Ugorji Nwoke
|
d75b2dcb6b
codec: regenerate fast-path.generated.go
|
7 years ago |
Ugorji Nwoke
|
6bfa3bf159
codec: do not pool codecFn helper - make it part of BasicHandle instead
|
7 years ago |
Ugorji Nwoke
|
856da096db
codec: implicitly call Close() by default
|
7 years ago |
Ugorji Nwoke
|
f6bb70d25a
codec: shared pool for large values used in Encode/Decode
|
7 years ago |
Ugorji Nwoke
|
e531a5003d
codec: refactor to afford bounds check elimination
|
7 years ago |
Ugorji Nwoke
|
3bfe5b406a
codec: more support for mid-stack inlining support - gives best performance with -l=4
|
7 years ago |
Ugorji Nwoke
|
f8fe4a359c
codec: set defEncByteBufSize=1024, used as length of input []byte or io.Writer buffer
|
7 years ago |
Ugorji Nwoke
|
0165389f8c
codec: bufioEncWriter.reset should not create a new []byte everytime, but reuse old one
|
7 years ago |
Ugorji Nwoke
|
16061014c5
codec: document concurrency guarantees and fix analyze script warnings
|
7 years ago |
Ugorji Nwoke
|
ce1d126566
codec: misc fixes - handle error from io.Reader.Read, fix MissingFielder support, refactor for better mid-stack inlining, use custom bufio.Writer impl, etc
|
7 years ago |
Ugorji Nwoke
|
52c929c5b8
codec: defer call with closure giving weird errors - revert
|
7 years ago |
Ugorji Nwoke
|
f623015d6c
codec: put defer calls in closure - as runtime optimizes defer with no params
|
7 years ago |
Ugorji Nwoke
|
92b7607989
codec: refactor for better inlining, improve intf for string encoding, and fix some vet/staticcheck/false-sharing concerns
|
7 years ago |
Ugorji Nwoke
|
99ea80c8b1
codec: change signature to MissingFielder interface
|
7 years ago |
Ugorji Nwoke
|
a65ca39637
codec: use types for encode and decode errors, which expose underlying error
|
7 years ago |
Ugorji Nwoke
|
e5646b99d2
codec: support ability to preserve missing fields
|
7 years ago |
Ugorji Nwoke
|
0a18df59e4
codec: use decReader and encWriter interfaces (not the struct decReaderSwitch/encWriterSwitch)
|
7 years ago |
Ugorji Nwoke
|
1322318ca6
codec: set up codebase so as to take advantage of mid-stack inlining support
|
7 years ago |
Ugorji Nwoke
|
eeb0478a81
codec: if json and the struct field name (or encode name) is ascii and alphanumeric, just write it out
|
7 years ago |
Ugorji Nwoke
|
6f05ff1f01
codec: En|Decode should have at most 1 defer call
|
7 years ago |
Ugorji Nwoke
|
9af07b2035
codec: streamline chan encoding/decoding
|
7 years ago |
Ugorji Nwoke
|
97ffc5827b
codec: update that mid-stack linling is expected in go 1.11 development cycle
|
7 years ago |
Allen Luce
|
b5b9495648
codec: propagate Flush() errors at end of encode
|
7 years ago |
Ugorji Nwoke
|
ccfe18359b
codecgen: fix decStruct to handle all variations (pointers, fields, slice/array elem)
|
8 years ago |
Ugorji Nwoke
|
d41456c8f4
codec: refactor codecgen, optimize isEmptyValue and cache more static type info
|
8 years ago |
Ugorji Nwoke
|
c93fc2cdd1
codec: refactor struct field loading and data structures, for performance and memory friendliness
|
8 years ago |
Ugorji Nwoke
|
7bdfbc6818
codec: clean up codebase to try to keep lines within 90 bytes.
|
8 years ago |
Ugorji Nwoke
|
d2b2650062
codec: support mapping non-interface type to an interface
|
8 years ago |
Ugorji Nwoke
|
4ad9ed71b1
codec: re-factor and streamline code for publishing standards, consistency and performance
|
8 years ago |
Ugorji Nwoke
|
6e9891a451
codec: clean up, improved symbol handling and numeric overflow support
|
8 years ago |