Content-Type: text/plain; charset=UTF-8
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
References: <0BFA7483-6E3E-45D0-9094-3DEE5D880A1E@gmail.com>
Importance: normal
Sensitivity: Normal
Message-ID:  <1816801436.1740507.1284759290533.JavaMail.fmail@mwmweb010>
Date:         Fri, 17 Sep 2010 23:34:50 +0200
Reply-To: Mailing list for the LaTeX3 project
              <LATEX-L@listserv.uni-heidelberg.de>
Sender: Mailing list for the LaTeX3 project <LATEX-L@listserv.uni-heidelberg.de>
From: Uwe Lueck <uwe.lueck@WEB.DE>
Subject: Re: Consistently active characters
To: LATEX-L@listserv.uni-heidelberg.de
In-Reply-To:  <0BFA7483-6E3E-45D0-9094-3DEE5D880A1E@gmail.com>
Precedence: list
Status: R

Will Robertson <wspr81@GMAIL.COM>, 17.09.2010 07:21:30:
>
> \def\foo{...
>   \lccode`#1=`\~

! Illegal parameter number in definition of \foo.

And according to my own experience, it must be \lccode`\~=`#1..., 
cf. doc.sty \@sverb/\do@noligs and below.

>   \lowercase{%
>     \def ~{...}
>   }
> }
>
> But this requires that ~ is naturally active; 

Sure? In my proposal below, ~ is only locally active.

> and that's not necessarily always a good assumption 
> (well, in expl3 isn't definitely not).

Arno Trautmann [, 17.09.2010 21:53:19
>
> \define_active_char:Nn <character> {<code>}
>
> which makes the <character> active and assigns the given code to it,
> internally using the code that Will wrote.

    \begingroup
      \catcode`\~=13
      \gdef\define_active_char:Nn#1#2{%
        \catcode`#113%
        \begingroup
          \lccode`\~`#1%
          \lowercase{\endgroup \def ~{#2}}%
      }
    \endgroup

Not tested exactly this way, sorry, but in a so far private package of mine, I have

    \providecommand*{\MakeActiveDef}[2]{%
      \catcode`#1\active
      \begingroup
        \lccode`\~`#1%
        \lowercase{\endgroup \def ~{#2}}%
    }

so (unless another definition of \MakeActiveDef occurs earlier) \MakeActiveDef and \define_active_char:Nn almost have the same meaning (\active vs. 13, in case LaTeX3 won't have \active).

The \endgroup trick here is from doc.sty's \do@noligs.

Admittedly \MakeActiveDef has a single instance (though vital) at present, but I have an analogous \MakeActiveLet in niceverb.sty (ctan.org/pkg/nicetext) that has several vital instances (active characters) and has been used in at least one large-scale project.

Actually, I prefer \MakeActiveLet to \MakeActiveDef, \MakeActiveDef only exists because I was (for some minutes) unable to use \MakeActiveLet in that other package. I favor \MakeActiveLet for the case that the meaning of a single active character might be wanted to change, so you store intended meanings as macros \@if@active@i, \@if@active@ii, ... and then switch

    \MakeActiveLet\<char>\@if@active@i

somewhere,

    \MakeActiveLet\<char>\@if@active@ii

somewhere else ...

Cheers, 

    Uwe.


Will Robertson <wspr81@GMAIL.COM>, 17.09.2010 07:21:30:
>
> \def\foo{...
>   \lccode`#1=`\~

! Illegal parameter number in definition of \foo.

And according to my own experience, it must be \lccode`\~=`#1..., 
cf. doc.sty \@sverb/\do@noligs and below.

>   \lowercase{%
>     \def ~{...}
>   }
> }
>
> But this requires that ~ is naturally active; 

Sure? In my proposal below, ~ is only locally active.

> and that's not necessarily always a good assumption 
> (well, in expl3 isn't definitely not).

Arno Trautmann [, 17.09.2010 21:53:19
>
> \define_active_char:Nn <character> {<code>}
>
> which makes the <character> active and assigns the given code to it,
> internally using the code that Will wrote.

    \begingroup
      \catcode`\~=13
      \gdef\define_active_char:Nn#1#2{%
        \catcode`#113%
        \begingroup
          \lccode`\~`#1%
          \lowercase{\endgroup \def ~{#2}}%
      }
    \endgroup

Not tested exactly this way, sorry, but in a so far private package of mine, I have

    \providecommand*{\MakeActiveDef}[2]{%
      \catcode`#1\active
      \begingroup
        \lccode`\~`#1%
        \lowercase{\endgroup \def ~{#2}}%
    }

so (unless another definition of \MakeActiveDef occurs earlier) \MakeActiveDef and \define_active_char:Nn almost have the same meaning (\active vs. 13, in case LaTeX3 won't have \active).

The \endgroup trick here is from doc.sty's \do@ligs.

Admittedly \MakeActiveDef has a single instance (though vital) at present, but I have an analogous \MakeActiveLet in niceverb.sty (ctan.org/pkg/nicetext) that has several vital instances (active characters).

Actually, I prefer \MakeActiveLet to \MakeActiveDef, \MakeActiveDef only exists because I was for some minutes unable to use \MakeActiveLet in that other package. I favor \MakeActiveLet for the case that the meaning of a single active character might be wanted to change, so you store intended meanings as macros \@if@active@i, \@if@active@ii, ... and then switch

    \MakeActiveLet<char>\@if@active@i

somewhere,

    \MakeActiveLet<char>\@if@active@ii

somewhere else ...

Cheers, 

    Uwe.