Translating method invocations

There are two cases to consider. The case where the result of the method is not a multiarray, and the case where the result is a multiarray. First we consider the case where the result is not a multiarray. The general scheme is illustrated in Figure A.31. This scheme is modified in trivial ways if the method is static, or is applied to the current object.

The macro TRANS_ARGS evaluates to a list of translated expressions. We will define it in terms of the simpler macro TRANS_ARG which operates on a single argument. If the invoked method (or, in later sections, constructor) is not an HPspmd method (or constructor) the definition is:

$$\begin{minipage}[t]{\linewidth}\begin{tabbing} $\mbox{\it TR... ...\it TRANS\_ARG}\left(e_{N-1}\right)$\end{tabbing}\end{minipage}$$

If the invoked method (or constructor) is an HPspmd method (or constructor), the definition is:

$$\begin{minipage}[t]{\linewidth}\begin{tabbing} $\mbox{\it TR... ...N-1}\right)$\verb$, $\mbox{\it apg} \end{tabbing}\end{minipage}$$

Now, if $e$ is not a multiarray, then

$$\begin{minipage}[t]{\linewidth}\begin{tabbing} $\mbox{\it TR... ...equiv \quad {\bf T}\left[{e}\right]$\end{tabbing}\end{minipage}$$

Otherwise, if $e$ is a multiarray expression of rank $R$, then

$$\begin{minipage}[t]{\linewidth}\begin{tabbing} $\mbox{\it TR... ...{\mbox{\small dat}}\left[{e}\right]$\end{tabbing}\end{minipage}$$

In other words, each multiarray argument is split into $2 + R$ arguments.

Figure A.32: Translation of multiarray-valued method invocation expression.

SOURCE:

$$\begin{minipage}[t]{\linewidth}\begin{tabbing} $v$ \verb$ =... ...\ldots$\verb$, $$e_{N-1}$\verb$) ;$ \end{tabbing}\end{minipage}$$

TRANSLATION:

$$\begin{minipage}[t]{\linewidth}\begin{tabbing} \verb$DAD $$d... ...$ = $$\mbox{\it TRANS\_DIM}(d, R-1)$\end{tabbing}\end{minipage}$$

where:

$$\begin{array}{l} \mbox{The expression $v$ is the assigned a... ...M}} \\ \quad \quad \mbox{are defined in the text.} \end{array}$$