xref: /original-bsd/lib/libc/db/hash/README (revision 73949c1b) 
1#	@(#)README	5.3 (Berkeley) 02/22/91
2
3This package implements a superset of the hsearch and dbm/ndbm libraries.
4
5Contents:
6
7Hashing Package:
8	dynahash.c
9	page.c
10	buf.c
11	big.c
12	hfunc.c
13	log2.c
14
15	hash.h
16	page.h
17	db.h
18
19Backward Compatibility Routines:
20	ndbm.c
21	ndbm.h
22	hsearch.c
23	search.h
24
25Misc:
26	byte_order.c
27
28Compatibility routines:
29	mkstemp.c
30
31	ansi.h
32	bsd.h
33	cdefs.h
34	endian.h
35	posix.h
36	unistd.h
37
38DIFFS:
39	These are diffs since the date of the file (i.e. a file labeled
40	DIFFS.2.12 are the diffs since the 2.12 version on arpa).  The
41	date of the DIFF file indicates when those diffs were installed.
42
43Test Programs:
44	All test programs which need key/data pairs expect them entered
45	with key and data on separate lines
46
47	tcreat3.c
48		Takes
49			bucketsize (bsize),
50			fill factor (ffactor), and
51			initial number of elements (nelem).
52		Creates a hash table named hashtest containing the
53		keys/data pairs entered from standard in.
54	thash4.c
55		Takes
56			bucketsize (bsize),
57			fill factor (ffactor),
58			initial number of elements (nelem)
59			bytes of cache (ncached), and
60			file from which to read data  (fname)
61		Creates a table from the key/data pairs on standard in and
62		then does a read of each key/data in fname
63	tdel.c
64		Takes
65			bucketsize (bsize), and
66			fill factor (ffactor).
67			file from which to read data (fname)
68		Reads each key/data pair from fname and deletes the
69		key from the hash table hashtest
70	tseq.c
71		Reads the key/data pairs in the file hashtest and writes them
72		to standard out.
73	tread2.c
74		Takes
75			butes of cache (ncached).
76		Reads key/data pairs from standard in and looks them up
77		in the file hashtest.
78	tverify.c
79		Reads key/data pairs from standard in, looks them up
80		in the file hashtest, and verifies that the data is
81		correct.
82
83NOTES:
84
85If you are not running a 4.3BSD-Reno or later system, you may need to use
86some of the compatibility files provided.  The files are as follows:
87
88	mkstemp.c	Mkstemp/mktemp library routine.
89
90	ansi.h		Map bcopy and friends to memcpy and friends.
91	bsd.h		Map various new BSD things to old things.
92	cdefs.h		Handle the function prototypes in other include files.
93	endian.h	Handle byte ordering.  Be sure to set BYTE_ORDER in
94			endian.h appropriately for your machine.  If you don't
95			know what "endian" your machine is, compile
96			byte_order.c and run it.  It should tell you.
97	posix.h		Map various POSIX 1003.1 things to old-style things.
98	unistd.h	POSIX 1003.1 definitions.
99
100If you are not running on the current BSD release (4.3BSD-Reno+), you will
101need to include bsd.h in hash.h.  Depending on what system you are running
102on, you will need to add the other compatibility h files in hash.h.
103
104The file search.h is provided for using the hsearch compatible interface
105on BSD systems.  On System V derived systems, search.h should appear in
106/usr/include.
107
108The man page db.3 explains the interface to the hashing system.
109The file hash.ps is a postscript copy of a paper explaining
110the history, implementation, and performance of the hash package.
111
112"bugs" or idiosyncracies
113
114If you have a lot of overflows, it is possible to run out of overflow
115pages.  Currently, this will cause a message to be printed on stderr.
116Eventually, this will be indicated by a return error code.
117
118If you are using the ndbm interface and exit without flushing or closing the
119file, you may lose updates since the package buffers all writes.  Also,
120the db interface only creates a single database file.  To avoid overwriting
121the user's original file, the suffix ".db" is appended to the file name
122passed to dbm_open.  Additionally, if your code "knows" about the historic
123.dir and .pag files, it will break.
124
125There is a fundamental difference between this package and the old hsearch.
126Hsearch requires the user to maintain the keys and data in the application's
127allocated memory while hash takes care of all storage management.  The down
128side is that the byte strings passed in the ENTRY structure must be null
129terminated (both the keys and the data).
130