Sent with Proton Mail secure email.
On Monday, January 19th, 2026 at 4:55 PM, Hans Hagen <[email protected]> wrote:
>
>
> On 1/19/2026 2:55 PM, luigi scarso wrote:
>
> > On Wed, 14 Jan 2026 at 10:56, Udi Fogiel <[email protected]
> > mailto:[email protected]> wrote:
> >
> > Hello,
> >
> > Here is what happens when \vcenter is encountered:
> >
> > staticvoidrun_vcenter(void){
> > scan_spec(vcenter_group);
> > normal_paragraph();
> > push_nest();
> > mode = -vmode;
> > prev_depth_par = ignore_depth;
> > if(every_vbox_par != null)
> > begin_token_list(every_vbox_par, every_vbox_text);
> > }
> >
> > So unlike other boxes is uses scan_spec and not scan_full_spec,
> > which means it cannot accept the attr and (b)dir keywords, and the
> > directions registers are not set to its direction
> > inside of it.
> >
> > Then when it finishes we have:
> >
> > voidfinish_vcenter(void)
> > {
> > halfword p;
> > unsave();
> > save_ptr--;
> > p = vpack(vlink(cur_list.head_field), saved_value(0),
> > saved_level(0), -1);
> > pop_nest();
> > p = math_vcenter_group(p);
> > tail_append(p);
> > }
> >
> > And because of the -1 in the last argument, its direction is
> > determined by \bodydirection,
> > unlike other boxes which in math context use \mathdirection if a
> > (b)dir keyword was not specified.
> >
> > Why is that? It feels a bit inconsistent, and makes it a bit hard to
> > follow directions.
> >
> > I guess since mathematical typesetting is quite complex and we want to
> > maintain some degree of compatibility with pdftex, \vcenter remains
> > largely unchanged
> > (on the other hand, I think using the token lib it is possible to get
> > something similar to \hbox/\vbox for \vcenter) .
>
>
> also, in luatex, vcenter is a math only command and injects a noad and
> not a box (vlist) so introducing attr and dir there would also mean
> figuring out how to deal with it in the perspective of math (like what
> should get the attr and when); as vcenter is often used low level and
> can wrap any content one can mess with that instead
>
> so here compatibility is indeed the verdict (changing this after two
> decades is also a bit dangerous wrt compatibility)
>
Thanks for the answers luigi and Hans.
The fact that \vcenter does not accept a (b)dir and attr keywords
does make sense, but what more bother me is if the side effects of that, i.e.
setting the vlist direction as \bodydirection rather than \mathdirection (which
I find it odd for math typesetting), and not setting \textdirection,
\pardirection and \bodydirection inside the group (which is a bit
inconsistent), are desired.
At the end these facts are not really bad, just peculiarities I guess. So
unless you think it should be changed
probably just more technical documentation about how the directions model
work in LuaTeX will make this less surprising. I don't mind working on that at
some point.
Udi
_______________________________________________
dev-luatex mailing list -- [email protected]
To unsubscribe send an email to [email protected]
