Re: system() vs. fork()+exec*()

[Valeriy E. Ushakov]

On Mon, Jul 16, 2001 at 23:22:13 +0100, Graham Douglas wrote:

> Suppose that one wanted to have a new primitive called, say,
> @Script{...} which allowed embedded scripting language commands [or
> other program commands] to be embedded, I would like to know how
> you'd actually communicate the output to Lout?

That's the crucial question.  I don't mean technical details, I mean
semantically.  What is this supposed to *mean*?

And pragmatically, if you need a lot of complex scripting it actually
makes sense to preprocess or generate entire file rather than embed
scripting into the source processed by Lout.

E.g. I wrote a small tool to use Tibetan with Lout.  I don't call
external program from lout on every tiny bit of transliterated Tibetan
(and this has to be done on *each* run that is necessary for the
document to get xrefs converged &c), instead I preprocess the
transliterated source once.

SY, Uwe
-- 
address@hidden                         |       Zu Grunde kommen
http://www.ptc.spbu.ru/~uwe/            |       Ist zu Grunde gehen