Changelog

Differences from Official J

The current differences from the latest J804 binary release are any undocumented differences from the official source release plus:

v0.3

• Add CONFIG_TALLOC option. Setting it on Linux replaces malloc calls with talloc calls which allows freeing the JT struct
• Add 9!:66 which returns the address of the JT struct which can be read with the 15!: family for advanced usage
• Add proof-of-concept csv reader.X 16!:1 Y reads a csv file into a boxed J table. Y is a boxed file name, X is a literal list of column types, 'l' for literal and 'n' for numeric, e.g., 'lnn' 16!:1 <'data.csv'

v0.2

• Allow unicode identifiers (UTF-8)
• Fix (13 :) for input like 'y+y' and 'y<y'
• Improve hashing for names, preventing O(n^2) performance for n names with simple patterns

v0.1

• Fix crash in ^: with nouns in gerund arguments (like ]^:(1:(<,'0';'string')) 0`)

• 9!:64 y Query Architecture. Returns numerical 64 or 32

• Fix wrong result from _2 1 E. 1 2 _1 _2 1 and http://code.jsoftware.com/wiki/System/Interpreter/Bugs/Crashes#numeric_dyad_E._corrupts_memory

• Fix crash from

       verb=: 0:`NOUN@.(0=#)
       NOUN =: 'string',LF
       verb ''