To: J3                                                     J3/19-243
From: Malcolm Cohen
Subject: Failed image detection
Date: 2019-October-15
Reference: 19-181

1. Introduction

Paper 19-181 contains five purported interpretation requests glommed into
a single paper. This paper discusses the fourth of them.


2. The alleged problem

From 19-181:
    QUESTION 4:
    In the definition of the FAILED_IMAGES intrinsic function, 16.9.77 p5
    states:
        "If the executing image has previously executed an image control
         statement whose STAT= specifier assigned the value
         STAT_FAILED_IMAGE from the intrinsic module ISO_FORTRAN_ENV, or
         referenced a collective subroutine whose STAT argument was set to
         STAT_FAILED_IMAGE, at least one image in the set of images
         participating in that image control statement or collective
         subroutine reference shall be known to have failed."
    Should the list include the cases of a reference to the intrinsic
    IMAGE_STATUS that returned STAT_FAILED_IMAGE, an atomic subroutine
    with a STAT argument was set to STAT_FAILED_IMAGE, and an image
    selector with a STAT= specifier that was set to STAT_FAILED_IMAGE?

    ANSWER 4:
    Yes. An edit is provided.

    EDIT 4:
    In 16.9.77 FAILED_IMAGES ([TEAM, KIND]) p5, change
        "or referenced a collective subroutine whose STAT argument was set
         to STAT_FAILED_IMAGE, at least one image in the set of images
         participating in that image control statement or collective
         subroutine reference"
    to
        "referenced an atomic or a collective subroutine with a STAT
         argument that assigned the value STAT_FAILED_IMAGE, referenced an
         image selector with a STAT= specifier that assigned the value
         STAT_FAILED_IMAGE, or referenced the intrinsic IMAGE_STATUS with
         value STAT_FAILED_IMAGE, at least one image in the set of images
         involved in that execution or reference"
    so that the sentence reads:
        "If the executing image has previously executed an image control
         statement whose STAT= specifier assigned the value
         STAT_FAILED_IMAGE from the intrinsic module ISO_FORTRAN_ENV,
         referenced an atomic or a collective subroutine with a STAT
         argument that assigned the value STAT_FAILED_IMAGE, referenced an
         image selector with a STAT= specifier that assigned the value
         STAT_FAILED_IMAGE, or referenced the intrinsic IMAGE_STATUS with
         value STAT_FAILED_IMAGE, at least one image in the set of images
         involved in that execution or reference shall be known to have
         failed."


2. Response from /INTERP

Firstly, it is noted that the standard is currently unambiguous and not
contradictory.

Secondly, this kind of item is frequently left up to "Quality of
Implementation".

Thirdly, the situations in question are situations with precisely one image
that might have failed, not a set of images.

Therefore, at this time, /INTERP politely declines to accept this request
at this precise point in time.

Instead, /HPC is asked to consider whether, in their opinion, it is
acceptable to leave this up to Quality of Implementation, or whether the
standard should require this "good" behaviour.

If /HPC considers that this would be good as a correction to the current
edition of the standard, /INTERP will process it accordingly.  However,
I will note that /INTERP is highly likely to propose a different edit, i.e.
one that does not lead to a sentence longer than my arm.

If /HPC considers that it is not necessary to adopt the proposed changes
immediately, they are encouraged to consider whether it should be processed
as a normal "new feature".

===END===