??=[arrows.targets]
In the general arrow constructions ||, ||, ||,
||, and ||, the target address can be given in a
large number of formats called "positions". The full range of
possibilities is described in the reference
manual [#!R94:XY-picRM!#]; here is a number of useful possibilities:
-
|[| r|,| c|]| , where r, c are integers, denotes the "relative
entry" found r rows below and c columns to the right of the
current entry (the current entry itself is thus |[0,0]|). Each such
pair corresponds to a |[|"hop"|]| as described in
section ??[basics.arrows], , |[1,2]| is the same as |[drr]|.
-
|"| r|,| c|"| , where r, c are positive integers, denotes the
"absolute entry" found in the rth row and cth column of the
diagram; the top left entry is |"1,1"|.
- t'|;| t , where t' is any target, changes the ??!w[base entry] of
the present arrow to t' and then sets the target to t relative to
the original base. For example,
![\begin{code}
\diagram
A \\
B & C \ulto <1ex>
\ulto;[] <1ex>
\enddiagram
\end{code}](img50.png)
typesets
the || arrow "starts" at the |[ul]| entry and ends in the
current entry (remember from ยง??[basics.arrows] that ||
is the same as |[ul]|).
Composite targets may be constructed: any complete target can be
followed by
- ??c![+]"vector" or ??c![-]"vector" which changes the target to be a
zero-sized one at the position obtained by adding or subtracting the
"vector" to its center, or
- ??c![!]"vector" which moves the center of the target by the "vector";
where a "vector" should have the form
-
| < | Dx|,| Dy| > | , where Dx, Dy are TEX dimensions, is the
vector with those coordinates,
- the following `corner offsets' of a target are vectors as shown:
0.(- 11, -6).(11, 6)!
C = "
box"*
.,"
box";"
box" +
L**
.,"
box" +
R**
.,"
box" +
D**
.,"
box" +
U**
.
L]
R]
D]
U]
DL]
DR]
UL]
UR]
(
theymustbespecifiedinuppercase),
and
| 0| isthezerovector.
These constructions are useful for pointing to corners of entries,
,
![\begin{code}
\diagram
\left[\sum^i\right] & j \lto+UR
\enddiagram
\end{code}](img55.png)
will typeset
The position language of "v3" is much richer than this: there it is
possible to build stacks of positions, typeset material in the middle
of locating a position, etc.
