Everyone on the LaTeX team is also on this list, right?
On 07/09/2008, at 11:14 PM, Joseph Wright wrote:
> Is there a reason expl3 has
> but no
> when the other (non-expanded) variants are all available as gpu and
> variations? As it stands, I have to use \clist_put_right:Nx for a
> global list: not good.
I've been looking into the seq, clist, toks, and tlp modules and
trying to normalise the commands that are provided for the put_left
and put_right cases in each. Seems to me like the sensible thing to do
is provide the same set in each for consistency's sake. Here's the
list I've ended up with:
% \seq_put_right:Nn |
% \seq_put_right:No |
% \seq_put_right:Nx |
% \seq_put_right:Nc |
% \seq_put_right:NC |
% \seq_put_right:cn |
% \seq_put_right:co |
% \seq_put_right:cc |
The idea is to support the whole lot for all of
in the four modules mentioned above.
This seems like too many to me.
\clist_put_right:NC is only used in xo-or.dtx , and the corresponding
global and/or left variants aren't used at all. I'm going to recommend
that the definition for this be moved into the xor package, then, and
we can omit defining it in expl3 whatsoever.
\seq_gput_right:Nc is only used in xo-alloc.dtx , so I'd suggest to
remove that variant from the expl3 bundle as well.
The :cc variant seems to be NEVER used, so lets drop that one too.
all come in with some level of frequency, so I'll recommend them as
the "base five" function variants to implement for the four modules.
I just noticed \toks_put_right:Nd and \toks_put_right:Nf .
The former is not used at all, the latter only in template.dtx.
Again, I'd recommend moving that to where it's used.
The alternative to moving these "one-shot" variants from expl3 is to
keep them in there but document them separately. I'm perfectly happy
with this solution, too.