Re: manunal page - uninterpretted dot

[Bjarni Ingi Gislason]

In <address@hidden> Jan =?utf-8?B?VsSNZWzDoWs=?= <address@hidden> writes:

> Hello,
> 
> some time ago I reported a bug about not working backslash-period at the 
> beginning of the line. This was evaluated as a documentation bug.
> 
> http://www.mail-archive.com/address@hidden/msg01499.html
> 
> I wanted to send a patch, but I'm really not sure what exactly should the 
> documentation say.
> 
> I believe this is the most confusing part in groff(7):
> > \.     An uninterpreted dot (period), even at start of line.
> 
> AFAIK the dot character is just the default control character and has a 
> special meaning only at the beginning of the line. Nowhere else. If the \. 
> sequence cannot be used to suppress the meaning of control character at the 
> beginning of the line, it makes no sense to use this sequence at all. Am I 
> right? So what about removing this line from documentation at all?
> 
> Jan

  The explanation in groff(7) is wrong.  '\.' is simply an another name
for the dot (.) (see the original troff manual).

  If the dot (period, full stop) shall be printed use "\&.", if the dot
starts a word (is after a blank or new line character).

  The groff(7) manual says: "The special behavior can be delayed by
using the \. escape."  '\\.' has to be used.  Or say "by using an
escaped \. escape (\\.)"

  "\." is only used in macros (and then escaped!), when the request
shall only be executed when the macro is executed (e.g. a macro is
defined in another macro).  Otherwise the request (the control
character) is executed when the macro is read ("copy mode").

  Example:

\.nr var 23
Some text.
\.tm var = \n[var]
Second sentence.
\\.tm var = \n[var]

  Output from "\.tm" is

var = 23


  Get the original troff manuals "54.ps.gz" from the server

"cm.bell-labs.com"

  and the file

/cm/cs/cstr/54.ps.gz

  See chapter 7.5 with the definition (explanation) of ".de xy":

  ".." can be concealed as \\.. which will copy as \.. and reread as "..".

  Which means an unescaped "\." is read as ".".