Thu, 4 Oct 90 18:01:29 +0100

Related articles |
---|

UD-chains for record fields jml@wally.altair.fr (1990-10-04) |

Newsgroups: | comp.compilers |

From: | jml@wally.altair.fr |

Posted-Date: | Thu, 4 Oct 90 18:01:29 +0100 |

Keywords: | optimize |

Organization: | Compilers Central |

Date: | Thu, 4 Oct 90 18:01:29 +0100 |

I wonder if there has been any research done on the behavior of

records in dataflow analysis. What I mean is that the algorithms

for the building of ud-chains or alias analysis that I have read

about consider only simple variables and arrays. But field

selection also introduces aliasing. For example, after an

assignment such as

x.a = v;

v.b becomes an alias for x.a.b.

And when pointer-induced aliasing is compounded with field

selection, it seems to lead to rather formidable complexity:

s1: y = x;

...

s2: y->a = v;

...

s3: return (x->a.b);

There seems to be no definition of x->a.b reaching s3, though s2

is such a definition, due to aliasing of x to y in s1, even if

the field b does not appear explicitly in s2.

I'd be grateful if somebody could direct me to literature

on this problem.

---

Jean-Marie (John) Larcheveque

<jml@bdblues.altair.fr> or <jml@nuri.inria.fr>

--

Post a followup to this message

Return to the
comp.compilers page.

Search the
comp.compilers archives again.