Commit Graph

28 Commits

Author SHA1 Message Date
Adrian Conlon
16752474d9 Correct a casting issue in the rapidjson implementation
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-19 11:43:06 +01:00
Adrian Conlon
1943db6f20 rapidjson can use an std::string constructor.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-19 10:57:27 +01:00
Adrian Conlon
710c8fb883 Refactor the ram_t class to be a container of byte_t
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-19 09:28:13 +01:00
Adrian Conlon
2be190af19 Couple of small consistency changes in the rapidjson implementation.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-19 00:59:02 +01:00
Adrian Conlon
91bd04a278 Add support for rapidjson parsing
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-19 00:39:26 +01:00
Adrian Conlon
dc0fc262a2 More micro-optimisations. 2021-10-18 20:40:13 +01:00
Adrian Conlon
357e51c09a Clarify some (no) exception specifications.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-18 17:19:28 +01:00
Adrian Conlon
f62e8c30a2 Remove some unneeded default constructors.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-18 13:22:06 +01:00
Adrian Conlon
b56a2450a1 Couple of blank lines removed.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-18 13:13:43 +01:00
Adrian Conlon
1234aaf9d9 Refactor RAM test values.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-18 13:12:22 +01:00
Adrian Conlon
9cd317624e Couple of extra safety checks.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-18 12:28:15 +01:00
Adrian Conlon
fa50f74508 Tidy the previous cycle parsing refactor and reset to a simdjson pukka run.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-18 12:07:24 +01:00
Adrian Conlon
79f3e3ac6c Refactor the code to isolate the cycles parsing. Interesting speed up!
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-18 11:54:01 +01:00
Adrian Conlon
50cea1b81a Some JSON C++ library example code "levelling"
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-17 18:36:27 +01:00
Adrian Conlon
500e65b895 Tidy up the code a little (including removing some no longer needed code).
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-13 23:33:08 +01:00
Adrian Conlon
4892ea95d3 Split load/parse of opcode tests.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-12 17:07:45 +01:00
Adrian Conlon
dcba8efc83 All experimentation on the effect of JSON parser reuser.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-12 10:10:45 +01:00
Adrian Conlon
31c3a57485 If I've bothered to try and implement the failing test, show the actual and expected events.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-11 23:56:20 +01:00
Adrian Conlon
77a2c08c20 Remove some build warnings.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-11 23:23:59 +01:00
Adrian Conlon
91221aa975 Try simdjson: bloody hell, that's fast!!
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-11 22:09:03 +01:00
Adrian Conlon
b5074ac48c Create an overtly performance checking build.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-11 19:52:22 +01:00
Adrian Conlon
59d1e2789e Whoops: missed reference to RapidJson
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-11 19:39:37 +01:00
Adrian Conlon
7a7b0046cd Add an implementation that uses JsonCpp (pretty slow!)
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-11 19:13:05 +01:00
Adrian Conlon
f3f6452119 Add nlohmann json parser as an option to build.
1/2 speed, compared to boost.json

Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-11 14:59:23 +01:00
Adrian Conlon
db106b1719 Performance: speed up message handling in TestRunner.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-11 11:41:50 +01:00
Adrian Conlon
5686906583 Add nodiscard attributes, where needed
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-11 10:43:33 +01:00
Adrian Conlon
4be61a9d54 Separate concerns a little between low level classes and high level classes in terms of report generation.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-11 10:20:18 +01:00
Adrian Conlon
6a59bfbcd8 First stab at using the Harte randomised processor tests. Some failures detected in the M6502 run.
Signed-off-by: Adrian Conlon <adrian.conlon@gmail.com>
2021-10-10 21:26:30 +01:00