R devel - Jul 2004 - Incorrect display of b[hat((a))] expression in plots

Hi, I am not sure if this is a bug or a non-implement feature of
text-drawing functions with TeX-style expression, but hat() and some of its
equals does not get the right "bounding boxes" if they are put in sub-
or
superscripts. For instance, for the expression 'b[hat(a)]' the hat()
seems
to shift 'hat(a)' too much to the right of 'b'. Try the below
example and
you'll see what I mean:

label <- list(
  expression((b[sqrt(a)])),
  expression((b[bar(a)])),
  expression((b[widehat(a)])),
  expression((b[widetilde(a)])),
  expression((tilde(a))),
  # Problematic:
  expression((b[dot(a)])),
  expression((b[ring(a)])),
  expression((b[hat(a)])),
  expression((b[tilde(a)])),
  expression((b^dot(a))),
  expression((b^ring(a))),
  expression((b^hat(a))),
  expression((b^tilde(a)))
)

plot(NA, xlim=c(0,2), ylim=c(-1,length(label)))
for (kk in seq(label))
  text(1,length(label)-kk, label=label[[kk]])

The problem occurs for the postscript and png devices too (I haven't tried
the others).

I'm on WindowsXP and "R version 1.9.1, 2004-06-21" (non-patched).

Best wishes

Henrik Bengtsson

Dept. of Mathematical Statistics @ Centre for Mathematical Sciences 
Lund Institute of Technology/Lund University, Sweden (+2h UTC)
+46 46 2229611 (off), +46 708 909208 (cell), +46 46 2224623 (fax)
h b @ m a t h s . l t h . s e, http://www.maths.lth.se/~hb/

Confirmed.

I might look at it during August, since the underlying code is not that 
new to me. If Paul (or anybody else from R Core) has not answered yet 
and there is not already a related bug report in the database, I'd 
suggest to submit a bug report.

Uwe

Henrik Bengtsson wrote:
> Hi, I am not sure if this is a bug or a non-implement feature of
> text-drawing functions with TeX-style expression, but hat() and some of its
> equals does not get the right "bounding boxes" if they are put in
sub- or
> superscripts. For instance, for the expression 'b[hat(a)]' the
hat() seems
> to shift 'hat(a)' too much to the right of 'b'. Try the
below example and
> you'll see what I mean:
> 
> label <- list(
>   expression((b[sqrt(a)])),
>   expression((b[bar(a)])),
>   expression((b[widehat(a)])),
>   expression((b[widetilde(a)])),
>   expression((tilde(a))),
>   # Problematic:
>   expression((b[dot(a)])),
>   expression((b[ring(a)])),
>   expression((b[hat(a)])),
>   expression((b[tilde(a)])),
>   expression((b^dot(a))),
>   expression((b^ring(a))),
>   expression((b^hat(a))),
>   expression((b^tilde(a)))
> )
> 
> plot(NA, xlim=c(0,2), ylim=c(-1,length(label)))
> for (kk in seq(label))
>   text(1,length(label)-kk, label=label[[kk]])
> 
> The problem occurs for the postscript and png devices too (I haven't
tried
> the others).
> 
> I'm on WindowsXP and "R version 1.9.1, 2004-06-21"
(non-patched).
> 
> Best wishes
> 
> Henrik Bengtsson
> 
> Dept. of Mathematical Statistics @ Centre for Mathematical Sciences 
> Lund Institute of Technology/Lund University, Sweden (+2h UTC)
> +46 46 2229611 (off), +46 708 909208 (cell), +46 46 2224623 (fax)
> h b @ m a t h s . l t h . s e, http://www.maths.lth.se/~hb/
> 
> ______________________________________________
> R-devel@stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-devel

Paul Murrell wrote:
> Hi
> 
> 
> Uwe Ligges wrote:
> 
>> Confirmed.
>>
>> I might look at it during August, since the underlying code is not 
>> that new to me. If Paul (or anybody else from R Core) has not answered 
>> yet and there is not already a related bug report in the database,
I'd
>> suggest to submit a bug report.
> 
> 
> 
> Yep, bug not lack-of-implementation.
> 
> Uwe:  sounds like you might get to this before me.  Some possibly useful 
> observations from a brief look yesterday:
> (i)  Looks like only the actual drawing that is wrong;  the bounding box 
> calculations appear correct (the parentheses around the outside look 
> about right)
> (ii)  The offending accents are all common to RenderAccent() (they all 
> appear in AccentTable)



Paul, Henrik,

here we go - later than promised, but not too late for 2.0.0, I hope.


The last lines of RenderAccent() need:

+   if(draw)
     PMoveTo(savedX + width, savedY, mc);

rather than just

     PMoveTo(savedX + width, savedY, mc);

and it took me more than 2 painful hours to get the point! :-(

(Do we have a PR#? I don't find it.)

Uwe


> Paul
> 
> 
>> Henrik Bengtsson wrote:
>>
>>> Hi, I am not sure if this is a bug or a non-implement feature of
>>> text-drawing functions with TeX-style expression, but hat() and
some
>>> of its
>>> equals does not get the right "bounding boxes" if they
are put in
>>> sub- or
>>> superscripts. For instance, for the expression 'b[hat(a)]'
the hat()
>>> seems
>>> to shift 'hat(a)' too much to the right of 'b'. Try
the below example
>>> and
>>> you'll see what I mean:
>>>
>>> label <- list(
>>>   expression((b[sqrt(a)])),
>>>   expression((b[bar(a)])),
>>>   expression((b[widehat(a)])),
>>>   expression((b[widetilde(a)])),
>>>   expression((tilde(a))),
>>>   # Problematic:
>>>   expression((b[dot(a)])),
>>>   expression((b[ring(a)])),
>>>   expression((b[hat(a)])),
>>>   expression((b[tilde(a)])),
>>>   expression((b^dot(a))),
>>>   expression((b^ring(a))),
>>>   expression((b^hat(a))),
>>>   expression((b^tilde(a)))
>>> )
>>>
>>> plot(NA, xlim=c(0,2), ylim=c(-1,length(label)))
>>> for (kk in seq(label))
>>>   text(1,length(label)-kk, label=label[[kk]])
>>>
>>> The problem occurs for the postscript and png devices too (I
haven't
>>> tried
>>> the others).
>>>
>>> I'm on WindowsXP and "R version 1.9.1, 2004-06-21"
(non-patched).
>>>
>>> Best wishes
>>>
>>> Henrik Bengtsson
>>>
>>> Dept. of Mathematical Statistics @ Centre for Mathematical Sciences
>>> Lund Institute of Technology/Lund University, Sweden (+2h UTC)
>>> +46 46 2229611 (off), +46 708 909208 (cell), +46 46 2224623 (fax)
>>> h b @ m a t h s . l t h . s e, http://www.maths.lth.se/~hb/
>>>
>>> ______________________________________________
>>> R-devel@stat.math.ethz.ch mailing list
>>> https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
>>
>>
>>
>> ______________________________________________
>> R-devel@stat.math.ethz.ch mailing list
>> https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
> 
> 
>