This document is intended to be an "All in One" guide for programming with ncurses and its sister libraries. We graduate from a simple "Hello World" program to more complex form manipulation. No prior experience in ncurses is assumed. Latest version of the document can always be found at my web site. Send comments to this address
In the olden days of teletype terminals, terminals were away from computers and were connected to them through serial cables. The terminals could be configured by sending a series of bytes to each of them. All the capabilities (such as moving the cursor to a new location, erasing part of the screen, scrolling the screen, changing modes, changing appearance, colors, brightness, blinking, underlining, reverse video etc.) of terminals could be accessed through these series of bytes which are usually called escape sequences because they start with an escape(0x1B) character. Even today, with proper emulation, we can send escape sequences to the emulator and achieve the same effect on the terminal window.
Suppose you wanted to print a line in color. Try typing this on your console.
The first character is an escape character, which looks like two characters ^ and [. To be able to print that you have to press CTRL+V and then the ESC key. All the others are normal printable characters. You should be able to see the string "In Color" in red. It stays that way and to revert back to the original mode type this.
Now, what do those magic characters mean? Difficult to comprehend? They might even be different for different terminals. So the designers of UNIX invented a mechanism named termcap. It is a file that lists all the capabilities of a particular terminal, along with the escape sequences needed to achieve a particular effect. In the later years, this was replaced by terminfo. Without delving too much into details, the concept of the mechanism is to allow application programs query the terminfo database and obtain the control characters to be sent to the terminal or terminal emulator.
You might be wondering, what the import of all this technical gibberish is. In the above scenario, every application program is supposed to query the terminfo and do the necessary stuff(sending control characters etc.). It soon became difficult to manage this complexity and this gave birth to 'CURSES'. Curses is a pun on the name "cursor optimization". The Curses library forms a wrapper over working with raw terminal codes, and provides highly flexible and efficient API (Application Programming Interface). It provides functions to move the cursor, create windows, produce colors, play with mouse etc. The Application programs need not worry about the underlying terminal capabilities.
So what is NCURSES? NCURSES is a clone of the original System V Release 4.0 (SVr4) curses. It is a freely distributable library, fully compatible with older version of curses. In short, it is a library of functions that manages an application's display on character-cell terminals. In the remainder of the document, the terms curses and ncurses are used interchangeably.
The ncurses package was originated by Pavel Curtis. The original maintainer of this package is Zeyd Ben-Halim <firstname.lastname@example.org>. Eric S. Raymond <email@example.com> wrote many of the new features in versions after 1.8.1. Jürgen Pfeifer wrote all of the menu and forms code as well as the Ada95 binding. Ongoing work is being done by Thomas Dickey and Jürgen Pfeifer. Florian La Roche acts as the maintainer for the Free Software Foundation, which holds the copyright on ncurses. Contact the current maintainers at firstname.lastname@example.org.
Ncurses not only creates a wrapper over terminal capabilities, but also gives a robust framework to create nice looking UI (User Interface)s in text mode. It provides functions to create windows etc. Its sister libraries panel, menu and form provide an extension to the basic curses library. These libraries usually come along with curses. One can create applications that contain multiple windows, menus, panels and forms. Windows can be managed independently, can provide 'scrollability' and even can be hidden.
Menus provide the user with an easy command selection option. Forms allow the creation of easy-to-use data entry and display windows. Panels extend the capabilities of ncurses to deal with overlapping and stacked windows.
These are just some of the basic things we can do with ncurses. As we move along, We will see all the capabilities of these libraries.
All right, now that you know what you can do with ncurses, you must be rearing to get started. Ncurses is usually shipped with your installation. In case you don't have the library or want to compile it on your own, read on.
Compiling the package
Read the README and INSTALL files for details on to how to install it. It usually involves the following operations.
Using the RPM
Ncurses RPM can be found and downloaded from http://rpmfind.net . The RPM can be installed with the following command after becoming root.
This document is intended to be a "All in One" guide for programming with ncurses and its sister libraries. We graduate from a simple "Hello World" program to more complex form manipulation. No prior experience in ncurses is assumed.
All the programs in the document are available in zipped form here. Unzip and untar it. The directory structure looks like this.
The individual directories contain the following files.
There is a top level Makefile included in the main directory. It builds all the files and puts the ready-to-use exes in demo/exe directory. You can also do selective make by going into the corresponding directory. Each directory contains a README file explaining the purpose of each c file in the directory.
For every example I have given the path name for the file relative to the ncurses directory.
If you prefer you can browse individual programs at http://tldp.org/HOWTO/NCURSES-Programming-HOWTO/ncurses_programs/
All the programs are released under GPL and you can use them for any thing you like.
This howto is also availabe in various other formats on the tldp.org site. Here are the links to other formats of this document.
If above links are broken or if you want to experiment with sgml read on.
I thank Sharath and Emre Akbas for helping me with few sections. The introduction was initially written by sharath. I rewrote it with few excerpts taken from his initial work. Emre helped in writing printw and scanw sections.
Then comes Ravi Parimi, my dearest friend, who has been on this project before even one line was written. He constantly bombarded me with suggestions and patiently reviewed the whole text. He also checked each program on Linux and Solaris. See his notes to check on your problems.
This is the wish list, in the order of priority. If you have a wish or you want to work on completing the wish, mail me.
Copyright (c) 2001 by Pradeep Padala. This document may be distributed under the terms set forth in the LDP license at linuxdoc.org/COPYRIGHT.html.
This HOWTO is free documentation; you can redistribute it and/or modify it under the terms of the LDP license. This document 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 LDP license for more details.
Welcome to the world of curses. Before we plunge into the library and look into its various features, bells and whistles, let's write a simple program and say hello to the world.
To use ncurses library functions, you have to include ncurses.h and to link the program with ncurses library the flag -lncurses should be added. ncurses.h already includes stdio.h.
Example 1. The Hello World !!! Program
The above program prints "Hello World !!!" to the screen and exits. This program shows how to initialize curses and do screen manipulation and end curses mode. Let's dissect it line by line.
The function initscr() initializes the terminal in curses mode. In some implementations it clears the screen and presents a blank screen. To do any screen manipulation using curses package this has to be called first. This function initializes the curses system and allocates memory for our present window which is called 'stdscr' and some other data-structures. Under extreme cases this function might fail due to insufficient memory to allocate memory for curses library's data structures.
After this is done we can do a variety of initializations to customize our curses settings. These details will be explained later .
The next line printw prints the string "Hello World !!!" on to the screen. This function is analogous to normal printf in all respects except that it prints the data in a window called stdscr at the current (y,x) co-ordinates. Since our present co-ordinates are at 0,0 the string is printed at the left hand corner of the window.
This brings us to that mysterious refresh(). Well, when we did printw actually the data is written to an imaginary window called stdscr, which is not updated on the screen yet. The job of printw is to update a few flags and data structures and write the data to a buffer corresponding to stdscr. In order to bring it to the screen we need to call refresh() and tell the curses system to dump the contents on the screen.
The philosophy behind all this is to allow the programmer to do multiple updates on the imaginary screen or windows and do a refresh once all his screen update is done. refresh() checks the window and updates only the portion which has been changed. This gives good response and offers greater flexibility too. But it is sometimes frustrating to beginners. A common mistake committed by beginners is to forget to call refresh() after they did some update through printw() class of functions. I still forget to add it sometimes :-)
And finally don't forget to end the curses mode. Otherwise your terminal might behave strangely after the program quits. endwin() frees the memory taken by curses sub-system and it's data structures and puts the terminal in normal mode. This function must be called after you are done with the curses mode.
Now that we have seen how to write a simple curses program let's get into the details. There are many functions that help customize what you see on screen and many features which can be put to full use.
Here we go...
We now know that to initialize curses system the function initscr() has to be called. There are functions which can be called after this initialization to customize our curses session. We may ask the curses system to set the terminal in raw mode or initialize color or initialize the mouse etc.. Let's discuss some of the functions that are normally called immediately after initscr();
Normally the terminal driver buffers the characters a user types until a new line or carriage return is encountered. But most programs require that the characters be available as soon as the user types them. The above two functions are used to disable line buffering. The difference between these two functions is in the way control characters like suspend (CTRL-Z), interrupt and quit (CTRL-C) are passed to the program. In the raw() mode these characters are directly passed to the program without generating a signal. In the cbreak() mode these control characters are interpreted as any other character by the terminal driver. I personally prefer to use raw() as I can exercise greater control over what the user does.
These functions control the echoing of characters typed by the user to the terminal. noecho() switches off echoing. The reason you might want to do this is to gain more control over echoing or to suppress unnecessary echoing while taking input from the user through the getch() etc. functions. Most of the interactive programs call noecho() at initialization and do the echoing of characters in a controlled manner. It gives the programmer the flexibility of echoing characters at any place in the window without updating current (y,x) co-ordinates.
This is my favorite initialization function. It enables the reading of function keys like F1, F2, arrow keys etc. Almost every interactive program enables this, as arrow keys are a major part of any User Interface. Do keypad(stdscr, TRUE) to enable this feature for the regular screen (stdscr). You will learn more about key management in later sections of this document.
This function, though not used very often, is a useful one at times. halfdelay()is called to enable the half-delay mode, which is similar to the cbreak() mode in that characters typed are immediately available to program. However, it waits for 'X' tenths of a second for input and then returns ERR, if no input is available. 'X' is the timeout value passed to the function halfdelay(). This function is useful when you want to ask the user for input, and if he doesn't respond with in certain time, we can do some thing else. One possible example is a timeout at the password prompt.
There are few more functions which are called at initialization to customize curses behavior. They are not used as extensively as those mentioned above. Some of them are explained where appropriate.
Let's write a program which will clarify the usage of these functions.
Example 2. Initialization Function Usage example
This program is self-explanatory. But I used functions which aren't explained yet. The function getch() is used to get a character from user. It is equivalent to normal getchar() except that we can disable the line buffering to avoid <enter> after input. Look for more about getch()and reading keys in the key management section . The functions attron and attroff are used to switch some attributes on and off respectively. In the example I used them to print the character in bold. These functions are explained in detail later.
Before we plunge into the myriad ncurses functions, let me clear few things about windows. Windows are explained in detail in following sections
A Window is an imaginary screen defined by curses system. A window does not mean a bordered window which you usually see on Win9X platforms. When curses is initialized, it creates a default window named stdscr which represents your 80x25 (or the size of window in which you are running) screen. If you are doing simple tasks like printing few strings, reading input etc., you can safely use this single window for all of your purposes. You can also create windows and call functions which explicitly work on the specified window.
For example, if you call
It prints the string on stdscr at the present cursor position. Similarly the call to refresh(), works on stdscr only.
Say you have created windows then you have to call a function with a 'w' added to the usual function.
As you will see in the rest of the document, naming of functions follow the same convention. For each function there usually are three more functions.
Usually the w-less functions are macros which expand to corresponding w-function with stdscr as the window parameter.
I guess you can't wait any more to see some action. Back to our odyssey of curses functions. Now that curses is initialized, let's interact with world.
There are three classes of functions which you can use to do output on screen.
These functions can be used interchangeably and it's a matter of style as to which class is used. Let's see each one in detail.
These functions put a single character into the current cursor location and advance the position of the cursor. You can give the character to be printed but they usually are used to print a character with some attributes. Attributes are explained in detail in later sections of the document. If a character is associated with an attribute(bold, reverse video etc.), when curses prints the character, it is printed in that attribute.
In order to combine a character with some attributes, you have two options:
Additionally, curses provides some special characters for character-based graphics. You can draw tables, horizontal or vertical lines, etc. You can find all avaliable characters in the header file ncurses.h. Try looking for macros beginning with ACS_ in this file.
mvaddch() is used to move the cursor to a given point, and then print. Thus, the calls:
waddch() is similar to addch(), except that it adds a character into the given window. (Note that addch() adds a character into the window stdscr.)
In a similar fashion mvwaddch() function is used to add a character into the given window at the given coordinates.
Now, we are familiar with the basic output function addch(). But, if we want to print a string, it would be very annoying to print it character by character. Fortunately, ncurses provides printf-like or puts-like functions.
These functions are similar to printf() with the added capability of printing at any position on the screen.
These two functions work much like printf(). mvprintw() can be used to move the cursor to a position and then print. If you want to move the cursor first and then print using printw() function, use move() first and then use printw() though I see no point why one should avoid using mvprintw(), you have the flexibility to manipulate.
These two functions are similar to above two except that they print in the corresponding window given as argument.
This function is similar to vprintf(). This can be used when variable number of arguments are to be printed.
Example 3. A Simple printw example
Above program demonstrates how easy it is to use printw. You just feed the coordinates and the message to be appeared on the screen, then it does what you want.
The above program introduces us to a new function getmaxyx(), a macro defined in ncurses.h. It gives the number of columns and the number of rows in a given window. getmaxyx() does this by updating the variables given to it. Since getmaxyx() is not a function we don't pass pointers to it, we just give two integer variables.
addstr() is used to put a character string into a given window. This function is similar to calling addch() once for each character in a given string. This is true for all output functions. There are other functions from this family such as mvaddstr(),mvwaddstr() and waddstr(), which obey the naming convention of curses.(e.g. mvaddstr() is similar to the respective calls move() and then addstr().) Another function of this family is addnstr(), which takes an integer parameter(say n) additionally. This function puts at most n characters into the screen. If n is negative, then the entire string will be added.
All these functions take y co-ordinate first and then x in their arguments. A common mistake by beginners is to pass x,y in that order. If you are doing too many manipulations of (y,x) co-ordinates, think of dividing the screen into windows and manipulate each one separately. Windows are explained in the windows section.
Well, printing without taking input, is boring. Let's see functions which allow us to get input from user. These functions also can be divided into three categories.
These functions read a single character from the terminal. But there are several subtle facts to consider. For example if you don't use the function cbreak(), curses will not read your input characters contiguously but will begin read them only after a new line or an EOF is encountered. In order to avoid this, the cbreak() function must used so that characters are immediately available to your program. Another widely used function is noecho(). As the name suggests, when this function is set (used), the characters that are keyed in by the user will not show up on the screen. The two functions cbreak() and noecho() are typical examples of key management. Functions of this genre are explained in the key management section .
These functions are similar to scanf() with the added capability of getting the input from any location on the screen.
The usage of these functions is similar to that of sscanf(), where the line to be scanned is provided by wgetstr() function. That is, these functions call to wgetstr() function(explained below) and uses the resulting line for a scan.
These are similar to above two functions except that they read from a window, which is supplied as one of the arguments to these functions.
This function is similar to vscanf(). This can be used when a variable number of arguments are to be scanned.
These functions are used to get strings from the terminal. In essence, this function performs the same task as would be achieved by a series of calls to getch() until a newline, carriage return, or end-of-file is received. The resulting string of characters are pointed to by str, which is a character pointer provided by the user.
Example 4. A Simple scanw example
We have seen an example of how attributes can be used to print characters with some special effects. Attributes, when set prudently, can present information in an easy, understandable manner. The following program takes a C file as input and prints the file with comments in bold. Scan through the code.
Example 5. A Simple Attributes example
Don't worry about all those initialization and other crap. Concentrate on the while loop. It reads each character in the file and searches for the pattern /*. Once it spots the pattern, it switches the BOLD attribute on with attron() . When we get the pattern */ it is switched off by attroff() .
The above program also introduces us to two useful functions getyx() and move(). The first function gets the co-ordinates of the present cursor into the variables y, x. Since getyx() is a macro we don't have to pass pointers to variables. The function move() moves the cursor to the co-ordinates given to it.
The above program is really a simple one which doesn't do much. On these lines one could write a more useful program which reads a C file, parses it and prints it in different colors. One could even extend it to other languages as well.
Let's get into more details of attributes. The functions attron(), attroff(), attrset() , and their sister functions attr_get() etc.. can be used to switch attributes on/off , get attributes and produce a colorful display.
The functions attron and attroff take a bit-mask of attributes and switch them on or off, respectively. The following video attributes, which are defined in <curses.h> can be passed to these functions.
The last one is the most colorful one :-) Colors are explained in the next sections.
We can OR(|) any number of above attributes to get a combined effect. If you wanted reverse video with blinking characters you can use
Then what is the difference between attron() and attrset()? attrset sets the attributes of window whereas attron just switches on the attribute given to it. So attrset() fully overrides whatever attributes the window previously had and sets it to the new attribute(s). Similarly attroff() just switches off the attribute(s) given to it as an argument. This gives us the flexibility of managing attributes easily.But if you use them carelessly you may loose track of what attributes the window has and garble the display. This is especially true while managing menus with colors and highlighting. So decide on a consistent policy and stick to it. You can always use standend() which is equivalent to attrset(A_NORMAL) which turns off all attributes and brings you to normal mode.
The function attr_get() gets the current attributes and color pair of the window. Though we might not use this as often as the above functions, this is useful in scanning areas of screen. Say we wanted to do some complex update on screen and we are not sure what attribute each character is associated with. Then this function can be used with either attrset or attron to produce the desired effect.
There are series of functions like attr_set(), attr_on etc.. These are similar to above functions except that they take parameters of type attr_t.
For each of the above functions we have a corresponding function with 'w' which operates on a particular window. The above functions operate on stdscr.
The function chgat() is listed in the end of the man page curs_attr. It actually is a useful one. This function can be used to set attributes for a group of characters without moving. I mean it !!! without moving the cursor :-) It changes the attributes of a given number of characters starting at the current cursor location.
We can give -1 as the character count to update till end of line. If you want to change attributes of characters from current position to end of line, just use this.
This function is useful when changing attributes for characters that are already on the screen. Move to the character from which you want to change and change the attribute.
Other functions wchgat(), mvchgat(), wchgat() behave similarly except that the w functions operate on the particular window. The mv functions first move the cursor then perform the work given to them. Actually chgat is a macro which is replaced by a wchgat() with stdscr as the window. Most of the "w-less" functions are macros.
Example 6. Chgat() Usage example
This example also introduces us to the color world of curses. Colors will be explained in detail later. Use 0 for no color.
Windows form the most important concept in curses. You have seen the standard window stdscr above where all the functions implicitly operated on this window. Now to make design even a simplest GUI, you need to resort to windows. The main reason you may want to use windows is to manipulate parts of the screen separately, for better efficiency, by updating only the windows that need to be changed and for a better design. I would say the last reason is the most important in going for windows. You should always strive for a better and easy-to-manage design in your programs. If you are writing big, complex GUIs this is of pivotal importance before you start doing anything.
A Window can be created by calling the function newwin(). It doesn't create any thing on the screen actually. It allocates memory for a structure to manipulate the window and updates the structure with data regarding the window like it's size, beginy, beginx etc.. Hence in curses, a window is just an abstraction of an imaginary window, which can be manipulated independent of other parts of screen. The function newwin() returns a pointer to structure WINDOW, which can be passed to window related functions like wprintw() etc.. Finally the window can be destroyed with delwin(). It will deallocate the memory associated with the window structure.
What fun is it, if a window is created and we can't see it. So the fun part begins by displaying the window. The function box() can be used to draw a border around the window. Let's explore these functions in more detail in this example.
Example 7. Window Border example
Don't scream. I know it's a big example. But I have to explain some important things here :-). This program creates a rectangular window that can be moved with left, right, up, down arrow keys. It repeatedly creates and destroys windows as user press a key. Don't go beyond the screen limits. Checking for those limits is left as an exercise for the reader. Let's dissect it by line by line.
The create_newwin() function creates a window with newwin() and displays a border around it with box. The function destroy_win() first erases the window from screen by painting a border with ' ' character and then calling delwin() to deallocate memory related to it. Depending on the key the user presses, starty or startx is changed and a new window is created.
In the destroy_win, as you can see, I used wborder instead of box. The reason is written in the comments (You missed it. I know. Read the code :-)). wborder draws a border around the window with the characters given to it as the 4 corner points and the 4 lines. To put it clearly, if you have called wborder as below:
it produces some thing like
You can also see in the above examples, that I have used the variables COLS, LINES which are initialized to the screen sizes after initscr(). They can be useful in finding screen dimensions and finding the center co-ordinate of the screen as above. The function getch() as usual gets the key from keyboard and according to the key it does the corresponding work. This type of switch- case is very common in any GUI based programs.
Above program is grossly inefficient in that with each press of a key, a window is destroyed and another is created. So let's write a more efficient program which uses other border related functions.
The following program uses mvhline() and mvvline() to achieve similar effect. These two functions are simple. They create a horizontal or vertical line of the specified length at the specified position.
Example 8. More border functions
Life seems dull with no colors. Curses has a nice mechanism to handle colors. Let's get into the thick of the things with a small program.
Example 9. A Simple Color example
As you can see, to start using color, you should first call the function start_color(). After that you can use color capabilities of your terminals by various functions. To find out whether terminal has color capabilities or not, you can use has_colors() function, which returns FALSE if the terminal does not support color.
Curses initializes all the colors supported by terminal when start_color() is called. These can be accessed by the define constants like COLOR_BLACK etc. Now to actually start using colors, you have to define pairs. Colors are always used in pairs. That means you have to use the function init_pair() to define the foreground and background for the pair number you give. After that that pair number can be used as a normal attribute with COLOR_PAIR()function. This may seem to be cumbersome at first. But this elegant solution allows us to manage color pairs very easily. To appreciate it, you have to look into the the source code of "dialog", a utility for displaying dialog boxes from shell scripts. The developers have defined foreground and background combinations for all the colors they might need and initialized at the beginning. This makes it very easy to set attributes just by accessing a pair which we already have defined as a constant.
The following colors are defined in curses.h. You can use these as parameters for various color functions.
The function init_color()can be used to change the rgb values for the colors defined by curses initially. Say you wanted to lighten the intensity of red color by a minuscule. Then you can use this function as
If your terminal cannot change the color definitions, the function returns ERR. The function can_change_color() can be used to find out whether the terminal has the capability of changing color content or not. The rgb content is scaled from 0 to 1000. Initially RED color is defined with content 1000(r), 0(g), 0(b).
The functions color_content() and pair_content() can be used to find the color content and foreground, background combination for the pair.
No GUI is complete without a strong user interface and to interact with the user, a curses program should be sensitive to key presses or the mouse actions done by the user. Let's deal with the keys first.
As you have seen in almost all of the above examples, it's very easy to get key input from the user. A simple way of getting key presses is to use getch() function. The cbreak mode should be enabled to read keys when you are interested in reading individual key hits rather than complete lines of text (which usually end with a carriage return). keypad should be enabled to get the Functions keys, arrow keys etc. See the initialization section for details.
getch() returns an integer corresponding to the key pressed. If it is a normal character, the integer value will be equivalent to the character. Otherwise it returns a number which can be matched with the constants defined in curses.h. For example if the user presses F1, the integer returned is 265. This can be checked using the macro KEY_F() defined in curses.h. This makes reading keys portable and easy to manage.
For example, if you call getch() like this
getch() will wait for the user to press a key, (unless you specified a timeout) and when user presses a key, the corresponding integer is returned. Then you can check the value returned with the constants defined in curses.h to match against the keys you want.
The following code piece will do that job.
Let's write a small program which creates a menu which can be navigated by up and down arrows.
Example 10. A Simple Key Usage example
Now that you have seen how to get keys, lets do the same thing from mouse. Usually each UI allows the user to interact with both keyboard and mouse.
Before you do any thing else, the events you want to receive have to be enabled with mousemask().
The first parameter to above function is a bit mask of events you would like to listen. By default, all the events are turned off. The bit mask ALL_MOUSE_EVENTS can be used to get all the events.
The following are all the event masks:
Once a class of mouse events have been enabled, getch() class of functions return KEY_MOUSE every time some mouse event happens. Then the mouse event can be retrieved with getmouse().
The code approximately looks like this:
getmouse() returns the event into the pointer given to it. It's a structure which contains
The bstate is the main variable we are interested in. It tells the button state of the mouse.
Then with a code snippet like the following, we can find out what happened.
That's pretty much interfacing with mouse. Let's create the same menu and enable mouse interaction. To make things simpler, key handling is removed.
Example 11. Access the menu with mouse !!!
The functions mouse_trafo() and wmouse_trafo() can be used to convert to mouse co-ordinates to screen relative co-ordinates. See curs_mouse(3X) man page for details.
The mouseinterval function sets the maximum time (in thousands of a second) that can elapse between press and release events in order for them to be recognized as a click. This function returns the previous interval value. The default is one fifth of a second.
In this section, we will look into some functions, which allow us to manage the screen efficiently and to write some fancy programs. This is especially important in writing games.
The function getyx() can be used to find out the present cursor co-ordinates. It will fill the values of x and y co-ordinates in the arguments given to it. Since getyx() is a macro you don't have to pass the address of the variables. It can be called as
The function getparyx() gets the beginning co-ordinates of the sub window relative to the main window. This is some times useful to update a sub window. When designing fancy stuff like writing multiple menus, it becomes difficult to store the menu positions, their first option co-ordinates etc. A simple solution to this problem, is to create menus in sub windows and later find the starting co-ordinates of the menus by using getparyx().
The functions getbegyx() and getmaxyx() store current window's beginning and maximum co-ordinates. These functions are useful in the same way as above in managing the windows and sub windows effectively.
While writing games, some times it becomes necessary to store the state of the screen and restore it back to the same state. The function scr_dump() can be used to dump the screen contents to a file given as an argument. Later it can be restored by scr_restore function. These two simple functions can be used effectively to maintain a fast moving game with changing scenarios.
To store and restore windows, the functions putwin() and getwin() can be used. putwin() puts the present window state into a file, which can be later restored by getwin().
The function copywin() can be used to copy a window completely onto another window. It takes the source and destination windows as parameters and according to the rectangle specified, it copies the rectangular region from source to destination window. It's last parameter specifies whether to overwrite or just overlay the contents on to the destination window. If this argument is true, then the copying is non-destructive.
Now you know enough features to write a good curses program, with all bells and whistles. There are some miscellaneous functions which are useful in various cases. Let's go headlong into some of those.
This function can be used to make the cursor invisible. The parameter to this function should be
Some times you may want to get back to cooked mode (normal line buffering mode) temporarily. In such a case you will first need to save the tty modes with a call to def_prog_mode() and then call endwin() to end the curses mode. This will leave you in the original tty mode. To get back to curses once you are done, call reset_prog_mode() . This function returns the tty to the state stored by def_prog_mode(). Then do refresh(), and you are back to the curses mode. Here is an example showing the sequence of things to be done.
Example 12. Temporarily Leaving Curses Mode
If you have ever programmed in DOS, you know about those nifty characters in extended character set. They are printable only on some terminals. NCURSES functions like box() use these characters. All these variables start with ACS meaning alternative character set. You might have noticed me using these characters in some of the programs above. Here's an example showing all the characters.
Example 13. ACS Variables Example
Apart from the curses library, there are few text mode libraries, which provide more functionality and a lot of features. The following sections explain three standard libraries which are usually distributed along with curses.
Now that you are proficient in curses, you wanted to do some thing big. You created a lot of overlapping windows to give a professional windows-type look. Unfortunately, it soon becomes difficult to manage these. The multiple refreshes, updates plunge you into a nightmare. The overlapping windows create blotches, whenever you forget to refresh the windows in the proper order.
Don't despair. There's an elegant solution provided in panels library. In the words of developers of ncurses
When your interface design is such that windows may dive deeper into the visibility stack or pop to the top at runtime, the resulting book-keeping can be tedious and difficult to get right. Hence the panels library.
If you have lot of overlapping windows, then panels library is the way to go. It obviates the need of doing series of wnoutrefresh(), doupdate() and relieves the burden of doing it correctly(bottom up). The library maintains information about the order of windows, their overlapping and update the screen properly. So why wait? Let's take a close peek into panels.
Panel object is a window that is implicitly treated as part of a deck including all other panel objects. The deck is treated as a stack with the top panel being completely visible and the other panels may or may not be obscured according to their positions. So the basic idea is to create a stack of overlapping panels and use panels library to display them correctly. There is a function similar to refresh() which, when called , displays panels in the correct order. Functions are provided to hide or show panels, move panels, change its size etc.. The overlapping problem is managed by the panels library during all the calls to these functions.
The general flow of a panel program goes like this:
Let's make the concepts clear, with some programs. The following is a simple program which creates 3 overlapping panels and shows them on the screen.
To use panels library functions, you have to include panel.h and to link the program with panels library the flag -lpanel should be added along with -lncurses in that order.
Example 14. Panel basics
As you can see, above program follows a simple flow as explained. The windows are created with newwin() and then they are attached to panels with new_panel(). As we attach one panel after another, the stack of panels gets updated. To put them on screen update_panels() and doupdate() are called.
A slightly complicated example is given below. This program creates 3 windows which can be cycled through using tab. Have a look at the code.
Example 15. Panel Window Browsing Example
In the above example I used user pointers to find out the next window in the cycle. We can attach custom information to the panel by specifying a user pointer, which can point to any information you want to store. In this case I stored the pointer to the next panel in the cycle. User pointer for a panel can be set with the function set_panel_userptr(). It can be accessed using the function panel_userptr() which will return the user pointer for the panel given as argument. After finding the next panel in the cycle It's brought to the top by the function top_panel(). This function brings the panel given as argument to the top of the panel stack.
The function move_panel() can be used to move a panel to the desired location. It does not change the position of the panel in the stack. Make sure that you use move_panel() instead mvwin() on the window associated with the panel.
Resizing a panel is slightly complex. There is no straight forward function just to resize the window associated with a panel. A solution to resize a panel is to create a new window with the desired sizes, change the window associated with the panel using replace_panel(). Don't forget to delete the old window. The window associated with a panel can be found by using the function panel_window().
The following program shows these concepts, in supposedly simple program. You can cycle through the window with <TAB> as usual. To resize or move the active panel press 'r' for resize 'm' for moving. Then use arrow keys to resize or move it to the desired way and press enter to end your resizing or moving. This example makes use of user data to get the required data to do the operations.
Example 16. Panel Moving and Resizing example
Concentrate on the main while loop. Once it finds out the type of key pressed, it takes appropriate action. If 'r' is pressed resizing mode is started. After this the new sizes are updated as the user presses the arrow keys. When the user presses <ENTER> present selection ends and panel is resized by using the concept explained. While in resizing mode the program doesn't show how the window is getting resized. It's left as an exercise to the reader to print a dotted border while it gets resized to a new position.
When the user presses 'm' the move mode starts. This is a bit simpler than resizing. As the arrow keys are pressed the new position is updated and pressing of <ENTER> causes the panel to be moved by calling the function move_panel().
In this program the user data which is represented as PANEL_DATA, plays very important role in finding the associated information with a panel. As written in the comments, the PANEL_DATA stores the panel sizes, label, label color and a pointer to the next panel in the cycle.
A Panel can be hidden by using the function hide_panel(). This function merely removes it form the stack of panels, thus hiding it on the screen once you do update_panels() and doupdate(). It doesn't destroy the PANEL structure associated with the hidden panel. It can be shown again by using the show_panel() function.
The following program shows the hiding of panels. Press 'a' or 'b' or 'c' to show or hide first, second and third windows respectively. It uses a user data with a small variable hide, which keeps track of whether the window is hidden or not. For some reason the function panel_hidden() which tells whether a panel is hidden or not is not working. A bug report was also presented by Michael Andres here
Example 17. Panel Hiding and Showing example
The functions panel_above() and panel_below() can be used to find out the panel above and below a panel. If the argument to these functions is NULL, then they return a pointer to bottom panel and top panel respectively.
The menus library provides a nice extension to basic curses, through which you can create menus. It provides a set of functions to create menus. But they have to be customized to give a nicer look, with colors etc. Let's get into the details.
A menu is a screen display that assists the user to choose some subset of a given set of items. To put it simple, a menu is a collection of items from which one or more items can be chosen. Some readers might not be aware of multiple item selection capability. Menu library provides functionality to write menus from which the user can chose more than one item as the preferred choice. This is dealt with in a later section. Now it is time for some rudiments.
To create menus, you first create items, and then post the menu to the display. After that, all the processing of user responses is done in an elegant function menu_driver() which is the work horse of any menu program.
The general flow of control of a menu program looks like this.
Let's see a program which prints a simple menu and updates the current selection with up, down arrows.
To use menu library functions, you have to include menu.h and to link the program with menu library the flag -lmenu should be added along with -lncurses in that order.
Example 18. Menu Basics
This program demonstrates the basic concepts involved in creating a menu using menus library. First we create the items using new_item() and then attach them to the menu with new_menu() function. After posting the menu and refreshing the screen, the main processing loop starts. It reads user input and takes corresponding action. The function menu_driver() is the main work horse of the menu system. The second parameter to this function tells what's to be done with the menu. According to the parameter, menu_driver() does the corresponding task. The value can be either a menu navigational request, an ascii character, or a KEY_MOUSE special key associated with a mouse event.
The menu_driver accepts following navigational requests.
Don't get overwhelmed by the number of options. We will see them slowly one after another. The options of interest in this example are REQ_UP_ITEM and REQ_DOWN_ITEM. These two options when passed to menu_driver, menu driver updates the current item to one item up or down respectively.
As you have seen in the above example, menu_driver plays an important role in updating the menu. It is very important to understand various options it takes and what they do. As explained above, the second parameter to menu_driver() can be either a navigational request, a printable character or a KEY_MOUSE key. Let's dissect the different navigational requests.
Each of the above requests will be explained in the following lines with several examples whenever appropriate.
Every menu created is associated with a window and a sub window. The menu window displays any title or border associated with the menu. The menu sub window displays the menu items currently available for selection. But we didn't specify any window or sub window in the simple example. When a window is not specified, stdscr is taken as the main window, and then menu system calculates the sub window size required for the display of items. Then items are displayed in the calculated sub window. So let's play with these windows and display a menu with a border and a title.
Example 19. Menu Windows Usage example
This example creates a menu with a title, border, a fancy line separating title and the items. As you can see, in order to attach a window to a menu the function set_menu_win() has to be used. Then we attach the sub window also. This displays the items in the sub window. You can also set the mark string which gets displayed to the left of the selected item with set_menu_mark().
If the sub window given for a window is not big enough to show all the items, then the menu will be scrollable. When you are on the last item in the present list, if you send REQ_DOWN_ITEM, it gets translated into REQ_SCR_DLINE and the menu scrolls by one item. You can manually give REQ_SCR_ operations to do scrolling. Let's see how it can be done.
Example 20. Scrolling Menus example
This program is self-explanatory. In this example the number of choices has been increased to ten, which is larger than our sub window size which can hold 6 items. This message has to be explicitly conveyed to the menu system with the function set_menu_format(). In here we specify the number of rows and columns we want to be displayed for a single page. We can specify any number of items to be shown, in the rows variables, if it is less than the height of the sub window. If the key pressed by the user is a PAGE UP or PAGE DOWN, the menu is scrolled a page due to the requests (REQ_SCR_DPAGE and REQ_SCR_UPAGE) given to menu_driver().
In the above example you have seen how to use the function set_menu_format(). I didn't mention what the cols variable (third parameter) does. Well, If your sub window is wide enough, you can opt to display more than one item per row. This can be specified in the cols variable. To make things simpler, the following example doesn't show descriptions for the items.
Example 21. Milt Columnar Menus Example
Watch the function call to set_menu_format(). It specifies the number of columns to be 3, thus displaying 3 items per row. We have also switched off the showing descriptions with the function menu_opts_off(). There are couple of functions set_menu_opts(), menu_opts_on() and menu_opts() which can be used to manipulate menu options. The following menu options can be specified.
All options are on by default. You can switch specific attributes on or off with menu_opts_on() and menu_opts_off() functions. You can also use set_menu_opts() to directly specify the options. The argument to this function should be a OR ed value of some of those above constants. The function menu_opts() can be used to find out a menu's present options.
You might be wondering what if you switch off the option O_ONEVALUE. Then the menu becomes multi-valued. That means you can select more than one item. This brings us to the request REQ_TOGGLE_ITEM. Let's see it in action.
Example 22. Multi Valued Menus example
Whew, A lot of new functions. Let's take them one after another. Firstly, the REQ_TOGGLE_ITEM. In a multi-valued menu, the user should be allowed to select or un select more than one item. The request REQ_TOGGLE_ITEM toggles the present selection. In this case when space is pressed REQ_TOGGLE_ITEM request is sent to menu_driver to achieve the result.
Now when the user presses <ENTER> we show the items he presently selected. First we find out the items associated with the menu using the function menu_items(). Then we loop through the items to find out if the item is selected or not. The function item_value() returns TRUE if an item is selected. The function item_count() returns the number of items in the menu. The item name can be found with item_name(). You can also find the description associated with an item using item_description().
Well, by this time you must be itching for some difference in your menu, with lots of functionality. I know. You want Colors !!!. You want to create nice menus similar to those text mode dos games. The functions set_menu_fore() and set_menu_back() can be used to change the attribute of the selected item and unselected item. The names are misleading. They don't change menu's foreground or background which would have been useless.
The function set_menu_grey() can be used to set the display attribute for the non-selectable items in the menu. This brings us to the interesting option for an item the one and only O_SELECTABLE. We can turn it off by the function item_opts_off() and after that that item is not selectable. It's like a grayed item in those fancy windows menus. Let's put these concepts in practice with this example
Example 23. Menu Options example
We can associate a user pointer with each item in the menu. It works the same way as user pointer in panels. It's not touched by menu system. You can store any thing you like in that. I usually use it to store the function to be executed when the menu option is chosen (It's selected and may be the user pressed <ENTER>);
Example 24. Menu User Pointer Usage
Well. If you have seen those forms on web pages which take input from users and do various kinds of things, you might be wondering how would any one create such forms in text mode display. It's quite difficult to write those nifty forms in plain ncurses. Forms library tries to provide a basic frame work to build and maintain forms with ease. It has lot of features(functions) which manage validation, dynamic expansion of fields etc.. Let's see it in full flow.
A form is a collection of fields; each field can be either a label(static text) or a data-entry location. The forms also library provides functions to divide forms into multiple pages.
Forms are created in much the same way as menus. First the fields related to the form are created with new_field(). You can set options for the fields, so that they can be displayed with some fancy attributes, validated before the field looses focus etc.. Then the fields are attached to form. After this, the form can be posted to display and is ready to receive inputs. On the similar lines to menu_driver(), the form is manipulated with form_driver(). We can send requests to form_driver to move focus to a certain field, move cursor to end of the field etc.. After the user enters values in the fields and validation done, form can be unposted and memory allocated can be freed.
The general flow of control of a forms program looks like this.
As you can see, working with forms library is much similar to handling menu library. The following examples will explore various aspects of form processing. Let's start the journey with a simple example. first.
To use forms library functions, you have to include form.h and to link the program with forms library the flag -lform should be added along with -lncurses in that order.
Example 25. Forms Basics
Above example is pretty straight forward. It creates two fields with new_field(). new_field() takes height, width, starty, startx, number of offscreen rows and number of additional working buffers. The fifth argument number of offscreen rows specifies how much of the field to be shown. If it is zero, the entire field is always displayed otherwise the form will be scrollable when the user accesses not displayed parts of the field. The forms library allocates one buffer per field to store the data user enters. Using the last parameter to new_field() we can specify it to allocate some additional buffers. These can be used for any purpose you like.
After creating the fields, back ground attribute of both of them is set to an underscore with set_field_back(). The AUTOSKIP option is turned off using field_opts_off(). If this option is turned on, focus will move to the next field in the form once the active field is filled up completely.
After attaching the fields to the form, it is posted. Here on, user inputs are processed in the while loop, by making corresponding requests to form_driver. The details of all the requests to the form_driver() are explained later.
Each form field is associated with a lot of attributes. They can be manipulated to get the required effect and to have fun !!!. So why wait?
The parameters we have given at the time of creation of a field can be retrieved with field_info(). It returns height, width, starty, startx, number of offscreen rows, and number of additional buffers into the parameters given to it. It is a sort of inverse of new_field().
The location of the field can be moved to a different position with move_field().
As usual, the changed position can be queried with field_infor().
The justification to be done for the field can be fixed using the function set_field_just().
The justification mode valued accepted and returned by these functions are NO_JUSTIFICATION, JUSTIFY_RIGHT, JUSTIFY_LEFT, or JUSTIFY_CENTER.
As you have seen, in the above example, display attribute for the fields can be set with set_field_fore() and setfield_back(). These functions set foreground and background attribute of the fields. You can also specify a pad character which will be filled in the unfilled portion of the field. The pad character is set with a call to set_field_pad(). Default pad value is a space. The functions field_fore(), field_back, field_pad() can be used to query the present foreground, background attributes and pad character for the field. The following list gives the usage of functions.
Though above functions seem quite simple, using colors with set_field_fore() may be frustrating in the beginning. Let me first explain about foreground and background attributes of a field. The foreground attribute is associated with the character. That means a character in the field is printed with the attribute you have set with set_field_fore(). Background attribute is the attribute used to fill background of field, whether any character is there or not. So what about colors? Since colors are always defined in pairs, what is the right way to display colored fields? Here's an example clarifying color attributes.
Example 26. Form Attributes example
Play with the color pairs and try to understand the foreground and background attributes. In my programs using color attributes, I usually set only the background with set_field_back(). Curses simply doesn't allow defining individual color attributes.
There is also a large collection of field option bits you can set to control various aspects of forms processing. You can manipulate them with these functions:
The function set_field_opts() can be used to directly set attributes of a field or you can choose to switch a few attributes on and off with field_opts_on() and field_opts_off() selectively. Anytime you can query the attributes of a field with field_opts(). The following is the list of available options. By default, all options are on.
A field's options cannot be changed while the field is currently selected. However, options may be changed on posted fields that are not current.
The option values are bit-masks and can be composed with logical-or in the obvious way. You have seen the usage of switching off O_AUTOSKIP option. The following example clarifies usage of some more options. Other options are explained where appropriate.
Example 27. Field Options Usage example
This example, though useless, shows the usage of options. If used properly, they can present information very effectively in a form. The second field being not O_PUBLIC, does not show the characters you are typing.
The field status specifies whether the field has got edited or not. It is initially set to FALSE and when user enters something and the data buffer gets modified it becomes TRUE. So a field's status can be queried to find out whether it has been modified or not. The following functions can assist in those operations.
It's better to check the field's status only after after leaving the field, as data buffer might not have been updated yet as the validation is still due. To guarantee that right status is returned, call field_status() either (1) in the field's exit validation check routine, (2) from the field's or form's initialization or termination hooks, or (3) just after a REQ_VALIDATION request has been processed by the forms driver
Every field structure contains one pointer that can be used by the user for various purposes. It is not touched by forms library and can be used for any purpose by the user. The following functions set and fetch user pointer.
If you want a dynamically changing field with variable width, this is the feature you want to put to full use. This will allow the user to enter more data than the original size of the field and let the field grow. According to the field orientation it will scroll horizontally or vertically to incorporate the new data.
To make a field dynamically growable, the option O_STATIC should be turned off. This can be done with a
But it's usually not advisable to allow a field to grow infinitely. You can set a maximum limit to the growth of the field with
The field info for a dynamically growable field can be retrieved by
Recall the library routine new_field; a new field created with height set to one will be defined to be a one line field. A new field created with height greater than one will be defined to be a multi line field.
A one line field with O_STATIC turned off (dynamically growable field) will contain a single fixed row, but the number of columns can increase if the user enters more data than the initial field will hold. The number of columns displayed will remain fixed and the additional data will scroll horizontally.
A multi line field with O_STATIC turned off (dynamically growable field) will contain a fixed number of columns, but the number of rows can increase if the user enters more data than the initial field will hold. The number of rows displayed will remain fixed and the additional data will scroll vertically.
The above two paragraphs pretty much describe a dynamically growable field's behavior. The way other parts of forms library behaves is described below:
Some of the above points make sense only after explaining form driver. We will be looking into that in next few sections.
The form windows concept is pretty much similar to menu windows. Every form is associated with a main window and a sub window. The form main window displays any title or border associated or whatever the user wishes. Then the sub window contains all the fields and displays them according to their position. This gives the flexibility of manipulating fancy form displaying very easily.
Since this is pretty much similar to menu windows, I am providing an example with out much explanation. The functions are similar and they work the same way.
Example 28. Form Windows Example
By default, a field will accept any data input by the user. It is possible to attach validation to the field. Then any attempt by the user to leave the field, while it contains data that doesn't match the validation type will fail. Some validation types also have a character-validity check for each time a character is entered in the field.
Validation can be attached to a field with the following function.
The form driver validates the data in a field only when data is entered by the end-user. Validation does not occur when
The following are the pre-defined validation types. You can also specify custom validation, though it's a bit tricky and cumbersome.
This field type accepts alphabetic data; no blanks, no digits, no special characters (this is checked at character-entry time). It is set up with:
The width argument sets a minimum width of data. The user has to enter at-least width number of characters before he can leave the field. Typically you'll want to set this to the field width; if it's greater than the field width, the validation check will always fail. A minimum width of zero makes field completion optional.
This field type accepts alphabetic data and digits; no blanks, no special characters (this is checked at character-entry time). It is set up with:
The width argument sets a minimum width of data. As with TYPE_ALPHA, typically you'll want to set this to the field width; if it's greater than the field width, the validation check will always fail. A minimum width of zero makes field completion optional.
This type allows you to restrict a field's values to be among a specified set of string values (for example, the two-letter postal codes for U.S. states). It is set up with:
The valuelist parameter must point at a NULL-terminated list of valid strings. The checkcase argument, if true, makes comparison with the string case-sensitive.
When the user exits a TYPE_ENUM field, the validation procedure tries to complete the data in the buffer to a valid entry. If a complete choice string has been entered, it is of course valid. But it is also possible to enter a prefix of a valid string and have it completed for you.
By default, if you enter such a prefix and it matches more than one value in the string list, the prefix will be completed to the first matching value. But the checkunique argument, if true, requires prefix matches to be unique in order to be valid.
The REQ_NEXT_CHOICE and REQ_PREV_CHOICE input requests can be particularly useful with these fields.
This field type accepts an integer. It is set up as follows:
Valid characters consist of an optional leading minus and digits. The range check is performed on exit. If the range maximum is less than or equal to the minimum, the range is ignored.
If the value passes its range check, it is padded with as many leading zero digits as necessary to meet the padding argument.
A TYPE_INTEGER value buffer can conveniently be interpreted with the C library function atoi(3).
This field type accepts a decimal number. It is set up as follows:
Valid characters consist of an optional leading minus and digits. possibly including a decimal point. The range check is performed on exit. If the range maximum is less than or equal to the minimum, the range is ignored.
If the value passes its range check, it is padded with as many trailing zero digits as necessary to meet the padding argument.
A TYPE_NUMERIC value buffer can conveniently be interpreted with the C library function atof(3).
This field type accepts data matching a regular expression. It is set up as follows:
The syntax for regular expressions is that of regcomp(3). The check for regular-expression match is performed on exit.
As in the menu system, form_driver() plays a very important role in forms system. All types of requests to forms system should be funneled through form_driver().
As you have seen some of the examples above, you have to be in a loop looking for user input and then decide whether it's a field data or a form request. The form requests are then passed to form_driver() to do the work.
The requests roughly can be divided into following categories. Different requests and their usage is explained below:
These requests cause page-level moves through the form, triggering display of a new form screen. A form can be made of multiple pages. If you have a big form with lot of fields and logical sections, then you can divide the form into pages. The function set_new_page() to set a new page at the field specified.
The following requests allow you to move to different pages
These requests treat the list as cyclic; that is, REQ_NEXT_PAGE from the last page goes to the first, and REQ_PREV_PAGE from the first page goes to the last.
These requests handle navigation between fields on the same page.
These requests treat the list of fields on a page as cyclic; that is, REQ_NEXT_FIELD from the last field goes to the first, and REQ_PREV_FIELD from the first field goes to the last. The order of the fields for these (and the REQ_FIRST_FIELD and REQ_LAST_FIELD requests) is simply the order of the field pointers in the form array (as set up by new_form() or set_form_fields()
It is also possible to traverse the fields as if they had been sorted in screen-position order, so the sequence goes left-to-right and top-to-bottom. To do this, use the second group of four sorted-movement requests.
Finally, it is possible to move between fields using visual directions up, down, right, and left. To accomplish this, use the third group of four requests. Note, however, that the position of a form for purposes of these requests is its upper-left corner.
For example, suppose you have a multi-line field B, and two single-line fields A and C on the same line with B, with A to the left of B and C to the right of B. A REQ_MOVE_RIGHT from A will go to B only if A, B, and C all share the same first line; otherwise it will skip over B to C.
These requests drive movement of the edit cursor within the currently selected field.
Each word is separated from the previous and next characters by whitespace. The commands to move to beginning and end of line or field look for the first or last non-pad character in their ranges.
Fields that are dynamic and have grown and fields explicitly created with offscreen rows are scrollable. One-line fields scroll horizontally; multi-line fields scroll vertically. Most scrolling is triggered by editing and intra-field movement (the library scrolls the field to keep the cursor visible). It is possible to explicitly request scrolling with the following requests:
For scrolling purposes, a page of a field is the height of its visible part.
When you pass the forms driver an ASCII character, it is treated as a request to add the character to the field's data buffer. Whether this is an insertion or a replacement depends on the field's edit mode (insertion is the default.
The following requests support editing the field and changing the edit mode:
The behavior of the REQ_NEW_LINE and REQ_DEL_PREV requests is complicated and partly controlled by a pair of forms options. The special cases are triggered when the cursor is at the beginning of a field, or on the last line of the field.
First, we consider REQ_NEW_LINE:
The normal behavior of REQ_NEW_LINE in insert mode is to break the current line at the position of the edit cursor, inserting the portion of the current line after the cursor as a new line following the current and moving the cursor to the beginning of that new line (you may think of this as inserting a newline in the field buffer).
The normal behavior of REQ_NEW_LINE in overlay mode is to clear the current line from the position of the edit cursor to end of line. The cursor is then moved to the beginning of the next line.
However, REQ_NEW_LINE at the beginning of a field, or on the last line of a field, instead does a REQ_NEXT_FIELD. O_NL_OVERLOAD option is off, this special action is disabled.
Now, let us consider REQ_DEL_PREV:
The normal behavior of REQ_DEL_PREV is to delete the previous character. If insert mode is on, and the cursor is at the start of a line, and the text on that line will fit on the previous one, it instead appends the contents of the current line to the previous one and deletes the current line (you may think of this as deleting a newline from the field buffer).
However, REQ_DEL_PREV at the beginning of a field is instead treated as a REQ_PREV_FIELD.
If the O_BS_OVERLOAD option is off, this special action is disabled and the forms driver just returns E_REQUEST_DENIED.
If the type of your field is ordered, and has associated functions for getting the next and previous values of the type from a given value, there are requests that can fetch that value into the field buffer:
Of the built-in field types, only TYPE_ENUM has built-in successor and predecessor functions. When you define a field type of your own (see Custom Validation Types), you can associate our own ordering functions.
Form requests are represented as integers above the curses value greater than KEY_MAX and less than or equal to the constant MAX_COMMAND. A value within this range gets ignored by form_driver(). So this can be used for any purpose by the application. It can be treated as an application specific action and take corresponding action.
Now that you have seen the capabilities of ncurses and its sister libraries, you are rolling your sleeves up and gearing for a project that heavily manipulates screen. But wait.. It can be pretty difficult to write and maintain complex GUI widgets in plain ncurses or even with the additional libraries. There are some ready-to-use tools and widget libraries that can be used instead of writing your own widgets. You can use some of them, get ideas from the code, or even extend them.
In the author's words
CDK stands for 'Curses Development Kit' and it currently contains 21 ready to use widgets which facilitate the speedy development of full screen curses programs.
The kit provides some useful widgets, which can be used in your programs directly. It's pretty well written and the documentation is very good. The examples in the examples directory can be a good place to start for beginners. The CDK can be downloaded from http://www.vexus.ca/release/cdk.tar.gz . Follow the instructions in README file to install it.
The following is the list of widgets provided with cdk and their description.
Apart from making our life easier with readily usable widgets, cdk solves one frustrating problem with printing multi colored strings, justified strings elegantly. Special formatting tags can be embedded in the strings which are passed to CDK functions. For Example
If the string
given as a parameter to newCDKLabel(), it prints the line with yellow foreground and blue background. There are other tags available for justifying string, embedding special drawing characters etc.. Please refer to the man page cdk_display(3X) for details. The man page explains the usage with nice examples.
All in all, CDK is a well-written package of widgets, which if used properly can form a strong frame work for developing complex GUI.
Long long ago, in September 1994, when few people knew linux, Jeff Tranter wrote an article on dialog in Linux Journal. He starts the article with these words..
Linux is based on the Unix operating system, but also features a number of unique and useful kernel features and application programs that often go beyond what is available under Unix. One little-known gem is "dialog", a utility for creating professional-looking dialog boxes from within shell scripts. This article presents a tutorial introduction to the dialog utility, and shows examples of how and where it can be used
As he explains, dialog is a real gem in making professional-looking dialog boxes with ease. It creates a variety of dialog boxes, menus, check lists etc.. It is usually installed by default. If not, you can find it at ibiblio linux archive.
The above-mentioned article gives a very good overview of its uses and capabilites. The man page has more details. It can be used in variety of situations. One good example is building of linux kernel in text mode. Linux kernel uses a modified version of dialog tailored for its needs.
dialog was initially designed to be used with shell scripts. If you want to use its functionality in a c program, then you can use libdialog. The documentation regarding this is sparse. Definitive reference is the dialog.h header file which comes with the library. You may need to hack here and there to get the required output. The source is easily customizable. I have used it on a number of occasions by modifying the code.
The perl module Curses, Curses::Form and Curses::Widgets give access to curses from perl. If you have curses and basic perl is installed, you can get these modules from CPAN All Modules page. Get the three zipped modules in the Curses category. Once installed you can use these modules from perl scripts like any other module. For more information on perl modules see perlmod man page. The above modules come with good documentation and they have some demo scripts to test the functionality. Though the widgets provided are very rudimentary, these modules provide good access to curses library from perl.
For more information see man pages Curses(3) , Curses::Form(3) and Curses::Widgets(3). These pages are installed only when the above modules are acquired and installed.
This section contains few programs written by me just for fun. They don't signify a better programming practice or the best way of using ncurses. They are provided here so as to allow beginners to get ideas and add more programs to this section. If you have written a couple of nice, simple programs in curses and want them to included here, contact me.
Game of life is a wonder of math. In Paul Callahan's words
This program starts with a simple inverted U pattern and shows how wonderful life works. There is a lot of room for improvement in the program. You can let the user enter pattern of his choice or even take input from a file. You can also change rules and play with a lot of variations. Search on google for interesting information on game of life.
File Path: JustForFun/life.c
Magic Square, another wonder of math, is very simple to understand but very difficult to make. In a magic square sum of the numbers in each row, each column is equal. Even diagnol sum can be equal. There are many variations which have special properties.
This program creates a simple magic square of odd order.
File Path: JustForFun/magic.c
The famous towers of hanoi solver. The aim of the game is to move the disks on the first peg to last peg, using middle peg as a temporary stay. The catch is not to place a larger disk over a small disk at any time.
File Path: JustForFun/hanoi.c
The objective of the famous N-Queen puzzle is to put N queens on a N X N chess board without attacking each other.
This program solves it with a simple backtracking technique.
File Path: JustForFun/queens.c
A fun game, if you have time to kill.
File Path: JustForFun/shuffle.c
A simple typing tutor, I created more out of need than for ease of use. If you know how to put your fingers correctly on the keyboard, but lack practice, this can be helpful.
File Path: JustForFun/tt.c