Table:

Dictionaries, definitions, and compilers all use "hash" tables. Hash tables are setup for storing and retrieving Dictionaries, Keywords, Primetypes, definitions (for each dictinary, DDS_DICT), and symbols (for each binary DDS_BIN). When practical, all related data is kept in the same table. Page management is more efficient than general heap support.

A hash table is summarized by a DDS_TABLE structure. Storage for DDS_ITEMs in DDS_TABLE is described by DDS_PAGE. Pages also contain name strings and other non-DDS_ITEM data. Data can be appended to the active DDS_PAGE in DDS_TABLE. When the active page overflows, a new page is allocated. Previous pages are chained together via their initial entry. Only the last DDS_ITEM in the active page can be deleted.

Each entry in a hash table begins with a DDS_ITEM_CLASS struct. It keeps the basic description for an individual entry. Extended entries must inherit DDS_ITEM_CLASS (first member). Additional attributes are appended to the structure as needed. Examples include DDS_DEFN, DDS_DICT, and DDS_KEYWORD.

Support functions operate on DDS_TABLE, DDS_ITEM and DDS_PAGE. Data in a hash page may be "named", "anonymous", or "un-linked".

named: definitions, dictionaries, typedefs, variables.

Data is found using hash table lookup (fast, extendable). The "named" entries can be used to delimit an anonymous list.

anonymous: nested blocks, array types, control info.

Anonymous entries have zero length string names. They do not appear in the hash table itself. They do appear in the double linked list ("older", "newer") of all entries, along with "named" hash table entries. Data is found by linearly searching double linked lists. Anonymous lists are delimited by named hash table items; they share the same list and chronological sequence. The "hash" and "chain" values may be overloaded, i.e. used for other purposes if the the item is anonymous.

un-linked: strings, constants, variables, arrays.

Arbitary data may be pushed onto the same stack used by the hash table to manage named and un-named items. Data is found using extended attributes on table entries and references from outside the table structure.

Structures: defined in dds_table.h
    DDS_TABLE - hash table descriptor (see DDS_STACK and DDS_ITEM)
    DDS_PAGE  - page descriptor for hash table items
    DDS_ITEM  - generic item descriptor for hash table entries 

Globals:
    DDS_TABLE dds_prime_table;    /* prime type symbol table     */
    DDS_TABLE *dds_symbol;        /* generic global symbol table */


Files:
    tableinit.c     - Initialize hash table descriptor
    tablepush.c     - Push "size" bytes ("align" offset) onto table
    tablepushstr.c  - Push string onto table
    tablepushchar.c - Push one char for contigous array onto table
    tablepushitem.c - Create and zero "named" item onto hash table
    tablelookup.c   - Lookup "named" hash table entry
    tablerehash.c   - Rehash "named" hash table entries
    tableunhash.c   - Unhash "indexed" hash table entry
                         Unlink item from hash table "chain" and overflow "chain".
                         Retain item on hash table list "newer" and "older".
    tableterm.c     - Terminate "named" hash table
    hash.c          - Compute hash key, given name string
    primeup.c       - Return prime number >= num