🚀 go-pugleaf

In article <BsJnKx.5yF@jrd.dec.com>, diamond@jit345.bad.jit.dec.com (Norman Diamond) writes:
|> 
|> I would guess that this one might have been addressed by an interpretation
|> ruling, if anyone might be kind enough finally to post the interpretation
|> rulings again.

No problem. Unfortunately, there are no rulings on 4.9.5.3.

May 30 22:22 1992  CIB#1 interpretations sorted by X3.159-1989 section  Page 1


RFI 33 Q  1  1.6       non-constraint "shall" violation doesn't need diagnostic
RFI 17 Q 35  2.1.1.2   where physical source lines occur in phase 1 matters not
RFI 17 Q  3  2.1.1.3   constraint violation always requires diagnostic
RFI 17 Q  2  2.1.2.2.1 if main does not exist behavior is implicitly undefined
RFI 17 Q 38  2.2.4.1   iteration/selection "control structure" means "statement"
RFI 17 Q 20  3.1.2     scope of macro identifiers is intentionally omitted
RFI 13 Q  5  3.1.2.1   enumeration tag is undefined until declaration complete
RFI 35 Q  2  3.1.2.1   parameter declarations have block scope
RFI 11 Q  2  3.1.2.2   linkage can be affected by scope
RFI 34 Q  2  3.1.2.2   incompatible declaration of object is undefined behavior
RFI 10 Q  1  3.1.2.5   typedef of an incomplete type is permitted
RFI 34 Q  1  3.1.2.5   size information is lost upon leaving scope
RFI 11 Q  1  3.1.2.6   composite type is not created across different scopes
RFI 13 Q  2  3.1.2.6   recursion does not preclude type compatibility
RFI 13 Q  3  3.1.2.6   composite type of enum and integer may not be unique
RFI 29 Q  1  3.1.2.6   tags are not used for type compatibility
RFI 17 Q  4  3.1.3.4   escape sequence value is not mapped
RFI 17 Q  5  3.1.3.4   constraint violation in example depends on 8-bit byte
RFI 17 Q 39  3.1.7     header name tokenization is context-dependent
RFI  3 Q  1  3.1.8     preprocessing number grammar is as intended
RFI 15 Q  1  3.2.1.1   bit-field treated as unsigned promotes as unsigned
RFI 28 Q  1  3.3       access rules also apply to dynamically allocated objects
RFI 12 Q  1  3.3.3.2   &*p is not valid when p has type void *
RFI 17 Q  6  3.3.3.2   storage class for object also applies to its parts
RFI 17 Q  7  3.3.3.4   type of result of sizeof is independent of size_t
RFI 17 Q 16  3.3.6     pointer arithmetic is bounded by specified array type
RFI 17 Q  8  3.3.9     void keyword versus void type can be resolved by context
RFI 17 Q  9  3.3.16.1  "assignment expression" refers to whole expression
RFI 31 Q  1  3.4       representability is required for constant subexpressions
RFI  7 Q  1  3.5       multiple declarations of tag are permitted
RFI 13 Q  4  3.5.2.3   structure is incomplete until content completely defined
RFI 17 Q 10  3.5.2.3   when size is needed is adequately specified
RFI 17 Q 11  3.5.2.3   multiple declaration of incomplete struct is allowed
RFI 16 Q  1  3.5.4     object may be declared without knowing its size
RFI  8 Q  2  3.5.4.3   volatile semantics are as intended
RFI  9 Q  1  3.5.4.3   typedef name takes precedence over ordinary identifier
RFI 13 Q  1  3.5.4.3   array parameter treated as pointer for composite type
RFI 17 Q 12  3.5.4.3   typedef name takes precedence over ordinary identifier
RFI 17 Q 13  3.5.4.3   storage class does not affect type compatibility
RFI 17 Q 14  3.5.4.3   void as parameter list is token, not type
RFI 17 Q 15  3.5.4.3   array parameter treated as pointer for composite type
RFI 17 Q 18  3.5.4.3   void as parameter list is not itself a parameter



















May 30 22:22 1992  CIB#1 interpretations sorted by X3.159-1989 section  Page 2


RFI 17 Q 36  3.5.4.3   qualified void function return type is undefined behavior
RFI 16 Q  2  3.5.7     implicit initialization of union is for first member
RFI 17 Q 17  3.5.7     explicit initialization of union skips unnamed members
RFI 17 Q 26  3.5.7     unnamed union members are not implicitly initialized
RFI  1 Q  1  3.6.6.4   function returns its value by copying
RFI  4 Q  1  3.7       can be at most one definition of unused extern identifier
RFI 20 Q  1  3.7       relaxed ref/def model can be standard conforming
RFI 35 Q  1  3.7.1     declared identifier is the one in the identifier list
RFI 11 Q  3  3.7.2     implicit initializer is effect, not token
RFI 11 Q  4  3.7.2     one implicit initializer supplied for incomplete extern
RFI 17 Q  1  3.8       first new-line terminates preprocessing directive
RFI  3 Q  2  3.8.3     macro replacement white space is as intended
RFI  3 Q  3  3.8.3     empty macro argument behavior is intentionally undefined
RFI  3 Q  4  3.8.3     preprocessor directive is forbidden in macro argument
RFI 17 Q 19  3.8.3     split-macro evaluation is intentionally fuzzy
RFI 17 Q 21  3.8.3     macro parentheses are those existing before replacement
RFI  2 Q  1  3.8.3.2   semantics of # are as intended
RFI 17 Q 22  3.8.3.3   ## produced by expansion is not the catenation operator
RFI 17 Q 23  3.8.3.4   split-macro non-replacement is intentionally fuzzy
RFI  5 Q  1  3.8.6     strictly conforming program cannot contain #pragma
RFI 17 Q 24  4.1.2     wording improvement is not necessary
RFI 17 Q 25  4.1.2.1   use of "must" in footnote is clear enough
RFI 19 Q  1  4.3       "printing character" is implementation-defined
RFI 25 Q  1  4.5.1     principles for C floating-point representation are given
RFI 30 Q  1  4.5.1     sin(DBL_MAX) must not set errno to EDOM
RFI 14 Q  1  4.6       setjmp can be either a macro or a function
RFI  8 Q  1  4.6.2.1   dead-store elimination is allowed when setjmp is used
RFI 17 Q 27  4.9.6.1   # flag takes precedence over %g in fprintf conversion
RFI 21 Q  1  4.9.6.1   precision is increased only if necessary
RFI 14 Q  2  4.9.6.2   %n is valid after fscanf encounters EOF
RFI 17 Q 29  4.9.6.2   fscanf matching failure can result in lost characters
RFI 18 Q  1  4.9.6.2   plain "characters" refers to one-byte characters
RFI 17 Q 28  4.9.9     no temporal ordering is implied for operations on return
RFI 17 Q 30  4.9.9     seek offset is obtained from successful previous call
RFI 22 Q  1  4.10.1.4  strtod and fscanf differ on conversion failure
RFI 23 Q  1  4.10.1.4  "0.0e99999" must be properly converted to zero
RFI 24 Q  1  4.10.1.4  LC_NUMERIC locale category affects strtod
RFI  6 Q  1  4.10.1.6  strtoul honors minus sign beginning subject sequence
RFI 17 Q 31  4.10.3    size of object is always measured in bytes
RFI 17 Q 32  4.11.4    string comparison interprets bytes as if unsigned char
RFI 17 Q 33  4.11.4    string terminator takes part in comparison
RFI 17 Q 34  4.11.5.8  description of calls to strtok is clear enough



-- 
==================
Chris Volpe
G.E. Corporate R&D
volpecr@crd.ge.com