Re: Numerals across code pages

From: spir (
Date: Fri Feb 19 2010 - 08:50:50 CST

  • Next message: Ed Trager: "Re: Numerals across code pages"

    On Fri, 19 Feb 2010 18:27:47 +0530
    "vainateya" <> wrote:
    > Different programming environments and applications seem to treat such
    > scenarios differently, most seem to expect an application level code /
    > routine which maps to equivalent ASCII before further processing or database
    > select query.

    I guess not.

    Let me situate the context as I understand it --I may be wrong. You are developping a "decoder" for a given source language, which should produce models representing decoded source code (abstract syntax trees AST) to be then either interpreted or compiled into a target language, using an implementation language.

    If no, then the rest of this post is irrelevant.
    If yes, then for sure your decoder misses a phase, namely the one that transforms expressions of literal values such as 1, 1.1, "hello", or (1,2,3) into actual values --in the implementation language. The said values beeing used as arguments for (arithmetic, logical...) operations. Concretely, this means that after identifying (matching) literals, you need to convert them using a function, that depends on their type. Eg for an integer:
       integer("<hindi 3>")
    should all return the same value --providing they are correct in the source language. Common parser generators allow directly specifying such post-matching actions in the grammar.

    What you were trying to do instead is to compare/process numbers according their string value -- which indeed makes no sense: basically "3">"123".

    Note that this is independent of the use of unicode in source (the first 3 literals above are plain ascii).
    Hope this helps.


    la vita e estrany

    This archive was generated by hypermail 2.1.5 : Fri Feb 19 2010 - 08:52:59 CST