[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Issue: COMPILER-WARNING-STREAM (Version 6)



This one had just revision -> Version and KMP -> Pitman.

Status:  ready for re-release! No ? about it.

!
Issue:        COMPILER-WARNING-STREAM
References:   COMPILE (p438), COMPILE-FILE (p439)
Category:     CLARIFICATION/CHANGE
Edit history: Version 1 by Pitman 02/27/87
              Version 2 at committee meeting 15-Mar-87
              Version 3 Masinter 15-Mar-87
              Version 4 by Fahlman, incorporate Dribble
              Version 5 by Masinter, 29-May-87, revert to Version 3
              Version 6 by Masinter,  5-Jun-87, minor formatting
              

Problem Description:

The description of the COMPILE and COMPILE-FILE functions does not
explicitly permit them to print warnings. If this is to be allowed, it
should be an explicitly expressed part of the contract.

Proposal (COMPILER-WARNING-STREAM:ERROR-OUTPUT)

COMPILE and COMPILE-FILE are permitted to output warnings; warnings
should go to the stream that is the value of *ERROR-OUTPUT*.

Rationale:

Compiler warning output is a widely accepted extension to the
compilation. Warnings that come via the WARN function will go to the
stream that is the value of *ERROR-OUTPUT*.

Current Practice:

Some implementations send compiler warning output to *ERROR-OUTPUT*.
Other implementations send it to *STANDARD-OUTPUT*.

Adoption Cost:

In most cases, the change to the compiler to redirect output is expected
to be very slight.

Benefits:

Currently, it is difficult to redirect the output of COMPILE and
COMPILE-FILE because it isn't clear where it's directed.

Conversion Cost:

Most user programs that care are probably already tolerant of both
situations or naively expect that output will go to *ERROR-OUTPUT*. As
such, most users will probably perceive this as a clarification.

Aesthetics:

Most users will probably perceive this change as a simplification
because it will allow the kind of warning that comes from WARN and the
kind of warning that comes from compilation to be conceptually grouped.

Discussion:

This was a problem in adapting MACSYMA to Common Lisp because Macsyma
provides alternate user interfaces to the compiler which it needs to be
able to control.

The committee considered extending the proposal to describe the
interaction with DRIBBLE on the warning output, but found that DRIBBLE
was so underspecified as to make the task impossible. DRIBBLE should be
considered in a separate proposal.

The cleanup committee supports this change as stated.