Empty objects that inherit from Sized would try to access internal
ByteBuffer when Sized::size was called. So we add a single byte in
the empty buffer, so when size() is called it would return 0
The real position of a string is calculated by using the indirect() method,
which should be based on parentWidth and not byteWidth, as it was implemented.
We are also fixing the flag BUILDER_FLAG_SHARE_STRINGS on FlexBuffersBuilder
that was set as '1', same value as BUILDER_FLAG_SHARE_KEYS.
* [FlexBuffers][Java] Implementation of FlexBuffers API
This is the initial attemp to implement FlexBuffer on Java.
There is some limitations as compared to the C++ implementation:
1 - No mutations implemented yet
2 - Does not parse from json
Also, this initial implementation is not focused and performance, but
get the basics write. So there is many opportunities for optimization, for instance,
remove all enums, return CharSequence instead of Strings and object pooling.
* [FlexBuffers][Java] Optimizations and simplification of the Builder API.
This change removes BitWidth enum in favor of static ints. Also
make all "reads" APIs closer to C++ implementation (try to cast or convert
as much as possible, assuming user knows what he is doing). Finally,
we remove the helper classes for building vectors and maps.
There is no official benchmarks, but the unit tests are running in less
than 50% for previous runs, which mean those optimizations are worth it.
* [FlexBuffers][Java] Fix Reference::asString behavior
There was a incorrect assumption that strings would be null-terminated, which
could lead to truncated strings. S now it relies size instead of null-termination.
Other minor improvements
* Add FLATBUFFERS_COMPATIBILITY string
- Add a new __reset method NET/JAVA which hides internal state
* Resolve PR notes
* Use operator new() to __init of Struct and Table
* Restrict visibility of C# Table/Struct to internal level
This is done on purpose, to avoid API version mismatches that
can cause bad decoding results, see:
https://github.com/google/flatbuffers/issues/5368
Change-Id: I2c857438377e080caad0e2d8bcc758c9b19bd6ec
vtable and vtable size depends only on `Table#bb_pos` but calculated in
`Table#_offset` method on each field lookup.
Doing this with every call of `Table#__offset` is redundant.
These values can be read once with change of `Table#bb_pos` and reused
for any field lookup.
Introduce a HeapByteBufferFactory singleton instance in order to reduce allocations.
Clarify the usage of LITTLE_ENDIAN ByteBuffers in ByteBufferFactory.
* initial changes to support size prefixed buffers in Java
* add slice equivalent to CSharp ByteBuffer
* resolve TODO for slicing in CSharp code generation
* add newly generated Java and CSharp test sources
* fix typo in comment
* add FinishSizePrefixed methods to CSharp FlatBufferBuilder as well
* add option to allow writing the prefix as well
* generate size-prefixed monster binary as well
* extend JavaTest to test the size prefixed binary as well
* use constants for size prefix length
* fuse common code for getRootAs and getSizePrefixedRootAs
* pulled file identifier out of if
* add FinishSizePrefixed, GetSizePrefixedRootAs support for Python
* Revert "extend JavaTest to test the size prefixed binary as well"
This reverts commit 68be4420dd.
* Revert "generate size-prefixed monster binary as well"
This reverts commit 2939516fdf.
* fix ByteBuffer.cs Slice() method; add proper CSharp and Java tests
* fix unused parameter
* increment version number
* pulled out generated methods into separate utility class
* pulled out generated methods into separate utility class for Python
* fix indentation
* remove unnecessary comment
* fix newline and copyright
* add ByteBufferUtil to csproj compilation
* hide ByteBuffer's internal data; track offset into parent's array
* test unsafe versions as well; compile and run in debug mode
* clarify help text for size prefix
* move ByteBuffer slicing behavior to subclass
* fix protection levels
* add size prefix support for text generation
* add ByteBufferSlice to csproj compilation
* revert size prefix handling for nested buffers
* use duplicate instead of slice for removing size prefix
* remove slice subclass and use duplicate for removing size prefix
* remove slice specific tests
* remove superfluous command line option
As Error does not inherit from Exception, a generic `catch(Exception ex)` will not catch this error.
A simple change from `Error` to `Exception` is not sufficient as an `Exception` is checked by the compiler, so in order to keep this issue unchecked, I am proposing raising a `RuntimeException` which is not checked, but is still a subclass of `Exception`.
The only possible breaking change would be if a consumer was explicitly catching `Error` already for this library.
https://github.com/google/flatbuffers/issues/4629
This is something the format supports, but none of the builders
were doing. Can save 10-20% on FlatBuffer binary size!
Also fixed the Go tests.
Change-Id: I616c56ce9bbcfcaee23aa24f0532fcb60b6a8c75
Tested: on Linux.
The ByteBufferFactory interface gives the user an option to specify
the method in which the internal ByteBuffer is allocated. This provides
flexibility in the type of ByteBuffer that can be used.
The sizedInputStream method is an alternative to sizedByteArray that
does not make a copy of the data in memory.
This avoids the following allocations over repeated use:
- A ``CharsetEncoder``
- A byte array to contain the contents of the string
- A wrapper, via ``ByteBuffer#wrap``, for the previously mentioned byte
array
This also removes a copy of the temporary byte array. This extra copy
is needed because ``String#getBytes`` returns a byte array and its
length must match the exact size of the contents.
This implementation requires that the builder retain a ``ByteBuffer``
and ``CharEncoder``. This is considered slower only for users that
will only allocate a single string in their buffers or none at all.
The improvement is especially potent if the builder is constantly reused
by its caller. This also eliminates the penalty for users that do not
use strings at all since the cost of allocating these fields is
now amortized.
The only allocation left in this code is a ``CharBuffer`` wrapper. It's
possible to eliminate this allocation using an additional copy but need
further profiling to see which is the bigger problem.
Removes the following allocations:
- ``CharsetDecoder`` is reused between calls
- ``CharBuffer#wrap`` removed in favor of heap
based char buffer that is reused
- Temporary ``char[]``, an intermediate copy inside ``StringCoding``
- Another ``char[]``, this is needed because ``StringCoding`` uses
a ``CharBuffer`` internally but returns a ``char[]``. Extra
characters need to be trimmed so this means yet another allocation
- Yet another ``char[]`` directly from ``__string`` for non-heap
based buffers
Removes the following copies
- No copy is performed to trim the allocation since a ``CharBuffer``
is used directly
- For non-heap based byte buffers, removes the copy that was
previously done in the __string function
This does need to get the TLS entry which implies at least some
contention on the thread object table and a fence.
Adding an API reference for the supported languages.
General docs cleanup, including a new `tutorial` section that
supports all of the supported languages.
Added samples for each supported language to mirror the new
tutorial page.
Cleaned up all the links by making them `@ref` style links,
instead of referencing the names of the generated `.html` files.
Removed all generated files that were unnecessarily committed.
Also fixed the C# tests (two were failing due to a missing file).
Bug: b/25801305
Tested: Tested all samples on Ubuntu, Mac, and Android. Docs were
generated using doxygen and viewed on Chrome.
Change-Id: I2acaba6e332a15ae2deff5f26a4a25da7bd2c954
Also checks for nesting of objects in vector construction.
This avoids common errors in FlatBuffer construction.
Change-Id: I5507c5d767684e20e94883a92448f05acefba4d6
Tested: on Linux.
