[splint-discuss] How do I treat a symbol table?

[Terry Colligan]

  I'm new to splint, and can't seem to solve a number of
  problems.  I've been reading (and re-reading) the manual, but
  I can't quite see the big picture.

  One of my problems is how to treat a symbol table.  I need
  mostly to add malloc'd copies of strings to the symbol table,
  but the symbol table is initialized with some string literals.

  The code:

// test to show splint problem

// enough declarations to keep splint happy
typedef struct key key;
key *this_key;

// make entry in table
void add_entry(key *p,  // lookup key
    /*@null@*//*@keep@*/char *value)    // value for hey
    {
    // whatever, including making entry in symbol table structure
    // (enough code to solve not-used issues)
    if (p)
        if (value)
            ;
        else ;
    else if (value)
        ;
    }

// remember entry if appropriate
void remember(bool appropriate, // decision to add
              char *value)      // value to add
    {
    if (appropriate)
        add_entry(this_key, value);
    }


// make static entry
void initialize(void)
    {
    add_entry(this_key, "Built-in Symbol");
    }


  The result:
  
~/C$ splint -ifempty test.c
Splint 3.1.1 --- 21 Aug 2005

test.c: (in function remember)
test.c:26:3: Variable value is kept in true branch, but not kept in
                continuation.
  The state of a variable is different depending on which branch is taken. This
  means no annotation can sensibly be applied to the storage. (Use -branchstate
  to inhibit warning)
   test.c:26:3: in true branch:
   test.c:26:23: Storage value becomes kept
test.c: (in function initialize)
test.c:33:22: Unqualified static storage "Built-in Symbol" passed as keep
                 param: add_entry (..., "Built-in Symbol")
  Static storage is transferred in an inconsistent way. (Use -statictrans to
  inhibit warning)

Finished checking --- 2 code warnings


  I added the /*@keep@*/ annotation to add_entry() to resolve
  warnings where splint (correctly, it seems to me) was worried
  that storage was being lost.

  Do I misunderstand what /*@keep@*/ is for?

  Can somebody suggest an alternative?

  I don't really want to turn off 'branchstate' or 'statictrans'
  in general -- I want to solve this issue so I don't have to
  turn off major checking areas.

  Thanks for considering my problem.
-- 
Terry

Terry Colligan                         terry-splint at tenberry.com
Tenberry Software, Inc.                  http://www.tenberry.com
info at tenberry.com     phone 1.480.767.8868    fax 1.480.767.8709