arny@wayback.UUCP (04/13/87)
In article <12293281394.11.MENDAL@Sierra.Stanford.EDU>, Mendal@SIERRA.STANFORD.EDU (Geoff Mendal) writes: > Is the following package specification semantically valid? > > package Static_Test is > subtype Non_Static is Integer range Integer (1) .. 100; > > type Must_Be_Static is range Non_Static'Base'First .. 10; > end Static_Test; > > Read the ARM, 3.3.3(8..9) and 4.9(1..2, 8, 11). Forgive me if I missed something, but after checking the LRM sections you mentioned above (especially 4.9:11), I believe the subtype "Non_Static" is indeed static. This subtype is formed by imposing a static range constraint (1..100) on a scalar base type (Integer). (Note that enclosing the numeral 1 in parentheses does not make it non-static according to 4.9:2,4,10). It follows that the declaration of "Must_Be_Static" is then valid according to 4.9:8. Comments? Arny B. Engelson {ihnp4|bonnie|clyde}!wayback!arny (201) 386-4816 AT&T Bell Laboratories, Whippany, N.J.
deller@vrdxhq.UUCP (04/14/87)
In article <1132@wayback.UUCP>, arny@wayback.UUCP (Arny B. Engelson) writes: > > Is the following package specification semantically valid? > > package Static_Test is > > subtype Non_Static is Integer range Integer (1) .. 100; > > type Must_Be_Static is range Non_Static'Base'First .. 10; > > end Static_Test; > > Read the ARM, 3.3.3(8..9) and 4.9(1..2, 8, 11). > > ... I believe the subtype "Non_Static" is > indeed static. This subtype is formed by imposing a static range constraint > (1..100) on a scalar base type (Integer). (Note that enclosing the numeral 1 > in parentheses does not make it non-static according to 4.9:2,4,10). > . . . I believe you missed the fact that instead of "1", the expression is "Integer (1)", not simply "(1)". "Integer (1)" is a type conversion, which is not listed in RM 4.9 as a static primary. Using a "non-modifying" type conversion is one of the typical "tricks" used to make expressions non-static. Note that "Integer'(1)" is a qualified expression, which _is_ a static expression primitive, and is more likely to be used when "casting" (What, how did C sneak in here?) a literal to a specific type. Steven Deller, Verdix -- <end_message> ::= <disclaimer> | <joke> | <witty_saying> | <cute_graphic> {verdix,seismo,umcp-cs}!vrdxhq!deller