Judy1(3X) Judy1(3X)
NAME
Judy1 macros - C library for creating and accessing a dynamic array of bits, using any value of a word as an index.
SYNOPSIS
cc [flags] sourcefiles -lJudy
#include <Judy.h>
int Rc_int; // return code - integer
Word_t Rc_word; // return code - unsigned word
Word_t Index, Index1, Index2, Nth;
Pvoid_t PJ1Array = (Pvoid_t) NULL; // initialize Judy1 array
J1S( Rc_int, PJ1Array, Index); // Judy1Set()
J1U( Rc_int, PJ1Array, Index); // Judy1Unset()
J1T( Rc_int, PJ1Array, Index); // Judy1Test()
J1C( Rc_word, PJ1Array, Index1, Index2); // Judy1Count()
J1BC(Rc_int, PJ1Array, Nth, Index); // Judy1ByCount()
J1FA(Rc_word, PJ1Array); // Judy1FreeArray()
J1MU(Rc_word, PJ1Array); // Judy1MemUsed()
J1F( Rc_int, PJ1Array, Index); // Judy1First()
J1N( Rc_int, PJ1Array, Index); // Judy1Next()
J1L( Rc_int, PJ1Array, Index); // Judy1Last()
J1P( Rc_int, PJ1Array, Index); // Judy1Prev()
J1FE(Rc_int, PJ1Array, Index); // Judy1FirstEmpty()
J1NE(Rc_int, PJ1Array, Index); // Judy1NextEmpty()
J1LE(Rc_int, PJ1Array, Index); // Judy1LastEmpty()
J1PE(Rc_int, PJ1Array, Index); // Judy1PrevEmpty()
DESCRIPTION
A Judy1 array is the equivalent of a bit array or bit map. A bit is addressed by an Index (key). The array may be sparse, and the Index
may be any word-sized value. If an index is present, it represents a set bit (a bit set represents an index present). If an index is
absent, it represents an unset bit (a bit unset represents an absent index).
A Judy1 array is allocated with a NULL pointer
Pvoid_t PJ1Array = (Pvoid_t) NULL;
Memory to support the array is allocated as bits are set, and released as bits are unset. If the Judy1 pointer (PJ1Array) is NULL, all
bits are unset (and the Judy1 array requires no memory).
As with an ordinary array, a Judy1 array contains no duplicate indexes.
Using the macros described here, rather than the Judy1 function calls, the default error handling sends a message to the standard error and
terminates the program with exit(1). For other error handling methods, see the ERRORS section.
Because the macro forms are faster and have a simpler error handling interface than the equivalent functions, they are the preferred way of
calling the Judy1 functions.
Set Index's bit in the Judy1 array PJ1Array.
Return Rc_int set to 1 if Index's bit was previously unset (successful), otherwise 0 if the bit was already set (unsuccess-
ful).
Unset Index's bit in the Judy1 array PJ1Array;
that is, remove Index from the Judy1 array.
Return Rc_int set to 1 if Index's bit was previously set (successful), otherwise 0 if the bit was already unset (unsuccess-
ful).
Test if Index's bit is set in the
Judy1 array PJ1Array.
Return Rc_int set to 1 if Index's bit is set (Index is present), 0 if it is unset (Index is absent).
Count the number of indexes present in the Judy1 array
PJ1Array between Index1 and Index2 (inclusive).
Return Rc_word set to the count. A return value of 0 can be valid as a count, or it can indicate a special case for fully
populated array (32-bit machines only). See Judy1Count() for ways to resolve this.
To count all indexes present in a Judy1 bit array, use:
J1C(Rc_word, PJ1Array, 0, -1);
Note: The -1 promotes to the maximum index, that is, all ones.
Locate the Nth index that is present in the Judy1 array
PJ1Array (Nth = 1 returns the first index present). To refer to the last index in a fully populated array (all indexes
present, which is rare), use Nth = 0.
Return Rc_int set to 1 and Index set to the Nth index if found, otherwise return Rc_int set to 0 (the value of Index con-
tains no useful information).
Free the entire Judy1 array PJ1Array (much faster than using a
J1N(), J1U() loop).
Return Rc_word set to the number of bytes freed, and PJ1Array set to NULL.
Return Rc_word set to the number of bytes of memory currently in use by
Judy1 array PJ1Array. This is a very fast routine, and may be used after a J1S() or J1U() call with little performance
impact.
The Judy1 search functions allow you to search for set or unset bits in the array.
You may search inclusively or exclusively, in either forward or reverse directions. All of the search functions use a simi-
lar calling sequence. Rc_int is returned set to 1 for a successful search and the found Index is returned. Rc_int is
returned set to 0 for an unsuccessful search, and Index contains no useful information. The return code Rc_int must be
checked prior to using the returned Index, since a search failure is possible.
Search (inclusive) for the first index present that is equal
to or greater than the passed Index. (Start with Index = 0 to find the first index in the array.) J1F() is typically used
to begin a sorted-order scan of the indexes present in a Judy1 array.
Search (exclusive) for the next index present that is
greater than the passed Index. J1N() is typically used to continue a sorted-order scan of the indexes present in a Judy1
array, or to locate a "neighbor" of a given index.
Search (inclusive) for the last index present that is equal
to or less than the passed Index. (Start with Index = -1, that is, all ones, to find the last index in the array.) J1L()
is typically used to begin a reverse-sorted-order scan of the indexes present in a Judy1 array.
Search (exclusive) for the previous index present that is
less than the passed Index. J1P() is typically used to continue a reverse-sorted-order scan of the indexes present in a
Judy1 array, or to locate a "neighbor" of a given index.
Search (inclusive) for the first absent index that is equal to
or greater than the passed Index. (Start with Index = 0 to find the first index absent in the array.)
Search (exclusive) for the next absent index that is
greater than the passed Index.
Search (inclusive) for the last absent index that is
equal to or less than the passed Index. (Start with Index = -1 to find the last index absent in the array.)
Search (exclusive) for the previous absent index that is
less than the passed Index.
ERRORS
There are two categories of Judy error returns:
1) User programming errors (bugs) such as memory corruption or invalid pointers.
2) Out-of-memory (malloc(3) failure) when modifying a Judy1 array with J1S() or J1U().
There are three methods of handling errors when using the macros:
1) Default error handling.
2) User-specified JUDYERROR() macro.
3) Disable macro error handling.
The default is to print error messages to stderr, for example:
File 'YourCfile.c', line 1234: Judy1Set(), JU_ERRNO_* == 2, ID == 321
This indicates that an error occurred in a J1S() call at line 1234 in 'YourCfile.c'. JU_ERRNO_* == 2 is JU_ERRNO_NOMEM (as
defined in the Judy.h file). The ID number indicates the Judy source line number where the error was detected.
Your program then terminates with an exit(1).
The JUDYERROR() macro provides flexibility for handling error returns
as needed to suit your program while still accessing Judy1 arrays using macros instead of function calls. You can modify
JUDYERROR() to distinguish between the two types of errors (described above), and explicitly test for the remaining
JU_ERRNO_NOMEM errors possible in your program.
// This is an example of Judy1 macro API to continue when out of memory.
#ifndef JUDYERROR_NOTEST
#include <stdio.h>
// This is the macro that the Judy macro APIs use for return codes of -1:
#define JUDYERROR(CallerFile, CallerLine, JudyFunc, JudyErrno, JudyErrID)
{
if ((JudyErrno) != JU_ERRNO_NOMEM) /* ! a malloc() failure */
{
(void) fprintf(stderr, "File '%s', line %d: %s(), "
"JU_ERRNO_* == %d, ID == %d
",
CallerFile, CallerLine,
JudyFunc, JudyErrno, JudyErrID);
exit(1);
}
}
#endif // JUDYERROR_NOTEST not defined
This error handling macro must be included before the #include <Judy.h> statement in your program.
When your program is "bug free",
the only errors occurring should be malloc(3) errors. You can turn off error handling included in the Judy1 macros by using
#define JUDYERROR_NOTEST 1
(in your program code), or
cc -DJUDYERROR_NOTEST sourcefile -lJudy
(on your command line).
EXAMPLE
In the following example, errors in the J1S() or J1U() calls go to a user-defined procedure, process_malloc_failure. This is not needed
when you use the default JUDYERROR() macro, since the default causes your program to exit on all failures, including malloc(3) failure.
#include <stdio.h>
#include <Judy.h>
int main() // Example program of Judy1 macro APIs
{
Word_t Index; // index (or key)
Word_t Rcount; // count of indexes (or bits set)
Word_t Rc_word; // full word return value
int Rc_int; // boolean values returned (0 or 1)
Pvoid_t PJ1Array = (Pvoid_t) NULL; // initialize Judy1 array
Index = 123456;
J1S(Rc_int, J1Array, Index); // set bit at 123456
if (Rc_int == JERR) goto process_malloc_failure;
if (Rc_int == 1) printf("OK - bit successfully set at %lu
", Index);
if (Rc_int == 0) printf("BUG - bit already set at %lu
", Index);
Index = 654321;
J1T(Rc_int, J1Array, Index); // test if bit set at 654321
if (Rc_int == 1) printf("BUG - set bit at %lu
", Index);
if (Rc_int == 0) printf("OK - bit not set at %lu
", Index);
J1C(Rcount, J1Array, 0, -1); // count all bits set in array
printf("%lu bits set in Judy1 array
", Rcount);
Index = 0;
J1F(Rc_int, J1Array, Index); // find first bit set in array
if (Rc_int == 1) printf("OK - first bit set is at %lu
", Index);
if (Rc_int == 0) printf("BUG - no bits set in array
");
J1MU(Rc_word, J1Array); // how much memory was used?
printf("%lu Indexes used %lu bytes of memory
", Rcount, Rc_word);
Index = 123456;
J1U(Rc_int, J1Array, Index); // unset bit at 123456
if (Rc_int == JERR) goto process_malloc_failure;
if (Rc_int == 1) printf("OK - bit successfully unset at %lu
", Index);
if (Rc_int == 0) printf("BUG - bit was not set at %lu
", Index);
return(0);
}
AUTHOR
Judy was invented and implemented by Hewlett-Packard.
SEE ALSO
Judy(3X), Judy1_funcs(3X), JudyL(3X), JudyL_funcs(3X), JudySL(3X), JudySL_funcs(3X),
malloc(3),
the Judy website, http://www.hp.com/go/judy/, for more information and Application Notes.
Judy1(3X)