"Denis M. Wilson" <d...@oxytropis.plus.com> wrote: |On Fri, 14 Nov 2014 15:16:47 +0100 |Steffen Nurpmeso <sdao...@yandex.com> wrote: |>|Since the evaluation is done at runtime I can't see a problem \ |>|with variables, |>|although it would enable some very obfuscated code... |> |> I don't understand this, actually. | |A string variable could expand to a subexpression. The sort of dirty |trick we used in writing assembler code.
For all-assembler i mostly worked with nasm: no dirty tricks there :) (Except sometimes to make C constant #define's compatible with it.) You mean that if there were a to_string() function that turns the nodes that have been parsed into strings then it wouldn't matter wether only the value to test or also the regular expression that shall be used for testing came from such an expansion. Most likely right, except for what Tadziu said regarding the syntax and the clash in between roff and regular expression syntax, asking for trouble / is a quoting nightmare. But first of all: there is yet no to_string() function, so all this is hypothetic. --steffen
