Re: [Groff] refer question

[joerg van den hoff]

Peter Schaffter wrote:
> On Wed, May 04, 2005, joerg van den hoff wrote:
>
> > hi all,
> >
> > is there a way to reset refer settings somewhere 'downstream' if the 
> > 'upstream' defaults are no good?
> >
> > specifically:
> >
> > in my document I source a certain defaults file in which, upon different 
> > other settings, one finds:
> >
> > ...
> > .R1
> > database the_standard_database
> > .R2
> > ...
> >
> > now, I want to set in the _current_ document
> >
> > ...
> > .R1
> > database a_different_database
> > .R2
> > ...
> >
> > which by itself of course _appends_ the new database to the set of 
> > databases searched (and is searched latest).
> >
> > can I enforce solely searching of the new database without eliminating 
> > the first one from the defaults file (which would be bad for other 
> > documents...)? i.e. is there a way to clear the variable holding the 
> > database string (well, what is it's name? :-) )
>
>
> .R1
> no-default-database
> database a_different_database
> .R2
>
> instructs refer not to search the default database.  If later you
> need to have refer include the default database again:
>
> .R1
> default-database <filename>
> .R2
>
> I think that should work.

well, yes I should have read (again...) the refer manpage. so: thank you 
very much for your hint (actually to solve my problem I had to eliminate 
the refer command "database the_standard_database" from the sourced 
defaults file and instead set $REFER to this value. then, one has to use
database <filename> in the document (default-database takes no 
arguments...).

another problem:
I'm fiddling with getting the references the way the editor has ordered. 
this is (as has been remarked on this list) not straightforward for the 
uninitiated both for the format of the reference as well as for the 
label in the text. so what I would like are labels of the form, e.g.:

Miller, 2002; Jones, 2003; Jones, 2003a

i.e. the character enumeration should take effect _only_ if there is 
more than one reference of the kind _and_ the enumeration should affect 
only reference #2 and following.

what I have achieved up to now is

Miller, 2002; Jones, 2003a; Jones, 2003b

by using:

label A.n',\~'D.y%a*

(by the way: why does label A.n', 'D.y%a* issue an error message "too 
many arguments for command `label'" ?).

how can I get it completely right?


any help appreciated

joerg