[Groff] Re: Error in meref.me?

[Werner LEMBERG]

> I've never tried it because I never really understood what \E did -
> am I right to say that this a backslash which is *only* interpreted
> at the very last moment, and is passed through however many levels
> of copy mode without interpretation?

Yes.

> So, conceivably, it could be used *whenever* traditional troff
> doubles backslashes?

Hmh, almost.  I think that `\E.' doesn't work as a replacement for
`\\.', but I haven't tested this yet.  It also doesn't work in the
`.output' command, IIRC, but I'm too lazy to verify this.

> If so, perhaps a note to that effect could be put in the 
> groff_tmac  manpage, and it could be made a little more explicit in 
> groff info.
> 
> Suggested change to groff_tmac.5.  At the end of the section on 
> "Copy-in mode":
> 
> .P
> As a GNU extension, the entire problem of when to double backslashes 
> can be circumvented by using the \[rs]E escape sequence.
> .
> This is, in effect, a backslash which is only interpreted at the last 
> possible moment.
> .
> Thus, in the example of positional parameters above, one can write 
> \[rs]E$* instead of \[rs]\[rs]$*.

I don't see a real benefit in writing `\En[foo]' instead of
`\\n[foo]'.  The proper solution IMHO is to disable the escape
character while defining macros:

  .eo
  ...
  \n[foo]
  ...
  .ec

> This alternative form becomes particularly useful when an escape 
> sequence must copied-in two or three times.

Here it really makes sense.

Please reformulate your change.


    Werner