Add Json writer for cp2077-nativedb #7
Conversation
…abase. Writer create individual files per definition. Writer create grouped files per type of definitions (enums/bitfields/classes/globals). Writer use an obfuscated Json schema to minimize memory footprint (even better when compressed with Brotli). Writer only include required fields. Writer reformat and compress flags with only known and reused data. Writer is to be used by webapp 'cp2077-nativedb'.
…abase. Writer create individual files per definition. Writer create grouped files per type of definitions (enums/bitfields/classes/globals). Writer use an obfuscated Json schema to minimize memory footprint (even better when compressed with Brotli). Writer only include required fields. Writer reformat and compress flags with only known and reused data. Writer is to be used by webapp 'cp2077-nativedb'.
…s/RED4.RTTIDumper into nativedb-json-writer
|
Thanks! Overall looks good but I have few questions:
|
Flags are "compressed" such as only known bits / used bits in web app are stored in Json. It writes primitive types / template types as numbers too. It also writes definitions in grouped files (enums.json / bitfields.json / classes.json / globals.json) so only a few files are requested in web app. Fullname / shortname is only stored when there is a difference for example. The main purpose is to make Json data as compressed as possible to minimize bandwidth usage.
I believe that reusing identical names improves compression ratio, as it should provide a minimum of different symbols, resulting in a smaller memory footprint (e.g. 10MB to 800KB with Brotli). |
|
Ah, I see. Thank you for the explanation! |
Custom AST Json writer to be used with webapp cp2077-nativedb.