Refactor serializer #137
Refactor serializer #137
Conversation
|
An automated preview of the documentation is available at https://137.http-proto.prtest.cppalliance.org/index.html |
|
GCOVR code coverage report https://137.http-proto.prtest.cppalliance.org/gcovr/index.html |
|
|
||
| //--------------------------------------------------------- | ||
|
|
||
| class serializer::const_buf_gen_base |
There was a problem hiding this comment.
this declaration could be in src/ ?
There was a problem hiding this comment.
We need to instantiate const_buf_gen<ConstBufferSequence> for user-provided buffers.
ashtum
force-pushed
the
refactor-serializer
branch
from
5620864 to
a5ee891
Compare
|
An automated preview of the documentation is available at https://137.http-proto.prtest.cppalliance.org/index.html |
|
GCOVR code coverage report https://137.http-proto.prtest.cppalliance.org/gcovr/index.html |
ashtum
force-pushed
the
refactor-serializer
branch
from
a5ee891 to
4183089
Compare
|
An automated preview of the documentation is available at https://137.http-proto.prtest.cppalliance.org/index.html |
|
GCOVR code coverage report https://137.http-proto.prtest.cppalliance.org/gcovr/index.html |
ashtum
force-pushed
the
refactor-serializer
branch
from
4183089 to
1318efe
Compare
|
An automated preview of the documentation is available at https://137.http-proto.prtest.cppalliance.org/index.html |
Improvements:
* User-provided buffers are now type-erased and used directly, rather
than being copied into a flat buffer. This enables efficient
handling of buffer sequences with many elements.
* Chunk encoding now occurs within the main buffers, allowing
serialization of multiple chunks. This also improves buffer
consumption efficiency by reducing the total number of buffers.
* The number of internal states maintained by the serializer has been
reduced. This lowers memory overhead and simplifies reasoning about
the implementation.
ashtum
force-pushed
the
refactor-serializer
branch
from
1318efe to
c44cb03
Compare
|
An automated preview of the documentation is available at https://137.http-proto.prtest.cppalliance.org/index.html |
|
GCOVR code coverage report https://137.http-proto.prtest.cppalliance.org/gcovr/index.html |
3 participants
Improvements:
User-provided buffers are now type-erased and used directly, rather than being copied into a flat buffer. This enables efficient handling of buffer sequences with many elements.
Chunk encoding now occurs within the main buffers, allowing serialization of multiple chunks. This also improves buffer consumption efficiency by reducing the total number of buffers.
The number of internal states maintained by the serializer has been reduced. This lowers memory overhead and simplifies reasoning about the implementation.