docs/APRDesign.html

<HTML>
<HEAD><TITLE>APR Design Document</TITLE></HEAD>
<BODY>
<h1>Design of APR</h1>

<p>The Apache Portable Run-time libraries have been designed to provide a common
interface to low level routines across any platform.  The original goal of APR
was to combine all code in Apache to one common code base.  This is not the
correct approach however, so the goal of APR has changed.  There are places 
where common code is not a good thing.  For example, how to map requests 
to either threads or processes should be platform specific.  APR's place 
is now to combine any code that can be safely combined without sacrificing 
performance.</p>

<p>To this end we have created a set of operations that are required for cross
platform development.  There may be other types that are desired and those
will be implemented in the future.</p>

<p>This document will discuss the structure of APR, and how best to contribute
code to the effort.</p>

<h2>APR On Windows and Netware</h2>

<p>APR on Windows and Netware is different from APR on all other systems, 
because those platforms don't use autoconf. On Unix, apr_private.h (private to 
APR) and apr.h (public, used by applications that use APR) are generated by 
autoconf from acconfig.h and apr.h.in respectively. On Windows (and Netware), 
apr_private.h and apr.h are created from apr_private.hw (apr_private.hwn) 
and apr.hw (apr.hwn) respectively.</p>

<p> <strong>
        If you add code to acconfig.h or tests to configure.in or aclocal.m4,
        please give some thought to whether or not Windows and Netware need 
	these additions as well.  A general rule of thumb, is that if it is 
	a feature macro, such as APR_HAS_THREADS, Windows and Netware need it.
	In other words, if the definition is going to be used in a public APR 
	header file, such as apr_general.h, Windows needs it.
        
        The only time it is safe to add a macro or test without also adding 
        the macro to apr*.h[n]w, is if the macro tells APR how to build.  For
        example, a test for a header file does not need to be added to Windows.
</strong></p>

<h2>APR Features</h2>

<p>One of the goals of APR is to provide a common set of features across all 
platforms.  This is an admirable goal, it is also not realistic.  We cannot
expect to be able to implement ALL features on ALL platforms.  So we are
going to do the next best thing.  Provide a common interface to ALL APR 
features on MOST platforms.</p>

<p>APR developers should create FEATURE MACROS for any feature that is not
available on ALL platforms.  This should be a simple definition which has
the form:</p>

<code>APR_HAS_FEATURE</code>

<p>This macro should evaluate to true if APR has this feature on this platform.
For example, Linux and Windows have mmap'ed files, and APR is providing an
interface for mmapp'ing a file.  On both Linux and Windows, APR_HAS_MMAP
should evaluate to one, and the ap_mmap_* functions should map files into
memory and return the appropriate status codes.</p>

<p>If your OS of choice does not have mmap'ed files, APR_HAS_MMAP should 
evaluate to zero, and all ap_mmap_* functions should not be defined.  The 
second step is a precaution that will allow us to break at compile time if a 
programmer tries to use unsupported functions.</p>

<h2>APR types</h2>

<p>The base types in APR</p>

<ul>
<li>dso<br>
	Shared library routines
<li>mmap<br>
	Memory-mapped files
<li>poll<br>
	Polling I/O
<li>time<br>
	Time
<li>user<br>
	Users and groups
<li>locks<br>
	Process and thread locks (critical sections)
<li>shmem<br>
	Shared memory
<li>file_io<br>
	File I/O, including pipes
<li>atomic<br>
	Atomic integer operations
<li>strings<br>
	String handling routines
<li>memory<br>
	Pool-based memory allocation
<li>passwd<br>
	Reading passwords from the terminal
<li>tables<br>
	Tables and hashes
<li>network_io<br>
	Network I/O
<li>threadproc<br>
	Threads and processes
<li>misc<br>
	Any APR type which doesn't have any other place to belong.  This
	should be used sparingly.
<li>support<br>
	Functions meant to be used across multiple APR types.  This area
	is for internal functions only.  If a function is exposed, it should
	not be put here.
</ul>

<h2>Directory Structure</h2>

<p>Each type has a base directory.  Inside this base directory, are
subdirectories, which contain the actual code.  These subdirectories are named
after the platforms the are compiled on.  Unix is also used as a common
directory.  If the code you are writing is POSIX based, you should look at the
code in the unix directory.  A good rule of thumb, is that if more than half
your code needs to be ifdef'ed out, and the structures required for your code
are substantively different from the POSIX code, you should create a new
directory.</p>

<p>Currently, the APR code is written for Unix, BeOS, Windows, and OS/2.  An
example of the directory structure is the file I/O directory:</p>

<pre>
apr
  |
   ->  file_io
          |
           -> unix            The Unix and common base code
          |
           -> win32           The Windows code
          | 
           -> os2             The OS/2 code
</pre>

<p>Obviously, BeOS does not have a directory.  This is because BeOS is currently
using the Unix directory for it's file_io.</p>

<p>There are a few special top level directories.  These are test and include.
Test is a directory which stores all test programs.  It is expected
that if a new type is developed, there will also be a new test program, to
help people port this new type to different platforms.  A small document
describing how to create new tests that integrate with the test suite can be
found in the test/ directory.  Include is a directory which stores all 
required APR header files for external use.</p>

<h2>Creating an APR Type</h2>

<p>The current design of APR requires that most APR types be incomplete.  
It is not possible to write flexible portable code if programs can access 
the internals of APR types.  This is because different platforms are 
likely to define different native types.  There are only two exceptions to
this rule:</p>

<ul>
<li>The first exception to this rule is if the type can only reasonably be 
implemented one way.  For example, time is a complete type because there 
is only one reasonable time implementation.

<li>The second exception to the incomplete type rule can be found in 
apr_portable.h.  This file defines the native types for each platform.  
Using these types, it is possible to extract native types for any APR type.</p>
</ul>

<p>For this reason, each platform defines a structure in their own directories. 
Those structures are then typedef'ed in an external header file.  For example
in file_io/unix/fileio.h:</p>

<pre>
    struct ap_file_t {
        apr_pool_t *cntxt;
        int filedes;
        FILE *filehand;
        ...
    }
</pre>

<p>In include/apr_file_io.h:</p>
    </pre>
    typedef struct ap_file_t    ap_file_t;
    </pre>

<p> This will cause a compiler error if somebody tries to access the filedes 
field in this structure.  Windows does not have a filedes field, so obviously, 
it is important that programs not be able to access these.</p>

<p>You may notice the apr_pool_t field.  Most APR types have this field.  This
type is used to allocate memory within APR.  Because every APR type has a pool,
any APR function can allocate memory if it needs to.  This is very important
and it is one of the reasons that APR works.  If you create a new type, you
must add a pool to it.  If you do not, then all functions that operate on that
type will need a pool argument.</p>

<h2>New Function</h2>

<p>When creating a new function, please try to adhere to these rules.</p>

<ul>
<li>  Result arguments should be the first arguments.
<li>  If a function needs a pool, it should be the last argument.
<li>  These rules are flexible, especially if it makes the code easier
      to understand because it mimics a standard function.
</ul>

<h2>Documentation</h2>

<p>Whenever a new function is added to APR, it MUST be documented.  New 
functions will not be committed unless there are docs to go along with them.
The documentation should be a comment block above the function in the header
file.</p>

<p>The format for the comment block is:</p>

<pre>
    /**
     * Brief description of the function
     * @param parma_1_name explanation
     * @param parma_2_name explanation
     * @param parma_n_name explanation
     * @tip Any extra information people should know.
     * @deffunc function prototype if required
     */ 
</pre>

<p>For an actual example, look at any file in the include directory.  The 
reason the docs are in the header files is to ensure that the docs always
reflect the current code.  If you change parameters or return values for a
function, please be sure to update the documentation.</p>

<h2>APR Error reporting</h2>

<p>Most APR functions should return an ap_status_t type.  The only time an
APR function does not return an ap_status_t is if it absolutely CAN NOT
fail.  Examples of this would be filling out an array when you know you are
not beyond the array's range.  If it cannot fail on your platform, but it
could conceivably fail on another platform, it should return an ap_status_t.
Unless you are sure, return an ap_status_t.</p>

<strong>
        This includes functions that return TRUE/FALSE values.  How that 
        is handled is discussed below
</strong>

<p>All platforms return errno values unchanged.  Each platform can also have
one system error type, which can be returned after an offset is added.  
There are five types of error values in APR, each with its own offset.</p>

<!--  This should be turned into a table, but I am lazy today -->
<pre>
    Name			Purpose
0) 			This is 0 for all platforms and isn't really defined
 			anywhere, but it is the offset for errno values.
			(This has no name because it isn't actually defined, 
                        but for completeness we are discussing it here).

1) APR_OS_START_ERROR	This is platform dependent, and is the offset at which
			APR errors start to be defined.  Error values are 
			defined as anything which caused the APR function to 
			fail.  APR errors in this range should be named 
			APR_E* (i.e. APR_ENOSOCKET)

2) APR_OS_START_STATUS	This is platform dependent, and is the offset at which
			APR status values start.  Status values do not indicate
			success or failure, and should be returned if 
			APR_SUCCESS does not make sense.  APR status codes in 
			this range should be name APR_* (i.e. APR_DETACH)

4) APR_OS_START_USEERR	This is platform dependent, and is the offset at which
			APR apps can begin to add their own error codes.

3) APR_OS_START_SYSERR	This is platform dependent, and is the offset at which
			system error values begin.
</pre>

<strong>The difference in naming between APR_OS_START_ERROR and 
APR_OS_START_STATUS mentioned above allows programmers to easily determine if
the error code indicates an error condition or a status condition.</strong>

<p>If your function has multiple return codes that all indicate success, but
with different results, or if your function can only return PASS/FAIL, you 
should still return an apr_status_t.  In the first case, define one
APR status code for each return value, an example of this is
<code>apr_proc_wait</code>, which can only return APR_CHILDDONE, 
APR_CHILDNOTDONE, or an error code.  In the second case, please return 
APR_SUCCESS for PASS, and define a new APR status code for failure, an 
example of this is <code>apr_compare_users</code>, which can only return
APR_SUCCESS, APR_EMISMATCH, or an error code.</p>

<p>All of these definitions can be found in apr_errno.h for all platforms.  When
an error occurs in an APR function, the function must return an error code.
If the error occurred in a system call and that system call uses errno to
report an error, then the code is returned unchanged.  For example: </p>

<pre>
    if (open(fname, oflags, 0777) < 0)
        return errno;
</pre>

<p>The next place an error can occur is a system call that uses some error value
other than the primary error value on a platform.  This can also be handled
by APR applications.  For example:</p>

<pre>
    if (CreateFile(fname, oflags, sharemod, NULL, 
                   createflags, attributes, 0) == INVALID_HANDLE_VALUE
        return (GetLAstError() + APR_OS_START_SYSERR);
</pre>

<p>These two examples implement the same function for two different platforms.
Obviously even if the underlying problem is the same on both platforms, this
will result in two different error codes being returned.  This is OKAY, and
is correct for APR.  APR relies on the fact that most of the time an error
occurs, the program logs the error and continues, it does not try to
programatically solve the problem.  This does not mean we have not provided
support for programmatically solving the problem, it just isn't the default
case.  We'll get to how this problem is solved in a little while.</p>

<p>If the error occurs in an APR function but it is not due to a system call,
but it is actually an APR error or just a status code from APR, then the
appropriate code should be returned.  These codes are defined in apr_errno.h
and should be self explanatory.</p>

<p>No APR code should ever return a code between APR_OS_START_USEERR and 
APR_OS_START_SYSERR, those codes are reserved for APR applications.</p>

<p>To programmatically correct an error in a running application, the error 
codes need to be consistent across platforms.  This should make sense.  APR
has provided macros to test for status code equivalency.  For example, to
determine if the code that you received from the APR function means EOF, you
would use the macro APR_STATUS_IS_EOF().</p>

<p>Why did APR take this approach?  There are two ways to deal with error 
codes portably.</p>

<ol type=1>
<li>  Return the same error code across all platforms.
<li>  Return platform specific error codes and convert them when necessary.  
</ol>

<p>The problem with option number one is that it takes time to convert error 
codes to a common code, and most of the time programs want to just output 
an error string.  If we convert all errors to a common subset, we have four 
steps to output an error string:</p>

<p>The second problem with option 1, is that it is a lossy conversion.  For
example, Windows and OS/2 have a couple hundred error codes, but POSIX errno
only defines about 50 errno values.  This means that if we convert to a
canonical error value immediately, there is no way for the programmer to
get the actual system error.</p>

<pre>
    make syscall that fails
        convert to common error code                 step 1
        return common error code
            check for success
            call error output function               step 2
                convert back to system error         step 3
                output error string                  step 4
</pre>

<p>By keeping the errors platform specific, we can output error strings in two
steps.</p>

<pre>
    make syscall that fails
        return error code
            check for success
            call error output function               step 1
                output error string                  step 2
</pre>

<p>Less often, programs change their execution based on what error was returned.
This is no more expensive using option 2 than it is using option 1, but we
put the onus of converting the error code on the programmer themselves.
For example, using option 1:</p>

<pre>
    make syscall that fails
        convert to common error code
        return common error code
            decide execution based on common error code
</pre>

<p>Using option 2:</p>
    
<pre>
    make syscall that fails
        return error code
            convert to common error code (using ap_canonical_error)
            decide execution based on common error code
</pre>

<p>Finally, there is one more operation on error codes.  You can get a string
that explains in human readable form what has happened.  To do this using 
APR, call ap_strerror().</p>