Dmitry Olshansky
19778b1c7a
address review comments
2016-03-28 12:54:59 +03:00
Dmitry Olshansky
df07aa7dea
Special case ASCII to use bit tables insted of 2-level tries
2016-03-27 20:26:49 +03:00
Dmitry Olshansky
0e55583fed
Start small - direct dispatch tables
2016-03-27 16:49:44 +03:00
Dmitry Olshansky
ec7b647d27
Segregate opcodes to functions; preparation to JIT
...
Then generate sequence of function calls to them..
2016-03-27 16:24:03 +03:00
Dmitry Olshansky
9d445d182d
Introduce filtered loops with bloom filters
...
Make use of bloom filter on dchar to predict if we should take the
out of loop branch.
2016-03-21 22:06:56 +03:00
Dmitry Olshansky
88c3a3d52f
2 minor tweaks to std.regex default engine
...
Roughly 10-15% faster on a few common patterns.
2016-03-07 15:46:30 +03:00
Martin Nowak
ce2ac192c9
fix incorrect imports/fqn usages
...
- mostly of the form `import a.b : sym; a.b.sym();`, which is wrong b/c
selective imports do not add the module to the current scope
2016-02-20 14:41:44 +01:00
k-hara
e7b3d834d7
detab and remove trailing whitespaces
2015-02-10 00:56:40 +09:00
Dmitry Olshansky
2b78074fc2
Split up the intertwined mess of std.regex.
...
The docs and API still stay in one file.
With time and refactoring more internals may be
exposed such as parser, each engine explicitly
and the sample generator (generator.d).
Also inclusdes changes prompted by reviews/pulls:
Convert spaces-->tabs in makefiles.
Move things (again) to std/regex/internal.
Use new package(std.regex) feature.
Remove C-style arrays (some other pull against regex.d).
2014-09-13 13:45:46 +04:00
