https://github.com/tensorflow/tflite-micro makes use of flatbuffers with
a variety of DSP toolchains.
Without the change from this PR, we can get a double-promotion warning
with some of these DSP toolchains:
```
flatbuffers/include/flatbuffers/util.h:104:11: error: implicit conversion increases floating-point precision: 'std::numeric_limits<float>::_Ty' (aka 'float') to 'double' [-Werror,-Wdouble-promotion]
T eps = std::numeric_limits<float>::epsilon();
~~~ ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
```
iomanip isn't available for our DSP. Luckily, we compile with FLATBUFFERS_PREFER_PRINTF, so moved the #include directive there.
ctype.h has to be included explicilty for tolower() and toupper().
* Add the file a symbol is declared in to Reflection
If we move a code-generator to depend on Reflection,
it may need to know which file something was declared in
to properly name generated files.
* Doc comments in reflection, and more precise tests
* Add --project-root flag to flatc, normalize declaraion_file to this root
* fix --project-root stuff
* posixpath
* fix scripts
* format
* rename --project-root to --bfbs-filenames
Also, make it optional, rather than defaulting to `./`, if its not
specified, then don't serialize the filenames.
* bfbs generation
* fix some tests
* uncomment a thing
* add to project root directory conditionally
* fix
* git clang format
* Added help description and removed != nullptr
* "
* Remove accidental change to docs
* Remove accidental change to docs
* Pool strings
Co-authored-by: Casper Neo <cneo@google.com>
* [idl_parser] Validate `force_align` on all possible paths
- added validation of `force_align` for `Vector`.
- added assertion to `flatbuffers::PreAlign` method
These changes should resolve following oss-fuzz issues:
- 6275816289861632
- 5713529908887552
- 4761242994606080
* size_t problem on Mac
* Fix review notes
* TS/ES6 modules spike iteration 1
* Initial modularized dasherized output
* Remove obsoleted parts and namespace wrapping
* Use _flatbuffers_ prefix
* First part of imports logic
* Second part of imports logic
* Fix TS/JS code removal mixup
* Alias imported symbols if same name from different namespaces and some fixes
* Use star import for bare imports
* Fix messed up string concat
* var to const and remove not needed semi
* Remove some cases of ns prefixing
* Add missing space
* Cleanups
* Completed initial import tracking logic
* Compilable output
* Adjust TypeScriptTest and dependents to work
* Use local flatbuffers package for tests
* Refactor away use of any
* Remove obsolete imported_fileset and reexport_map
* Still need any and fix JavaScriptTest.sh
* Fix test runs out of the box
* Temp add generated files
* TypeScriptTest replaces JavaScriptTest and cleanups
* Also remove reference to JavaScriptTest in TestAll.sh
* Remove old generated ts/js files
* Remove use of --js in generate_code scripts
* idl_gen_js_ts to idl_gen_ts and removal of js gen
* Remove obsoleted options
* Fix obsolete ts test detection
* Tweak ts compilation be as strict as possible
* Remove jsdoc type annotation generation
* Generated test ts files
* Fix search and replace messup
* Regenerated ts test output
* Use CharToLower
* Use normal for loop
* Rework namespacedir
* Revert "Rework namespacedir"
This reverts commit 6f4eb0104ceeb86011bb076ebca901138c48e068.
* Revert "Use normal for loop"
This reverts commit 676b2135bfaa1853dfbb06c92b5c16a0d81bb13a.
* Revert "Use CharToLower"
This reverts commit 2d08648d0d72d0af201fad80d54cdc76412b35e9.
* Again do rework but correct
* Avoid runtime cast
* Fix test runs
* Also add npm install to get tsc
* Bump node test versions
* for range to std for loop
* Clang format
* Missed one clang format
* Move accessor to later
* Attempt to make windows version of TypeScriptTest
* Want to see the output
* Try to get newer node at appveyor
* Style changes
This commit adds replacement of `::tolower` and `::toupper`.
Added CharToLower and CharToUpper routines reduce the number of cast operators
that required for correct usage of standard C/C++ `::tolower/toupper` routines.
* Fixed refractoring issue in reflection/generate_code.sh. Also, mv deletes the original file, so I don't need to clean it up manually in that case.
* Added --filename-suffix and --filename-ext to flatc
* Fixed typo and added example generation of suffix and extension for C++
* Removed extra ;
* Removed clang-format block from a region that didn't need it. Fixed an auto format of another clang-format block
* Added docs, fixed pointer alignment, removed suffix test file
* Added the code to embed the binary schema to the source.
This is pulled forward from a old PR #5162 that will be closed.
* Update idl_gen_cpp.cpp
Added a small comment to trigger a new build. The build was failing in a strange location and doesn't look like it has anything to do with the code.
* Moved the EscapeAndWrapBuffer to util.cpp and did some formating.
* One more camelCases removed and renamed some variables.
* wrapped_line_xxx should have been passed as a const reference in the first place.
* Moved the bfbs embed sample to it's own file.
* Missed moving the namespace back.
* Moved the embedded bfbs to test.cpp instead of using a sample.
* Missed adding the generation of embedded bfbs to the build.
* See if this makes the build happier.
* Fixed a in-compatable cpp output of the generated header.
* Did some changes to reflect the code review comments.
1. Update the EscapeAndWrapBuffer to BufferToHexText and fixed a variable name.
2. Moved the include of the embedded binary schema to all the other includes.
* Moved some code to inline the instead of using a local variable.
* Moved the BufferToHexText back to be a inline function in util.h
* Automatic refractor of C++ headers to Google C++ style guide
* Automatic refractor of C++ source to Google C++ style guide
* Automatic refractor of C++ tests to Google C++ style guide
* Fixed clang-format issues by running clang-format twice to correct itself. Kotlin was missing clang-format on after turning it off, so it was changed,
On Solaris Sparc, calling NumToString() with a char called the primary
version, not the signed char or unsigned char specializations, which
caused integer to string conversions to be missed.
* Fix high certainty warnings from PVS-studio
- Introduced FLATBUFFERS_ATTRIBUTE macro to use [[attribute]] if modern C++ compiler used
* Update the note about __cplusplus usage in the MSVC
* Make the Parser independent from the global C-locale
* Set a specific test locale using the environment variable FLATBUFFERS_TEST_LOCALE
* Remove redundant static qualifiers
* Refactoring of numbers parser
More accurate parse of float and double.
Hexadecimal floats.
Check "out-of-range" of uint64 fields.
Check correctness of default values and metadata.
* Remove locale-independent code strtod/strtof from PR #4948.
* small optimization
* Add is_(ascii) functions
* is_ascii cleanup
* Fix format conversation
* Refine number parser
* Make code compatible with Android build
* Remove unnecessary suppression of warning C4127
* Add define/ifdef blocks for FLATBUFFERS_PREFER_PRINTF to avoid using std::*streams for idl_parser
* Use string::size() as limit in snprintf
* Refactored FLATBUFFERS_PREFER_PRINTF guarded feature into NumToStringImplWrapper around sprintf
* Remove '.0' where not needed from IntToDigitCount
* Remove leading dot from name in GetFullyQualifiedName when FLATBUFFERS_PREFER_PRINTF is enabled
* Return string directly from conversion functions where possible when FLATBUFFERS_PREFER_PRINTF is enabled
* Use string instead of stringstream for GetFullyQualifiedName
* Revert removing leading dot from GetFullyQualifiedName, it does need to be there for parity with the stringstream implementation
* Dot is single char in Namespace::GetFullyQualifiedName
* Remove trailing (duplicate) null-byte from NumToStringImplWrapper when using FLATBUFFERS_PREFER_PRINTF.
* Update preprocessor indenting (and use clang-format off/on) for FLATBUFFERS_PREFER_PRINTF
* Reduce whitespace, unneeded braces in FLATBUFFERS_PREFER_PRINTF string width functions
* Remove unneeded use of iostream from idl_parser.cpp, std::string is used instead
* Tell snprintf to use the trailing null byte expected at the end of std::string buffer
* Added support for the non-escaped print of utf-8 string.
* EscapeString: the first invalid symbol resets print_natural_utf8 flag to false.
* Move the test to ParseAndGenerateTextTest. Fixes.
* Removed dependence between `natural_utf8` and `allow_non_utf8` flags.
In file included from include/flatbuffers/flexbuffers.h:24,
from src/idl_gen_text.cpp:20:
include/flatbuffers/util.h: In function 'int flatbuffers::FromUTF8(const char**)':
include/flatbuffers/util.h:324:45: error: type qualifiers ignored on cast result type [-Werror=ignored-qualifiers]
if ((static_cast<const unsigned char>(**in) << len) & 0x80) return -1; // Bit after leading 1's must be 0.
^
cc1plus: all warnings being treated as errors
make[2]: *** [CMakeFiles/flatbuffers_shared.dir/build.make:92: CMakeFiles/flatbuffers_shared.dir/src/idl_gen_text.cpp.o] Error 1
Const does not make sense here, and compiler actually throws warning
(error with -Werror) when you would try to compile it.
In file included from include/flatbuffers/flexbuffers.h:24,
from include/flatbuffers/idl.h:26,
from include/flatbuffers/code_generators.h:22,
from src/code_generators.cpp:17:
include/flatbuffers/util.h: In function ‘int flatbuffers::FromUTF8(const char**)’:
include/flatbuffers/util.h:325:44: error: type qualifiers ignored on cast result type [-Werror=ignored-qualifiers]
if ((static_cast<const unsigned char>(tmp) << len) & 0x80) return -1; // Bit after leading 1's must be 0.
^
cc1plus: all warnings being treated as errors
This warning caught by gcc8:
$ g++ --version
g++ (GCC) 8.0.1 20180222 (Red Hat 8.0.1-0.16)
* Added internal - limited - implementation of flatbuffers::unique_ptr
for STLs that don't ship with std::unique_ptr. In C++11 and beyond
this is just an alias for std::unique_ptr.
* Aliased used type traits structs is_scalar is_floating_point is_unsigned
into flatbuffers namespace so they can be replaced in C++98 implementations.
Right now these point at stlport's TR1 implementations.
* Wrapped vector::data() in vector_data().
* Wrapped vector::emplace_back() in vector_emplace_back().
* Wrapper string::back() in string_back().
* Added variants of FlatBufferBuilder::CreateVector() and
FlatBufferBuilder::CreateVectorOfStructs() that allow the use of plain
function pointers.
Generated code has also been modified to use plain functions to build objects
rather than std::function() so all generated code will work in C++98
applications.
* Added flexbuffers::Builder::Vector(), flexbuffers::Builder::TypedVector()
and flexbuffers::Builder::Map() methods that allow the use of plain function
pointers.
* Changed Parser to internally use plain function pointers when parsing table
and vector delimiters.
* Added specializations of NumToString() for 64-bit types that aren't supported
by stringstream in stlport.
* Overloaded numeric_limits for 64-bit types not supported by stlport.
* Replaced build_apk.sh (which was broken by deprecation of the
"android" tool in the Android SDK) with build.gradle and the
appropriate gradle wrapper to build an APK.
* Switched Android build to build against all STL variants.
* Updated travis configuration to build Android test and sample.
Tested:
* Verified all tests continue to work on Linux, OSX and Android.
* Verified Travis build is green.
Change-Id: I9e634363793f85b9f141d21454b10686020a2065
FlexBuffers are actually completely unrelated to FlatBuffers and can be used separately.
However, since they utilize several utility functions from flatbuffer.h, they require this header file, which creates unreasonable dependency.
By moving those utility functions to separate base header, both libraries can use same code but still be independent
There were several possible bugs involving paths not being
recognized as being the same on Windows. Rather than trying
to ensure all code deals with / and \ correctly, paths now
get transformed to / on input, fixing all current and
future such bugs.
Tested: on OS X.
* Fix Visual Studio 2017 new warning (C4244: 'argument': conversion from 'int' to 'const char', possible loss of data)
* Fix Visual Studio 2017 pedantic warnings
* Fix Visual Studio 2017 pedantic warnings
It passed all scalar ints thru a int64_t, which would truncate
uint64_t values with the upper bit set.
Change-Id: I38fb8c68c911ae44d9863f8e35c2429ca0ab51e5
Tested: on Linux.
