Pipe-Sort (due 11/12)
Learning Objectives
Upon completion of this assignment, you should be able:
1. Work in a multiprocess environment
2. Combine newly written processes with existing system utility processes
3. Gain facility with interprocess communication
New mechanisms you will see and use include:
system calls: fork, exec, wait, pipe, dup2
/usr/bin/sort, buffered I/O, String Routines
NAME
pipesort – An exercise in plumbing
SYNOPSIS
pipesort [n:s:l:]
DESCRIPTION
-n count Count is the number of sorters.
-s short Words must have more than “short” letters
-l long “long” is the maximum number of letters in a word.Letters beyond that are discarded
pipesort reads text from standard input and writes the unique words to standard output sorted in
alphabetic order, using “count sorters”
Exit Status
pipesort returns 0 on success and 1 on failure
DISCUSSION
The output should have one word per line preceded with a count of how
many times it appeared in the input.
Internally, the program must be organized into 3 types of processes. A single
process reads the input parsing the lines into words, a group of processes
does the sorting, and a single process suppresses and counts duplicate
words and writes the output. You should organize the processes so that
every parent waits for its children, and if any process exits in error the main
process returns the error code in its exit.
NOTES:
• You must use the system sort command (/usr/bin/sort) with no
arguments to do the actual sorting,
• You must not use any disk files, or large amounts of memory. There is
NO LIMIT on how large the input may be, and you should test with
large input files.
• Parse distributes the words round robin to the sorts
• All the I/O is done using the buffered I/O library (fget, fputc, etc), you
will need fdopen for attaching to the pipes.
• Words are all alphabetic and case insensitive with the parser
converting all alphabetic characters to lower case. Any non-alphabetic
characters delimit words and are discarded.
• The “short” and “long” parameters affect the parsing. Words with
“short” characters or less are discarded, words with “long” characters
or greater are truncated to “long” characters.
• The output MUST have the count left justified in a 10 character field
followed immediately by the word.
Example short=2 long=6
This is nonsense abc-def$@#abc
some more of this
2 abc
1 def
1 more
1 nonsen
1 some
2 this
HAND in pipesort.c, makefile (README) in
Pipe-Sort (due 11/12)
Learning Objectives