Part 1: Message database search program ---------------------------------------

(a)

The directory "/home/jae/cs3157-pub/bin" contains a simple database into which you can put records, each consisting of a name and a short message. Here are the files in that directory:

mdb-cs3157

- This is the database file. It's a binary file that contains each record one after another. Each record is 40 bytes, so the size of this database file is always a multiple of 40.

mdb-add-cs3157

- This is a program that inserts a record into the mdb-cs3157 database file. It will ask for a name and a short message, and then fills up the following structure with them:

struct MdbRec { char name[16]; char msg[24]; };

The structure in memory is then written at the end of the database file. Note that the name and the message will be truncated to 15 and 23 characters, respectively, in order to fit them into the structure.

mdb-lookup-cs3157 - This is the program you use to see what's in the database and search for a particular name or a message. It will prompt for a string to search for. If you simply press ENTER, it will show you all the records in the database. If you type something, it will show you those records that contain what you typed either in the name field or in the msg field.

Only the first 5 letters are used in the search. So searching for "hello" and "helloooooo" will yield the same result. The match is case-sensitive.

The program keeps running, prompting you for another string to search for. You can press Ctrl-D to terminate the program.

Part 1(a) is easy. You play with mdb-add-cs3157 and mdb-lookup-cs3157, inserting a couple of records and seeing the entertaining messages that your classmates have put in.

List the name and message pairs you have inserted into the database in your README.txt file. You are required to insert at least one record into the database.

(b)

There are two more programs in /home/jae/cs3157-pub/bin. mdb-add and mdb-lookup are similar to the mdb-add-cs3157 and mdb-lookup-cs3157, but they let you work with your own database file. For example, you can insert records into your own database file by running:

mdb-add my-mdb

It will create a database file named my-mdb in your current directory if the file is not there already.

Your job is to write mdb-lookup program that behaves the same way as my mdb-lookup. You don't have to write mdb-add.

As usual, here are some hints and programming requirements:

- Put the MdbRec structure definition in a header file called mdb.h.

- Name your executables "mdb-lookup" and "mdb-add".

- mdb-lookup take the database file name as the sole command line argument.

- When the program starts, you must first read all records from the database file into memory. Moreover, the records must be kept in a linked list using the mylist library from lab 3.

- Just like in part 2 of lab 3, you must not bring any of the mylist library files into lab4 directory. Use -I and -L flags to use the header file and the library file directly from the "../../lab3/solutions/part1" directory. Note the use of a relative path. Do NOT use an absolute path or a path that include your UNI. They will not be accessible when the TAs try to build your code so your build will fail if you use them. Also, make sure you use the solution version of mylist library, not your own version.

- You must keep the records in the linked list in the same order that they were in the database file. addFront() function is not a good choice since it has the effect of inverting the order as you insert the records. Use addAfter() instead. See mylist-test.c from lab 3 for the example of using addAfter() to append items at the end of a linked list.

- Lab 3 solution may contain some additional functions that were not required such as addBack(). Do NOT use addBack() for lab 4. - When reading from the keyboard, you must read in the entire line first, and then take only the first n characters of it. A common mistake would be to read only n characters from the keyboard. Then, the input in the next iteration will begin at the (n+1)th character of the previous line, which is not what you want.

- You may assume that an input line typically won't exceed some large length such as 1000 characters. But your code must NOT hang, produce memory errors, leak memory, or exit unexpectedly if a user enters a string longer than the upper bound above. This behavior is otherwise undefined and you can handle this any other way you see fit.

- mdb-lookup should truncate the search string at 5 characters.

- Some useful functions for taking user input are:

strncpy(), strlen(), fgets(), isprint()

Note that strncpy() may or may not null-terminate the string, and fgets() may or may not include newline character in the buffer. Read their descriptions in the book or man page very carefully.

- Don't forget to print those records that contain the given search string in any of the two fields. The library function strstr() may come in handy.

- Don't forget to print the record number when you print out the records in the same way that my version does; i.e., the record numbers are the positions of the records in the database file, starting from 1.

- Remember that you terminate mdb-lookup by pressing Ctrl-D. Ctrl-D generates the EOF on the standard input. You should design your loop to detect it, so that you can get out of the loop and clean up. In particular, don't forget to free all memories you allocated and close all files you opened.

- Don't forget to check for memory error using valgrind, and include valgrind output in your README.txt.