You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At the moment OZ80 uses a very simple linear list of word tokens. This is efficient at walking, but it means that we are doing a lot of structural validation as we execute the user-program. Execution and assembling are already very fast compared to tokenisation, but to speed things up more (especially when we introduce loops and macros) I want to be able to do away with the validation during execution.
The solution to this is to build an Abstract Syntax Tree instead of a linear token list and do a large amount of structural validation during the tokenisation stage. This will also allow us to pre-allocate Label and Value names and link those to the tokens in question, sparing us from the painful Dictionary object lookups currently used throughout.
The text was updated successfully, but these errors were encountered:
At the moment OZ80 uses a very simple linear list of word tokens. This is efficient at walking, but it means that we are doing a lot of structural validation as we execute the user-program. Execution and assembling are already very fast compared to tokenisation, but to speed things up more (especially when we introduce loops and macros) I want to be able to do away with the validation during execution.
The solution to this is to build an Abstract Syntax Tree instead of a linear token list and do a large amount of structural validation during the tokenisation stage. This will also allow us to pre-allocate Label and Value names and link those to the tokens in question, sparing us from the painful Dictionary object lookups currently used throughout.
The text was updated successfully, but these errors were encountered: