gperf
2.7.2
gperf
Utility
gperf
gperf
gperf
gperf
gperf
Version 2, June 1991
Copyright (C) 1989, 1991 Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA. Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed.
The licenses for most software are designed to take away your freedom to share and change it. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change free software--to make sure the software is free for all its users. This General Public License applies to most of the Free Software Foundation's software and to any other program whose authors commit to using it. (Some other Free Software Foundation software is covered by the GNU Library General Public License instead.) You can apply it to your programs, too.
When we speak of free software, we are referring to freedom, not price. Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for this service if you wish), that you receive source code or can get it if you want it, that you can change the software or use pieces of it in new free programs; and that you know you can do these things.
To protect your rights, we need to make restrictions that forbid anyone to deny you these rights or to ask you to surrender the rights. These restrictions translate to certain responsibilities for you if you distribute copies of the software, or if you modify it.
For example, if you distribute copies of such a program, whether gratis or for a fee, you must give the recipients all the rights that you have. You must make sure that they, too, receive or can get the source code. And you must show them these terms so they know their rights.
We protect your rights with two steps: (1) copyright the software, and (2) offer you this license which gives you legal permission to copy, distribute and/or modify the software.
Also, for each author's protection and ours, we want to make certain that everyone understands that there is no warranty for this free software. If the software is modified by someone else and passed on, we want its recipients to know that what they have is not the original, so that any problems introduced by others will not reflect on the original authors' reputations.
Finally, any free program is threatened constantly by software patents. We wish to avoid the danger that redistributors of a free program will individually obtain patent licenses, in effect making the program proprietary. To prevent this, we have made it clear that any patent must be licensed for everyone's free use or not licensed at all.
The precise terms and conditions for copying, distribution and modification follow.
TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
END OF TERMS AND CONDITIONS
If you develop a new program, and you want it to be of the greatest possible use to the public, the best way to achieve this is to make it free software which everyone can redistribute and change under these terms.
To do so, attach the following notices to the program. It is safest to attach them to the start of each source file to most effectively convey the exclusion of warranty; and each file should have at least the "copyright" line and a pointer to where the full notice is found.
one line to give the program's name and an idea of what it does. Copyright (C) year name of author This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA.
Also add information on how to contact you by electronic and paper mail.
If the program is interactive, make it output a short notice like this when it starts in an interactive mode:
Gnomovision version 69, Copyright (C) year name of author Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'. This is free software, and you are welcome to redistribute it under certain conditions; type `show c' for details.
The hypothetical commands `show w' and `show c' should show the appropriate parts of the General Public License. Of course, the commands you use may be called something other than `show w' and `show c'; they could even be mouse-clicks or menu items--whatever suits your program.
You should also get your employer (if you work as a programmer) or your school, if any, to sign a "copyright disclaimer" for the program, if necessary. Here is a sample; alter the names:
Yoyodyne, Inc., hereby disclaims all copyright interest in the program `Gnomovision' (which makes passes at compilers) written by James Hacker. signature of Ty Coon, 1 April 1989 Ty Coon, President of Vice
This General Public License does not permit incorporating your program into proprietary programs. If your program is a subroutine library, you may consider it more useful to permit linking proprietary applications with the library. If this is what you want to do, use the GNU Library General Public License instead of this License.
gperf
Utilitygperf
perfect hash function generator utility was
originally written in GNU C++ by Douglas C. Schmidt. It is now also
available in a highly-portable "old-style" C version. The general
idea for the perfect hash function generator was inspired by Keith
Bostic's algorithm written in C, and distributed to net.sources around
1984. The current program is a heavily modified, enhanced, and extended
implementation of Keith's basic idea, created at the University of
California, Irvine. Bugs, patches, and suggestions should be reported
to both <bug-gnu-utils@gnu.org>
and
<gperf-bugs@lists.sourceforge.net>
.
gperf
.
gperf
is a perfect hash function generator written in C++. It
transforms an n element user-specified keyword set W into a
perfect hash function F. F uniquely maps keywords in
W onto the range 0..k, where k >= n. If k
= n then F is a minimal perfect hash function.
gperf
generates a 0..k element static lookup table and a
pair of C functions. These functions determine whether a given
character string s occurs in W, using at most one probe into
the lookup table.
gperf
currently generates the reserved keyword recognizer for
lexical analyzers in several production and research compilers and
language processing tools, including GNU C, GNU C++, GNU Pascal, GNU
Modula 3, and GNU indent. Complete C++ source code for gperf
is
available via anonymous ftp from ftp://ftp.gnu.org/pub/gnu/gperf/
.
A paper describing gperf
's design and implementation in greater
detail is available in the Second USENIX C++ Conference proceedings.
gperf
A static search structure is an Abstract Data Type with certain
fundamental operations, e.g., initialize, insert,
and retrieve. Conceptually, all insertions occur before any
retrievals. In practice, gperf
generates a static
array
containing search set keywords and any associated attributes specified
by the user. Thus, there is essentially no execution-time cost for the
insertions. It is a useful data structure for representing static
search sets. Static search sets occur frequently in software system
applications. Typical static search sets include compiler reserved
words, assembler instruction opcodes, and built-in shell interpreter
commands. Search set members, called keywords, are inserted into
the structure only once, usually during program initialization, and are
not generally modified at run-time.
Numerous static search structure implementations exist, e.g., arrays, linked lists, binary search trees, digital search tries, and hash tables. Different approaches offer trade-offs between space utilization and search time efficiency. For example, an n element sorted array is space efficient, though the average-case time complexity for retrieval operations using binary search is proportional to log n. Conversely, hash table implementations often locate a table entry in constant time, but typically impose additional memory overhead and exhibit poor worst case performance.
Minimal perfect hash functions provide an optimal solution for a particular class of static search sets. A minimal perfect hash function is defined by two properties:
For most applications it is far easier to generate perfect hash
functions than minimal perfect hash functions. Moreover,
non-minimal perfect hash functions frequently execute faster than
minimal ones in practice. This phenomena occurs since searching a
sparse keyword table increases the probability of locating a "null"
entry, thereby reducing string comparisons. gperf
's default
behavior generates near-minimal perfect hash functions for
keyword sets. However, gperf
provides many options that permit
user control over the degree of minimality and perfection.
Static search sets often exhibit relative stability over time. For
example, Ada's 63 reserved words have remained constant for nearly a
decade. It is therefore frequently worthwhile to expend concerted
effort building an optimal search structure once, if it
subsequently receives heavy use multiple times. gperf
removes
the drudgery associated with constructing time- and space-efficient
search structures by hand. It has proven a useful and practical tool
for serious programming projects. Output from gperf
is currently
used in several production and research compilers, including GNU C, GNU
C++, GNU Pascal, and GNU Modula 3. The latter two compilers are not yet
part of the official GNU distribution. Each compiler utilizes
gperf
to automatically generate static search structures that
efficiently identify their respective reserved keywords.
gperf
The perfect hash function generator gperf
reads a set of
"keywords" from a keyfile (or from the standard input by
default). It attempts to derive a perfect hashing function that
recognizes a member of the static keyword set with at most a
single probe into the lookup table. If gperf
succeeds in
generating such a function it produces a pair of C source code routines
that perform hashing and table lookup recognition. All generated C code
is directed to the standard output. Command-line options described
below allow you to modify the input and output format to gperf
.
By default, gperf
attempts to produce time-efficient code, with
less emphasis on efficient space utilization. However, several options
exist that permit trading-off execution time for storage space and vice
versa. In particular, expanding the generated table size produces a
sparse search structure, generally yielding faster searches.
Conversely, you can direct gperf
to utilize a C switch
statement scheme that minimizes data space storage size. Furthermore,
using a C switch
may actually speed up the keyword retrieval time
somewhat. Actual results depend on your C compiler, of course.
In general, gperf
assigns values to the characters it is using
for hashing until some set of values gives each keyword a unique value.
A helpful heuristic is that the larger the hash value range, the easier
it is for gperf
to find and generate a perfect hash function.
Experimentation is the key to getting the most from gperf
.
gperf
You can control the input keyfile format by varying certain command-line
arguments, in particular the `-t' option. The input's appearance
is similar to GNU utilities flex
and bison
(or UNIX
utilities lex
and yacc
). Here's an outline of the general
format:
declarations %% keywords %% functions
Unlike flex
or bison
, all sections of
gperf
's input are optional. The following sections describe the
input format for each section.
struct
Declarations and C Code Inclusion
The keyword input file optionally contains a section for including
arbitrary C declarations and definitions, as well as provisions for
providing a user-supplied struct
. If the `-t' option
is enabled, you must provide a C struct
as the last
component in the declaration section from the keyfile file. The first
field in this struct must be a char *
or const char *
identifier called `name', although it is possible to modify this
field's name with the `-K' option described below.
Here is a simple example, using months of the year and their attributes as input:
struct months { char *name; int number; int days; int leap_days; }; %% january, 1, 31, 31 february, 2, 28, 29 march, 3, 31, 31 april, 4, 30, 30 may, 5, 31, 31 june, 6, 30, 30 july, 7, 31, 31 august, 8, 31, 31 september, 9, 30, 30 october, 10, 31, 31 november, 11, 30, 30 december, 12, 31, 31
Separating the struct
declaration from the list of keywords and
other fields are a pair of consecutive percent signs, `%%',
appearing left justified in the first column, as in the UNIX utility
lex
.
Using a syntax similar to GNU utilities flex
and bison
, it
is possible to directly include C source text and comments verbatim into
the generated output file. This is accomplished by enclosing the region
inside left-justified surrounding `%{', `%}' pairs. Here is
an input fragment based on the previous example that illustrates this
feature:
%{ #include <assert.h> /* This section of code is inserted directly into the output. */ int return_month_days (struct months *months, int is_leap_year); %} struct months { char *name; int number; int days; int leap_days; }; %% january, 1, 31, 31 february, 2, 28, 29 march, 3, 31, 31 ...
It is possible to omit the declaration section entirely. In this case the keyfile begins directly with the first keyword line, e.g.:
january, 1, 31, 31 february, 2, 28, 29 march, 3, 31, 31 april, 4, 30, 30 ...
The second keyfile format section contains lines of keywords and any associated attributes you might supply. A line beginning with `#' in the first column is considered a comment. Everything following the `#' is ignored, up to and including the following newline.
The first field of each non-comment line is always the key itself. It
can be given in two ways: as a simple name, i.e., without surrounding
string quotation marks, or as a string enclosed in double-quotes, in
C syntax, possibly with backslash escapes like \"
or \234
or \xa8
. In either case, it must start right at the beginning
of the line, without leading whitespace.
In this context, a "field" is considered to extend up to, but
not include, the first blank, comma, or newline. Here is a simple
example taken from a partial list of C reserved words:
# These are a few C reserved words, see the c.gperf file # for a complete list of ANSI C reserved words. unsigned sizeof switch signed if default for while return
Note that unlike flex
or bison
the first `%%' marker
may be elided if the declaration section is empty.
Additional fields may optionally follow the leading keyword. Fields
should be separated by commas, and terminate at the end of line. What
these fields mean is entirely up to you; they are used to initialize the
elements of the user-defined struct
provided by you in the
declaration section. If the `-t' option is not enabled
these fields are simply ignored. All previous examples except the last
one contain keyword attributes.
The optional third section also corresponds closely with conventions
found in flex
and bison
. All text in this section,
starting at the final `%%' and extending to the end of the input
file, is included verbatim into the generated output file. Naturally,
it is your responsibility to ensure that the code contained in this
section is valid C.
gperf
Several options control how the generated C code appears on the standard
output. Two C function are generated. They are called hash
and
in_word_set
, although you may modify their names with a command-line
option. Both functions require two arguments, a string, char *
str, and a length parameter, int
len. Their default
function prototypes are as follows:
hash
function returns an integer value
created by adding len to several user-specified str key
positions indexed into an associated values table stored in a
local static array. The associated values table is constructed
internally by gperf
and later output as a static local C array
called `hash_table'; its meaning and properties are described below
(see section 7 Implementation Details of GNU gperf
). The relevant key positions are specified via
the `-k' option when running gperf
, as detailed in the
Options section below(see section 4 Invoking gperf
).
NULL
.
If the option `-c' is not used, str must be a NUL terminated string of exactly length len. If `-c' is used, str must simply be an array of len characters and does not need to be NUL terminated.
The code generated for these two functions is affected by the following options:
struct
.
switch
statement rather than use a large,
(and potentially sparse) static array. Although the exact time and
space savings of this approach vary according to your C compiler's
degree of optimization, this method often results in smaller and faster
code.
If the `-t' and `-S' options are omitted, the default action
is to generate a char *
array containing the keys, together with
additional null strings used for padding the array. By experimenting
with the various input and output options, and timing the resulting C
code, you can determine the best option choices for different keyword
set characteristics.
By default, the code generated by gperf
operates on zero
terminated strings, the usual representation of strings in C. This means
that the keywords in the input file must not contain NUL characters,
and the str argument passed to hash
or in_word_set
must be NUL terminated and have exactly length len.
If option `-c' is used, then the str argument does not need
to be NUL terminated. The code generated by gperf
will only
access the first len, not len+1, bytes starting at str.
However, the keywords in the input file still must not contain NUL
characters.
If option `-l' is used, then the hash table performs binary
comparison. The keywords in the input file may contain NUL characters,
written in string syntax as \000
or \x00
, and the code
generated by gperf
will treat NUL like any other character.
Also, in this case the `-c' option is ignored.
gperf
There are many options to gperf
. They were added to make
the program more convenient for use with real applications. "On-line"
help is readily available via the `-h' option. Here is the
complete list of options.
struct
type declaration for generated
code. Any text before a pair of consecutive `%%' is considered
part of the type declaration. Keywords and additional fields may follow
this, one group of fields per line. A set of examples for generating
perfect hash tables and functions for Ada, C, C++, Pascal, Modula 2,
Modula 3 and JavaScript reserved words are distributed with this release.
gperf
to generate code in the language specified by the
option's argument. Languages handled are currently:
#define const
to empty
for compilers which don't know about this keyword.
gperf
. It does not do anything.
gperf
. It does not do anything.
struct
.
Perfect_Hash
.
isalnum
and isgraph
do
not guarantee that a character is in this range. Only an explicit
test like `c >= 'A' && c <= 'Z'' guarantees this.) This was the
default in versions of gperf
earlier than 2.7; now the default is
to assume 8-bit characters.
strncmp
function to perform
string comparisons. The default action is to use strcmp
.
<jjc@ai.mit.edu>
.
<string.h>
, at the
beginning of the code. By default, this is not done; the user must
include this header file himself to allow compilation of the code.
switch
statement scheme,
rather than an array lookup table. This can lead to a reduction in both
time and space requirements for some keyfiles. The argument to this
option determines how many switch
statements are generated. A
value of 1 generates 1 switch
containing all the elements, a
value of 2 generates 2 tables with 1/2 the elements in each
switch
, etc. This is useful since many C compilers cannot
correctly generate code for large switch
statements. This option
was inspired in part by Keith Bostic's original C program.
gperf
. It does not do anything.
gperf
strcmp
.
However, using `-l' might greatly increase the size of the
generated C code if the lookup table range is large (which implies that
the switch option `-S' is not enabled), since the length table
contains as many elements as there are entries in the lookup table.
This option is mandatory for binary comparisons (see section 3.3 Use of NUL characters).
gperf
does not backtrack it is possible for it to process
all your input keywords without finding a unique mapping for each word.
However, frequently only a very small number of duplicates occur, and
the majority of keys still require one probe into the table.
gperf
treats all these keys as
part of an equivalence class and generates a perfect hash function with
multiple comparisons for duplicate keys. It is up to you to completely
disambiguate the keywords by modifying the generated C code. However,
gperf
helps you out by organizing the output.
gperf
to work on
keyword sets that it otherwise could not handle.
gperf
's running time at the cost of minimizing generated
table-size. The iteration amount represents the number of times to
iterate when resolving a collision. `0' means iterate by the number of
keywords. This option is probably most useful when used in conjunction
with options `-D' and/or `-S' for large keyword sets.
gperf
jumps by random amounts.
gperf
's execution time, since collisions will
begin earlier and continue throughout the remainder of keyword
processing. See Cichelli's paper from the January 1980 Communications
of the ACM for details.
gperf
has difficultly with a certain keyword set try using
`-r' or `-D'.
gperf
's runtime, since it must search through a much larger range
of values. Judicious use of the `-f' option helps alleviate this
overhead, however.
gperf
is executing. It is useful both for
maintaining the program and for determining whether a given set of
options is actually speeding up the search for a solution. Some useful
information is dumped at the end of the program when the `-d'
option is enabled.
gperf
The following are some limitations with the current release of
gperf
:
gperf
utility is tuned to execute quickly, and works quickly
for small to medium size data sets (around 1000 keywords). It is
extremely useful for maintaining perfect hash functions for compiler
keyword sets. Several recent enhancements now enable gperf
to
work efficiently on much larger keyword sets (over 15,000 keywords).
When processing large keyword sets it helps greatly to have over 8 megs
of RAM.
However, since gperf
does not backtrack no guaranteed solution
occurs on every run. On the other hand, it is usually easy to obtain a
solution by varying the option parameters. In particular, try the
`-r' option, and also try changing the default arguments to the
`-s' and `-j' options. To guarantee a solution, use
the `-D' and `-S' options, although the final results are not
likely to be a perfect hash function anymore! Finally, use the
`-f' option if you want gperf
to generate the perfect hash
function fast, with less emphasis on making it minimal.
It should be "relatively" easy to replace the current perfect hash function algorithm with a more exhaustive approach; the perfect hash module is essential independent from other program modules. Additional worthwhile improvements include:
switch
option can minimize the data size, at the expense of slightly longer
lookups (note that the gcc compiler generally produces good code for
switch
statements, reducing the need for more complex schemes).
gperf
A paper describing the high-level description of the data structures and
algorithms used to implement gperf
will soon be available. This
paper is useful not only from a maintenance and enhancement perspective,
but also because they demonstrate several clever and useful programming
techniques, e.g., `Iteration Number' boolean arrays, double
hashing, a "safe" and efficient method for reading arbitrarily long
input from a file, and a provably optimal algorithm for simultaneously
determining both the minimum and maximum elements in a list.
[1] Chang, C.C.: A Scheme for Constructing Ordered Minimal Perfect Hashing Functions Information Sciences 39(1986), 187-195. [2] Cichelli, Richard J. Author's Response to "On Cichelli's Minimal Perfect Hash Functions Method" Communications of the ACM, 23, 12(December 1980), 729. [3] Cichelli, Richard J. Minimal Perfect Hash Functions Made Simple Communications of the ACM, 23, 1(January 1980), 17-19. [4] Cook, C. R. and Oldehoeft, R.R. A Letter Oriented Minimal Perfect Hashing Function SIGPLAN Notices, 17, 9(September 1982), 18-27.
[5] Cormack, G. V. and Horspool, R. N. S. and Kaiserwerth, M. Practical Perfect Hashing Computer Journal, 28, 1(January 1985), 54-58. [6] Jaeschke, G. Reciprocal Hashing: A Method for Generating Minimal Perfect Hashing Functions Communications of the ACM, 24, 12(December 1981), 829-833.
[7] Jaeschke, G. and Osterburg, G. On Cichelli's Minimal Perfect Hash Functions Method Communications of the ACM, 23, 12(December 1980), 728-729.
[8] Sager, Thomas J. A Polynomial Time Generator for Minimal Perfect Hash Functions Communications of the ACM, 28, 5(December 1985), 523-532
[9] Schmidt, Douglas C. GPERF: A Perfect Hash Function Generator Second USENIX C++ Conference Proceedings, April 1990.
[10] Sebesta, R.W. and Taylor, M.A. Minimal Perfect Hash Functions for Reserved Word Lists SIGPLAN Notices, 20, 12(September 1985), 47-53.
[11] Sprugnoli, R. Perfect Hashing Functions: A Single Probe Retrieving Method for Static Sets Communications of the ACM, 20 11(November 1977), 841-850.
[12] Stallman, Richard M. Using and Porting GNU CC Free Software Foundation, 1988.
[13] Stroustrup, Bjarne The C++ Programming Language. Addison-Wesley, 1986.
[14] Tiemann, Michael D. User's Guide to GNU C++ Free Software Foundation, 1989.
switch
, switch
This document was generated on 26 September 2000 using the texi2html translator version 1.51.