constraint violations and compilation

[Richard Weeks]

My question, stated as simply as possible, is: must a constraint
violation, as defined by the standard, cause compilation to halt?

I believe that in the presence of a constraint violation, at
least one diagnostic is required, which, at the discretion of the
compiler writer, may be either an "error" or a "warning."
Generally however, an error halts compilation and a warning does
not. This would seem to imply that it is up to the compiler
writer to decide which constraint violations halt compilation and
which do not.

Richard

[=?iso-2022-kr?q?Harald_van_D=0E=29=26=0Fk?=]

On Mon, 01 Oct 2007 22:06:37 +0000, Richard Weeks wrote:

My question, stated as simply as possible, is: must a constraint
violation, as defined by the standard, cause compilation to halt?

No.

I believe that in the presence of a constraint violation, at least one
diagnostic is required, which, at the discretion of the compiler writer,
may be either an "error" or a "warning."

It may be any diagnostic. Several compilers have diagnostic categories
other than "error" and "warning", for example "fatal error",
"notification", or "remark".

Generally however, an error
halts compilation and a warning does not. This would seem to imply that
it is up to the compiler writer to decide which constraint violations
halt compilation and which do not.

This is correct.

[Richard Heathfield]

Richard Weeks said:

My question, stated as simply as possible, is: must a constraint
violation, as defined by the standard, cause compilation to halt?

No. It does mean that a diagnostic message must be produced, but the
Standard does not require translation to be halted.

I believe that in the presence of a constraint violation, at
least one diagnostic is required, which, at the discretion of the
compiler writer, may be either an "error" or a "warning."

<shrugThat's just custom and usage. The Standard doesn't require either a
warning or an error. It only requires a diagnostic message.

Generally however, an error halts compilation and a warning does
not. This would seem to imply that it is up to the compiler
writer to decide which constraint violations halt compilation and
which do not.

It is entirely up to the implementation, yes - all it *has* to do is
diagnose at least one of any syntax errors and constraint violations it
encounters. The implementation is then free to choose whether to continue
or to halt translation.

--
Richard Heathfield <http://www.cpax.org.uk>
Email: -http://www. +rjh@
Google users: <http://www.cpax.org.uk/prg/writings/googly.php>
"Usenet is a strange place" - dmr 29 July 1999

[Keith Thompson]

Richard Weeks <rw****@nomail.netwrites:

My question, stated as simply as possible, is: must a constraint
violation, as defined by the standard, cause compilation to halt?

[...]

No, and if a compiler generates an executable (after producing the
required diagnostic), the behavior is undefined.

--
Keith Thompson (The_Other_Keith) ks***@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <* <http://users.sdsc.edu/~kst>
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"