Fast I/O procedures specialized for different types. Instead of using the Scheme procedure read, specialized procedures can be used, so:
- I/O is faster, particularly if you have loaded the numbers egg
- There is no need to check the result of read
The table below is a simple comaprison done in one system. (This is not a real benchmark; it is just meant to give an approximate idea of what the speedup could be, and even then, the results on a different machine could be diferent!)
scheme specialized-io speedup ------------------------------------------------- write fixnum 6.516 4.396 1.4823 read fixnum 27.98 6.304 4.4385 write flonum 9.596 6.076 1.5793 read flonum 99.396 5.776 17.208 write complex 155.092 20.776 7.4650 read complex 212.08 22.776 9.3116 -------------------------------------------------
The source for the benchmarks is in the subversion repository.
(use specialized-io) (let ((v (read-flonum (current-input-port)))) (write-flonum (fp* 2.5 v) (current-output-port)))
Note that you can use fp* with a value returned from read-flonum (it's guaranteed to be a flonum).
Documentation[procedure] (read-fixnum PORT)
[procedure] (read-flonum PORT)
[procedure] (read-complex PORT)
[procedure] (read-one-char PORT)
These procedures read one object of a specific type from an input-port. There are no runtime checks in the Scheme way; the procedures will just call error if anything goes wront (EOF, lexical error etc).[procedure] (write-fixnum VALUE PORT)
[procedure] (write-flonum VALUE PORT)
[procedure] (write-complex VALUE PORT)
[procedure] (write-one-char VALUE PORT)
These procedures write one object of a specific type to an output-port. There are no runtime checks in the Scheme way; the procedures will just call error if anything goes wront (EOF, lexical error etc).[procedure] (write-one-string STRING PORT)
This procedure writes a string to the output port.[procedure] (read-string-until CHAR PORT)
[procedure] (read-string-between LEFTCHAR RIGHTCHAR PORT)
These procedures read a string from the input port. read-string-until will read from the current position of the stream until CHAR is found (and CHAR will be included in the resulting string). read-string-between will first search for LEFTCHAR then read until RIGHTCHAR is found.[parameter] specialized-io-error
This variable should be bound to a procedure that will be called when parsing fails. By default it's bound to error. To change it, call (specialized-io some-other-error-procedure).[parameter] +specialized-io-string-buffer-size+
This is the size of the buffer in which strings are read. When the buffer becomes full, its size is doubled. This default value is 500. To change it, call (+specialized-io-string-buffer-size+ N), where N is the desired buffer size.
The procedures in this egg are not safe. They may crash if called with arguments different form the expected. In particular, the ports that may be passed to these procedures are those bound to C file descriptors: files, standard input/ourput/error, etc. They do not work, for example, with string ports.
Create a file named "in", with the following content:
10 1.2 1.3 .14 4.2 .5 6.6 7.0 00 0.0 10.0
The code below will read the numbers in the file and put them in a vector.
(use specialized-io) (use srfi-4) (define port (open-input-file "in")) (let ((n (read-fixnum port))) (let ((vec (make-f64vector n))) (do ((i 0 (fx+ 1 i))) ((fx= i n)) (f64vector-set! vec i (read-flonum port))) vec)) ==> #f64(1.2 1.3 0.14 4.2 0.5 6.6 7.0 0.0 0.0 10.0)
- 1.5 Incompatible change! specialized-io-error and +specialized-io-string-buffer-size+ are actually parameters now
- 1.4 Fixed version number; don't shadow write-string; fix other minor issues
- 1.3 Fixed minor glitch
- 1.2 Fixed crash due to wrong return value
- 1.1 String reading and writing added
- 1.0 Initial Release
Copyright (c) 2010, 2011, 2012 Jeronimo C. Pellegrini All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. Neither the name of the author nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.