Ferdinand Majerech
|
755eb4e468
|
Moved common FastCharSearch instantiations to Scanner body to minimize bloat.
|
2014-08-05 21:30:59 +02:00 |
|
|
Ferdinand Majerech
|
d505728824
|
Moved a branch outside of aloop in scanPlain() to improve performance.
|
2014-08-05 20:58:05 +02:00 |
|
|
Ferdinand Majerech
|
57d936ed0f
|
Scanner using prefixBytes() for optimization.
|
2014-08-05 20:57:30 +02:00 |
|
|
Ferdinand Majerech
|
3b303f6e82
|
An ASCII (bytes) version of prefix().
|
2014-08-05 20:56:59 +02:00 |
|
|
Ferdinand Majerech
|
8f94a40730
|
Doc fixes.
|
2014-08-05 20:56:30 +02:00 |
|
|
Ferdinand Majerech
|
568e75d3de
|
Removed decodeCount_, as it's no longer used.
|
2014-08-05 20:53:04 +02:00 |
|
|
Ferdinand Majerech
|
2b7ea42199
|
Removed the old (obsolete) Reader decoding method.
|
2014-08-05 20:52:43 +02:00 |
|
|
Ferdinand Majerech
|
92396b4cae
|
An optimized version of forward() with length == 1.
|
2014-08-05 20:52:05 +02:00 |
|
|
Ferdinand Majerech
|
34e6f55bd9
|
forward() now uses upcomingASCII_
|
2014-08-05 20:46:00 +02:00 |
|
|
Ferdinand Majerech
|
c828c6b132
|
peek()/slice() now use upcomingASCII()
|
2014-08-05 20:44:15 +02:00 |
|
|
Ferdinand Majerech
|
d9079de427
|
get() now calls slice() directly instead of through prefix().
|
2014-08-05 20:42:51 +02:00 |
|
|
Ferdinand Majerech
|
bfa2f1bd5c
|
Using checkASCII in Reader.
|
2014-08-05 20:42:22 +02:00 |
|
|
Ferdinand Majerech
|
e01c40ede5
|
Func to count consecutive ASCII chars starting at current Reader position.
|
2014-08-05 20:36:33 +02:00 |
|
|
Ferdinand Majerech
|
7409f3bbd9
|
ASCII optimizations for isPrintableValidUTF8.
|
2014-08-05 19:34:28 +02:00 |
|
|
Ferdinand Majerech
|
1c0702f3cd
|
Func to count the num of ASCII chars in string before the first UTF-8 sequence
|
2014-08-05 19:12:54 +02:00 |
|
|
Ferdinand Majerech
|
8902ea8806
|
Minor optimization.
|
2014-08-05 18:21:42 +02:00 |
|
|
Ferdinand Majerech
|
b2d0c74e56
|
Minor style fix.
|
2014-08-05 18:21:29 +02:00 |
|
|
Ferdinand Majerech
|
a89f9e93f7
|
Removed unnecessary memory allocations in Queue.
|
2014-08-05 18:20:57 +02:00 |
|
|
Ferdinand Majerech
|
a9333e3dd3
|
Fixed another 'in' parameter.
|
2014-08-05 18:15:07 +02:00 |
|
|
Ferdinand Majerech
|
9e573d85c7
|
.gitignore update.
|
2014-08-05 13:16:24 +02:00 |
|
|
Ferdinand Majerech
|
9ea269de87
|
Removed obsolete 'final' from Reader methods.
|
2014-08-05 13:14:42 +02:00 |
|
|
Ferdinand Majerech
|
3d8de67771
|
Using peekByte() where possible in Scanner.
|
2014-08-05 13:14:15 +02:00 |
|
|
Ferdinand Majerech
|
078269be36
|
Queue no longer supports types with destructors.
|
2014-08-05 13:12:07 +02:00 |
|
|
Ferdinand Majerech
|
7539b40d3d
|
(optimization) Mark ctor now doesn't check file column for overflow.
|
2014-08-05 13:10:52 +02:00 |
|
|
Ferdinand Majerech
|
57afd47bb5
|
Doc fixes.
|
2014-08-05 01:53:17 +02:00 |
|
|
Ferdinand Majerech
|
fb9525bb00
|
Obsoleting decodeCount_.
|
2014-08-05 01:52:21 +02:00 |
|
|
Ferdinand Majerech
|
75ed314dd6
|
More FastCharSearch.
|
2014-08-05 01:51:32 +02:00 |
|
|
Ferdinand Majerech
|
0424ff5e77
|
Style.
|
2014-08-05 01:51:20 +02:00 |
|
|
Ferdinand Majerech
|
44885cde4e
|
Optimized fetchToken()
|
2014-08-04 02:26:14 +02:00 |
|
|
Ferdinand Majerech
|
7360e85a3a
|
More FastCharSearch based on profiling results.
|
2014-08-04 02:24:26 +02:00 |
|
|
Ferdinand Majerech
|
5a1e6e994d
|
Fixed a nasty rare bug caused by an assumption that 32 chars take 32 bytes.
|
2014-08-04 02:23:08 +02:00 |
|
|
Ferdinand Majerech
|
20048ea995
|
Using peekByte() in heavily used Scanner methods.
|
2014-08-04 02:22:09 +02:00 |
|
|
Ferdinand Majerech
|
8e63f62d7e
|
An optimized version of peek() that reads a byte, without decoding.
|
2014-08-04 02:20:13 +02:00 |
|
|
Ferdinand Majerech
|
a4befdd866
|
An optimized version of Reader.peek() with index == 0.
|
2014-08-04 02:19:34 +02:00 |
|
|
Ferdinand Majerech
|
063d9754d7
|
Queue now uses a freelist to minimize allocations.
|
2014-08-04 02:16:34 +02:00 |
|
|
Ferdinand Majerech
|
1f55fccc26
|
Updated yaml_bench to benchmark parsing, not Loader construction.
|
2014-08-04 02:14:37 +02:00 |
|
|
Ferdinand Majerech
|
97e717df1b
|
Loader creates Constructor/Resolver lazily to avoid garbage when user-provided
|
2014-08-04 02:14:01 +02:00 |
|
|
Ferdinand Majerech
|
46a574c0b4
|
Fixed a typo.
|
2014-08-04 02:13:06 +02:00 |
|
|
Ferdinand Majerech
|
760e39479e
|
yaml_gen now gens strings with user-specified alphabet, with Unicode support.
|
2014-08-02 23:29:55 +02:00 |
|
|
Ferdinand Majerech
|
af4245811a
|
yaml_gen optimization.
|
2014-08-02 23:28:45 +02:00 |
|
|
Ferdinand Majerech
|
85b89d3323
|
yaml_gen should no longer generate maps/omaps with duplicate keys.
|
2014-08-02 23:28:33 +02:00 |
|
|
Ferdinand Majerech
|
0d89c2beb1
|
yaml_bench Scanner benchmark, and files are not realoaded from HDD by default.
|
2014-08-02 23:27:44 +02:00 |
|
|
Ferdinand Majerech
|
6aa50b8898
|
A benchmark Loader method that scans a file but throws away the tokens.
|
2014-08-02 23:26:46 +02:00 |
|
|
Ferdinand Majerech
|
c160156346
|
Fixed the string->char[] Token value move.
|
2014-08-02 23:25:56 +02:00 |
|
|
Ferdinand Majerech
|
aeee0758a7
|
Refactored FastCharSearch with more modern string mixin code.
|
2014-08-02 02:35:03 +02:00 |
|
|
Ferdinand Majerech
|
d32addacda
|
Slices now nonconst in all layers up to Parser, where they get cast to string.
|
2014-08-02 01:58:20 +02:00 |
|
|
Ferdinand Majerech
|
7b699c5903
|
UTF-8 validation now uses UTF-8 decoding code.
|
2014-08-02 01:37:16 +02:00 |
|
|
Ferdinand Majerech
|
b5da695d6b
|
More @nogc in Scanner.
|
2014-08-02 01:19:29 +02:00 |
|
|
Ferdinand Majerech
|
e6fdade4a6
|
Scanner now uses @nogc UTF decoding.
|
2014-08-02 01:16:29 +02:00 |
|
|
Ferdinand Majerech
|
e1209711af
|
UTF-8 decoding now has versions for validated and unvalidated strings.
|
2014-08-02 01:15:57 +02:00 |
|