OR operations if they can be
known to be an identity operation, or if both operands happen to be known
immediates, in which case we can just store the result of the
exclusive-OR directly.
An arithmetic shift is by definition a signed shift, and a logical shift is by definition an unsigned shift.
- Rename VectorLogicalVShiftS* -> VectorArithmeticVShift*
- Rename VectorLogicalVShiftU* -> VectorLogicalVShift*
Fixes decoding for 64-bit instructions
Does not help/apply to any currently supported ARM versions (since
all are 32-bit length or below), it's for future-proofing should
such an arch be supported.
Makes the comparison interface consistent by providing all of the
relevant members. This also modifies the comparison operators to take
the Imm instance by value, as it's really only a u32 under the covers,
and it's cheaper to shuffle around a u32 than a 64-bit pointer address.
Moves the declaration of datasize to the scope that it's used within.
This also takes the opportunity to apply const where applicable, and
make early-exits all vertically consistent with one another.
* Return both the upper and lower parts of the multiply if required
* SSE2 does not support the pmuldq instruction, do sign correction to an unsigned result instead
* Improve port utilisation where possible (punpck instructions were a bottleneck)
This was disabled for debugging, but never re-enabled. Just to be sure,
testing was done downstream in yuzu to make sure this didn't happen to
break anything (which seems to be the case).
We can optimize these cases further for with the use of a fair bit of
shuffling via pshufb and the use of masks, but given the uncommon use of
this instruction, I wouldn't consider it to be beneficial in terms of
amount of code to be worth it over a simple manageable naive solution
like this.
If we ever do hit a case where vectorized CLZ happens to be a
bottleneck, then we can revisit this. At least with AVX-512CD, this can
be done with a single instruction for the 32-bit word case.
We can simply modify the general function made for SMULL{2} and
UMULL{2}'s by-element variants to also handle the other multiply-based
by-element variants.
enum classes are still considered complete types when forward declared
(as the compiler knows the exact size of the type from the declaration
alone). The only difference in this case being that the members of the
enum class aren't visible. Given we don't use the members within this
header in any way, we can simply forward declare them here and remove
the inclusions.
In the event position == 0, we can just treat it as a simple movq,
clearing the upper half of the XMM register. This also makes that case
use only one register.
* The MSB for each byte in cpsr_ge were not being appropriately set.
* We also expand test coverage to test this case.
* We fix the disassembly of the MSR (imm) and MSR (reg) instructions as well.
While we're at it we can join the code for SADDLP and UADDLP with these
instructions, since the only difference is we do an accumulate at the
end of the operation.
Currently we implement these in terms of the scalar variants. Falling
back to the interpreter is slow enough to make it more effective than
doing that.
This avoids a few redundant atomic increments and decrements,
considering the UserConfig instance contains a std::array of
std::shared_ptr<Coprocessor> instances.
x64 rounds before flushing to zero
AArch64 rounds after flushing to zero
This difference of behaviour is noticable if something would round to a smallest normalized number
Bugs:
* DenormalsAreZero flushed to positive zero instead of preserving sign.
* FMAXNM/FMINNM (scalar) should perform DAZ *before* special zero handling.
* FMAX/FMIN/FMAXNM/FMINNM (vector) did not DAZ.
For the time being, we can just do a naive implementation which avoids
falling back to the interpreter a bit. Horizontal operations aren't
necessarily x86 SIMD's forte anyways.
Gets rid of a bit of duplication regarding the early-out cases and makes
all helpers functions consistent (previously some had a return type of
bool, while others had a return type of void).
Discovered by @Subv.
Fixes incomplete fix begun in 5a91c94dca47c9702dee20fbd5ae1f4c07eef9df.
That fix fails to take into account that LinkBlock doesn't update ticks until there
are no remaining ticks to be executed.
Test added to confirm fix.
We can just use the same register the data is in as the result register,
eliminating the need to use a completely separate register to store the
result.
For example, say the converted signed type is s64, shifting left by 63
bits would be undefined behavior.
However, given an ASL is essentially the same behavior as an LSL
we can just use an unsigned type instead of converting to a signed type.
Negation of values such as -9223372036854775808 can't be represented in
signed equivalents (such as long long), leading to signed overflow.
Therefore, we can just invert bits and add 1 to perform this behavior
with unsigned arithmetic.
This code was bugged in that we were comparing if the rounding mode was
not equal to rounding towards zero. Fortunately, however, nothing uses
this function anymore, and there's already the more general
FPSCR_RMode() available, so this can be removed entirely.
Commit 755adef62e504a8d616de9dda8937d2428a9471b introduced a helper
alias for std::array, eliminating the need to manually type out sizes
for them, however I forgot to add the include for <array>
Avoids needing to remember to specify the proper size of the arrays, all
that's needed is to specify the type of the array and the size will
automatically be deduced from it. This helps prevent potential oversized
or undersized arrays from being specified.
We already have RotateRight() in our common code, so we can remove this
function and replace it with it. We can also implement ArmExpandImm_C()
in terms of ArmExpandImm().
Generally, the only difference between the signed variants and the
unsigned variants is whether or not we use a sign-extension or
zero-extension, so we can simply use common functions to implement both
cases without totally duplicating code twice here.
In the non-AVX512 path, the following code is present:
code.mov(from.cvt32(), from.cvt32());
since this potentially modifies 'from', we should be using
UseScratchGpr() instead.
Now that we have the helper function set up for the signed variants, we
can also modify it to be used with the unigned ones by performing a zero
extension instead of a sign extension.
Fixes a typo for the InvalidateByVAToPoC enum entry. Given yuzu is the
only known user of 64-bit mode and it doesn't use this value, we can get
away with changing this.
Now that the utility function is all set up from implementing SRSRA, the
unsigned variants can now be trivially implemented by modifying the
utility function to perform a logical shift right instead of an
arithmetical shift right for the unsigned case.
Uses the instruction that does what it says in its name if available. Allows avoiding the use
of a scratch register in EmitVectorBroadcast8() and EmitVectorBroadcastLower8()'s SSSE3 path.
Now that the vector rotation helpers are in, replace the explicit
shifting with the relevant helper function that does the same thing.
Simply tidies up code; no behavioral changes are made.
Gets rid of partial duplication of the same code for instructions that only have a small behavior difference to them.
e.g. The only difference between SSHR and SSRA is that SSRA adds an accumulator before storing the result.
EmitZeroExtendByteToLong() can be implemented in terms of EmitZeroExtendByteToWord() and
EmitZeroExtendHalfToLong() can be implemented in terms of EmitZeroExtendHalfToWord().
Makes behavior identical to the literal variant of PRFM. Given this is simply a hint instruction,
this is valid behavior. The upside is that we don't fall back to Unicorn unnecessarily whenever
the instruction is encountered.
Without this sanitizing it's possible to perform a shift with a shift
amount that's the same size as the type being shifted. This actually
occurs when decoding ORR variants.
We could get fancier here and make this branchless, but we don't
really use RotateRight in any performance intensive areas.
This uses the ARM terminology for sizes (Halfword -> 2 bytes, Word -> 4 bytes)
as opposed to the x86 terminology of (Word -> 2 bytes, Double word -> 4 bytes)
