> Ross very speedily fixed an alltt related problem for me a few months
> ago by introducing the $alltt_rx variable. This works fine in most
> cases but I've just come across a very strange case that I am at a
> loss to explain. Take the following file
>
...
> %\chcol{firebrick4}
> \begin{doutput}
> Enter a key: john
> No information about `john'
> Enter a key: John
> Info for `John' is
> Name: John Latham Phone: 6250 Room: CB-2.113
> Enter a key: ^D
> \end{doutput}
> \end{document}
> This (using l2h982b8) produces a <TT> environment but without the
> <BR>'s at the end of each line, which was the problem I reported to
> you in the first place. However, if I delete the **comment** line
> %\chcol{firebrick4}
> that occurs before the start of the douput environment everything
> works fine!
>
> Any ideas what's going on here?
Yes. It is the test for an alltt-like environment bein hidden
behind comments that is failing; e.g.
% \begin{alltt}
something here
and some more ...
The above must *not* put in the <BR>s.
However the following should:
% a comment
\begin{alltt}
something here
and some more
...
\end{alltt}
The test for this is being done as follows:
local(@check) = split("\n",$before);
local($lastline) = pop @check;
$alltt = &alltt_helper($alltt) # shield special chars
unless ($lastline =~ /(^|[^\\])(\\\\)*%.*$/m); # unless commented out
The intention is that the \n at the end of $before in the 2nd case,
should leave an empty last element in @check , so that the 'unless' test
of $lastline will fail and the <BR>s will be put in.
But this doesn't happen; instead the pop @check; is picking up
the commented line as $lastline; ergo, no <BR>s.
Fix it by inserting an extra conditional, as follows:
local(@check) = split("\n",$before);
local($lastline) = pop @check unless ($before =~ s/\n$//s);
> Thanks
>
Hope this helps.
Ross Moore
PS, http://www-texdev.mpce.mq.edu.au/ptest/tabtest/
has the result of testing this.
It also shows my latest efforts to clean up enumerated lists.
