Marking translatable strings

[François Pinard]

Greg Ewing <greg.ewing at compaq.com> writes:

> Tim Peters wrote:
> > 
> > > Experience shows that horizontal space is often a valuable resource while
> > > writing code, so we should try to be economical there for overall
> > > legibility, much more than typability.

> How about hacking the unary "+" operator on strings to
> do translation? It would be hard to get any more horizontally
> economical than that, and it wouldn't require any change to
> the scanner or parser.

It would need a change to Python, as currently:

   TypeError: bad operand type(s) for unary +

but indeed, if such a change was accepted, it would be especially neat.
For example:

        "printing %s" % +"hello"
        +"printing %s" % "hello"
        +"printing %s" % +"hello"

would allow precise, yet unobtrusive control over exactly what gets
translated.  I'm presuming, here, that unary `+' binds very tightly.

Would you foresee any difficulty while writing a _simple_ tool, to be
integrated in `xgettext' (or `xpot':-) that would properly identify unary
`+' operators before strings, I mean, never mistaking a binary `+' as unary?

-- 
François Pinard   http://www.iro.umontreal.ca/~pinard