The Be Book The Support Kit The Support Kit Index

Functions and Macros

Library: libbe.so
Summary:  more...


This section lists the Support Kit's general-purpose functions and macros.


add_system_beep_event()  see beep()


atomic_add() , atomic_and() , atomic_or()

Declared in:  be/support/SupportDefs.h


int32 atomic_add(vint32 *atomicVariable, int32 addValue)
int32 atomic_and(vint32 *atomicVariable, int32 andValue)
int32 atomic_or(vint32 *atomicVariable, int32 orValue)

These functions perform the named operations (addition, bitwise AND, or bitwise OR) on the 32-bit value found in atomicVariable, thus:

*atomicVariable += addValue
*atomicVariable &= andValue
*atomicVariable |= orValue

Each function returns the previous value of the vint32 variable that atomicVariable points to (in other words, they each return the value that was in *atomicVariable before the operation was performed).

These functions are guaranteed to be atomic: if two threads attempt to access the same atomic variable at the same time (through these functions), one of the two threads will be made to wait until the other thread has completed the operation and updated the atomicVariable value.


beep() , system_beep() , add_system_beep_event()

Declared in:  be/support/Beep.h

status_t beep(void)
status_t system_beep(const char *eventName)
status_t add_system_beep_event(const char *eventName,
      uint32 flags = _BEEP_FLAGS)

beep() produces the basic system beep. This function engages the Media Server, but doesn't wait for the sound to play. If it can't contact the server to play the beep, it returns B_ERROR. If it can make contact but can't get a satisfactory reply back from the server, it returns B_BAD_REPLY. Otherwise, it returns B_OK.

system_beep() plays the system beep configured for the event specified by the given eventName.

Applications can add new system beep events by calling add_system_beep_event(), passing the name to assign the event; the flags argument is currently unused and should be zero. Once this has been done, the user can use the Sounds preference application to configure a sound for the event.

See also: play_sound() in the Media Kit


cast_as()  see class_name()


class_name() , is_instance_of() , is_kind_of() , cast_as()

Declared in:  be/support/ClassInfo.h

const char *class_name(object)
bool is_instance_of(object, class)
bool is_kind_of(object, class)
class *cast_as(object, class)

These macros deliver information about an object's type, including the name of its class and its standing in the class hierarchy. In each case, the object argument is a pointer to an object; it can be an object of any type (it doesn't have to descend from any particular class). The class argument is a class name—not a string such as "BApplication", but the type name itself (literally BApplication).

class_name() returns a pointer to the name of object's class.

is_instance_of() returns true if object is an instance of class, and false otherwise.

is_kind_of() returns true if object is an instance of class or an instance of any class that inherits from class, and false if not.

cast_as() returns a pointer to object cast as a pointer to an object of class, but only if object is a kind of class. If not, object cannot be safely cast as a pointer to class, so cast_as() returns NULL.

For example, given this slice of the inheritance hierarchy from the Interface Kit,

and code like this that creates an instance of the BPictureButton class,

BButton *anObject = new BPictureButton(...);

the first three macros would work as follows:

const char *s = class_name(anObject);

if ( is_instance_of(anObject, BView) )
   printf("The object is an instance of BView.n");

if ( is_kind_of(anObject, BView) )
   printf("The object is a kind of BView.n");

Note that class names are not passed as strings, but class_name() returns the name as a string.

The cast_as() macro is most useful when you want to treat a generic object as an instance of a more specific class. Suppose, for example, that the BPictureButton mentioned above becomes the focus view for a window and you retrieve it by calling the BWindow's CurrentFocus() function:

BView *focus = myWindow->CurrentFocus();

Since the focus view might be any type of view, CurrentFocus() returns a pointer to an object of the base BView class. Unless you know otherwise, you cannot treat the object as anything more specific than a BView instance. However, you can ask the object if it's a kind of BPictureButton and, if it is, cast it to the BPictureButton type:

if ( is_kind_of(focus, BPictureButton) ) {
   BPictureButton *picbutton = (BPictureButton *)focus);
   if ( picbutton->Behavior() == B_TWO_STATE_BUTTON )
      . . .
}

The cast_as() macro does the same thing, but more efficiently. It casts the object to the target class if it is safe to do so—if the object is an instance of a class that inherits from the target class or an instance of the target class itself—and returns NULL if not.

BPictureButton *picbutton = cast_as(focus, BPictureButton);
if ( picbutton ) {
   if ( picbutton->Behavior() == B_TWO_STATE_BUTTON )
      . . .
}

cast_as() is often used in place of the cast operator to assure code safety even where an expected result is anticipated and there's no need for an intermediate variable (like focus):

BPictureButton *picbutton =
            cast_as(myWindow->CurrentFocus(), BPictureButton);
if ( picbutton ) {
      . . .
}

The cast_as() and is_kind_of() macros work alike; they're both based on the C++ dynamic_cast operator and they reflect its behavior. To describe that behavior more precisely, let's adopt the following shorthand terms for an object's type:

Either of these types can be compared to a target type, the type you want to cast the object to or test it against. The target type is the class argument passed to the macros.

In the best of all possible worlds, you'd want to ignore the declared type of an object and compare only the real type to the target type. However, the dynamic_cast operator—and by extension cast_as() and is_kind_of()—considers the real type only if it has to. It first compares the object's declared type to the target type. It assumes that the declared type is accurate (that the object is truly the kind of object it's represented to be) and it summarily handles the obvious cases: If the target type is the same as the declared type or if it's a class that the declared type inherits from, the operation will succeed. Consequently, cast_as() will cast the object to the target type and is_kind_of() will return true, regardless of the object's real type. In other words, if the target class is above or at the same level as the declared class in the inheritance hierarchy, the real class is ignored.

However, if the declared type doesn't match or derive from the target type, dynamic_cast and the macros look at the real type: If the target class is identical to the real type, or if it's a class that the real type derives from, the operation succeeds. If not, it fails.

Therefore, the is_kind_of() and cast_as() macros will produce reliable results as long as objects are not arbitrarily cast to types that may not be accurate. For example, you should not cast an object to a target type and then attempt to use is_kind_of() to determine if the cast was correct. This code is unreliable:

BPictureButton *picbutton =
                  (BPictureButton *)myWindow->CurrentFocus();
if ( is_kind_of(picbutton, BPictureButton) ) {
      . . .
}

In this example, is_kind_of() will always return true, no matter what the class of the current focus view. The general rule is that the declared type of an object must always be accurate; an object should be typed only to its own class or to a class that it inherits from. The macros cannot rescue you from an inaccurate cast.


convert_to_utf8() , convert_from_utf8()

Declared in:  be/support/UTF8.h

status_t convert_to_utf8(uint32 sourceEncoding,
      const char *source,
      int32 *sourceLength,
      char *destination,
      int32 *destinationLength,
      int32 *state,
      char substitute = B_SUBSTITUTE)
status_t convert_from_utf8(uint32 destinationEncoding,
      const char *source,
      int32 *sourceLength,
      char *destination,
      int32 *destinationLength,
      int32 *state,
      char substitute = B_SUBSTITUTE)

These functions convert text to and from the Unicode UTF-8 encoding that's standard for the Be operating system and is assumed in most contexts. UTF-8 is described under "Character Encoding" in "Responding to the User" section of The Interface Kit chapter.

convert_to_utf8() permits you to take text that's encoded according to another standard and convert it to UTF-8 for the BeOS. convert_from_utf8() lets you convert text from UTF-8 to other encodings for other venues (for example, to the encodings commonly used for displaying text on the World Wide Web).

The first argument passed to these functions names the other encoding—the source encoding for convert_to_utf8() and the destination encoding for convert_from_utf8(). It can be any of the following constants:

B_ISO1_CONVERSION
B_ISO2_CONVERSION
B_ISO3_CONVERSION
B_ISO4_CONVERSION
B_ISO5_CONVERSION
B_ISO6_CONVERSION
B_ISO7_CONVERSION
B_ISO8_CONVERSION
B_ISO9_CONVERSION
B_ISO10_CONVERSION
B_MAC_ROMAN_CONVERSION
B_SJIS_CONVERSION
B_EUC_CONVERSION
B_JIS_CONVERSION
B_MS_WINDOWS_CONVERSION
B_UNICODE_CONVERSION
B_KOI8R_CONVERSION
B_MS_WINDOWS_1251_CONVERSION
B_MS_DOS_866_CONVERSION

Most of these constants designate encoding schemes that are supported by the BFont class in the Interface Kit and its SetEncoding() function. They parallel the constants that are passed to that function. For example, B_ISO1_CONVERSION (for these functions) and B_ISO_8859_1 (for SetEncoding()) both designate the extended ASCII encoding defined in part one of ISO 8859 (Latin 1). Similarly, B_ISO2_CONVERSION matches B_ISO_8859_2, B_ISO3_CONVERSION matches B_ISO_8859_3, and so on. B_MAC_ROMAN_CONVERSION matches B_MACINTOSH_ROMAN. (B_ISO10_CONVERSION is not implemented for this release.)

B_SJIS_CONVERSION stands for the Shift-JIS (Japanese Industrial Standard) encoding of Japanese and B_EUC_CONVERSION stands for the EUC (Extended UNIX Code) encoding of Japanese in packed format.

Both functions convert up to sourceLength bytes of text from the source buffer. They write up to destinationLength bytes of converted text into the destination buffer. The amount of text that they actually convert is therefore constrained both by the amount of source text (sourceLength) and the capacity of the output buffer (destinationLength). Neither function stops at a null terminator ('0') when reading the input buffer nor adds one to the text in the output buffer; they depend only on sourceLength and destinationLength for guidance.

When finished, these functions modify the variable that sourceLength refers to so that it reports the number of bytes of source text actually converted. They also modify the variable that destinationLength refers to so that it reports the number of bytes actually written to the destination buffer. Neither function will stop in the middle of a multibyte source character; they're guaranteed to convert only full characters.

The state argument serves as a cookie that lets you use multiple calls to these functions to convert a batch of text (if, for example, the text is stored in multiple buffers). Pass 0 for the first buffer, and pass the value returned in state to the next call to continue processing the same text; this will cause the conversion to continue where it left off. For example:

/* buffer is a pointer to a small source buffer */
/* destBuffer is a pointer to a small destination buffer */
/* destBufferLen is the size of the destination buffer in bytes */
int32 state = 0;
int32 bufferLen;

while ((bufferLen = GetNextBuffer(file, buffer))) {
   convert_to_utf8(buffer, &bufferLen, destBuffer, &destBufferLen, &state);
   /* do stuff with the buffer in destBuffer, which has the converted text */
}

In this example, text is fetched using a call to a function called GetNextBuffer(),whose implementation depends on the application's needs. Each buffer is converted, and the converted text is stored in destBuffer, which can then be used by the application.

If either function encounters a character in the source that the destination format doesn't allow, it puts the character specified by the substitute argument (or a question mark ('?') if substitute isn't specified) in its place in the output text. This is much more likely to occur when converting from UTF-8 than when converting to it, since Unicode represents a very large number of characters.

If successful in converting at least one source character, both functions return B_OK. If unsuccessful, for example, if they don't recognize the source or destination encoding, they return B_ERROR. If there's an error, you should not trust any of the output arguments.

These functions are found in the libtextencoding.so library. If you use them, be sure to include this library file.

See also: BFont::SetEncoding(), "Character Encoding" in the "Responding to the User" section of the The Interface Kit chapter


find_instantiation_func()

Declared in:  be/support/Archivable.h

instantiation_func find_instantiation_func(const char *className)
instantiation_func find_instantiation_func(BMessage *archive)

Returns a pointer to the Instantiate() function that can create instances of the className class, or NULL if the function can't be found. If passed a BMessage archive, find_instantiation_func() gets the name of the class from a B_STRING_TYPE field called "class" in the BMessage.

The instantiation_func type is defined as follows:


BArchivable *(*instantiation_func) (BMessage *)

In other words, the function has the same syntax as the Instantiate() function declared in the BArachivable class and replicated in derived classes (with class-specific return values).

The function that's returned can be called like any C function; you don't need the class name or another object of the class. For example:

instantiation_func func;
if ( func = find_instantiation_func(arhiveMessage) ) {
   BArchivable *object = func(archiveMessage);
}

instantiate_object() will do this work for you.

See also: BArchivable::Instantiate(), instantiate_object()


instantiate_object()

Declared in:  be/support/Archivable.h

BArchivable *instantiate_object(BMessage *archive)

Creates and returns a new instance of an archived object, or returns NULL if the object can't be constructed. The object is created by calling the Instantiate() function of...

1.    ...the class that was last added to the "class" field (an array of class names) of the archive BMessage. This will be the "most derived" class in the array.

2.    If the named class isn't recognized by the app (or doesn't define Instantiate()), the function tries to load the add-on that's identified (by signature) in archive's "add_on" field. This add-on should contain the code for the class.

3.    If the add-on method doesn't work, instantiate_object() tries the next class in the "class" array, and so works its way up the class hierarchy.

4.    If the appropriate class is never found, the function returns NULL.

When successful, instantiate_object() returns the object that Instantiate() created, but typed to the base BArchivable class. The cast_as() macro can type it to a the proper class.

BArchivable *base = instantiate_object(archive);
if ( base ) {
   TheClass *object = cast_as(base, TheClass);
   if ( object ) {
      . . .
   }
}

See also: the BArchivable class, find_instantiation_func()


is_instance_of()  see class_name()
is_kind_of()  see class_name()
is_type_swapped()  see swap_data()


swap_data() , is_type_swapped()

Declared in:  be/support/ByteSwap.h

status_t swap_data(type_code type, void *data, size_t length, swap_action action)
bool is_type_swapped(type_code type)

swap_data is a general byte swapping function, converting data of type type and converting its endianness according to action (defined in "Byte Swapping Constants"). The length field can be used to specify an array of items to be converted. For example, you can convert an array of BMessengers from little endian to the host endianness:

BMessenger messengers[10];
...
swap_data(B_MESSENGER_TYPE, messengers, 10*sizeof(BMessengers)
         B_SWAP_LENDIAN_TO_HOST);

The function can swap most data types with type constants defined in <be/support/TypeConstants.h>. It returns B_OK on success and B_BAD_VALUE on failure.

is_type_swapped takes a type code and determines whether or not it has been swapped. This only works for types defined in <be/support/TypeConstants.h>. It returns true if the type code is in the host's native endianness and false otherwise.

See also: "Byte Swapping Constants"


min() , max() , min_c() , max_c()

Declared in:  be/support/SupportDefs.h

min(a, b)
min_c(a, b)
max(a, b)
max_c(a, b)

These macros compare two integers or floating-point numbers. min() and min_c() return the lesser of the two (or b if they're equal); max() and max_c() return the greater of the two (or a if they're equal). min() and max() can only be used in straight C programs. min_c() and max_c() are defined for C and C++.


system_  see beep()


validate_instantiation()

Declared in:  be/support/Archivable.h

bool validate_instantiation(BMessage *archive, const char *className)

Returns true if the archive BMessage contains data for an object belonging to the className class, and false if not. The determination is made by looking for the class name in a "class" array in the archive. If the class name appears anywhere in the array, this function returns true. If not, it returns false.


The Be Book The Support Kit The Support Kit Index

The Be Book,
...in lovely HTML...
for BeOS Release 5.

Copyright © 2000 Be, Inc. All rights reserved..