J3/01-376R2 Date: December 6, 2001 To: j3 From: Subgroup C Subject: Chapter 6 comments, updated from meeting 158 This is an update of 282/R2 from meeting 158. I deleted the items that were taken care of at the last meeting to make it easier for the B team to process. I kept the same numbering scheme as in 282 so that notes from last meeting should still be valid. Page, etc., numbers have been updated to the current 007/R4 PDF version The edit has been changed from 01-376 by subgroup C. 10) Page 110[11-12], 6.3.3.2, 2nd paragraph, second sentence. Seems to restate the first, or how can a portion be independent of any other portion? The R1 version of this paper gave rise to a need for clarification. The following example was given: TYPE T REAL, POINTER :: P(:) END TYPE T TYPE(T), ALLOCATABLE:: A REAL, POINTER:: Q(:) ALLOCATE(A(10)) ALLOCATE(A(3)%P(9)) Q => A(3)%P DEALLOCATE(Q) ! is this legal While A(3)%P is a subjobject of A, the target of A(3)%P is not a subobject of A, and since Q points to the whole of the target of A(3)%P the deallocate statement is legal. Edit: in 007R4 [110:10-12] Replace first two sentences of the paragraph with "A pointer shall not be deallocated if it is currently associated with a subobject of an allocated target that is not the whole of that target."