<!-- {{{ Banner                 -->

<!-- =============================================================== -->
<!--                                                                 -->
<!--     package.sgml                                                -->
<!--                                                                 -->
<!--     How to write a package.                                     -->
<!--                                                                 -->
<!-- =============================================================== -->
<!-- ####ECOSDOCCOPYRIGHTBEGIN####                                       -->
<!-- ===============================================================     -->
<!-- Copyright (C) 2000, 2001, 2002 Free Software Foundation, Inc.       -->
<!-- This material may be distributed only subject to the terms          -->
<!-- and conditions set forth in the Open Publication License, v1.0      -->
<!-- or later (the latest version is presently available at              -->
<!-- http://www.opencontent.org/openpub/)                                -->
<!-- Distribution of the work or derivative of the work in any           -->
<!-- standard (paper) book form is prohibited unless prior               -->
<!-- permission obtained from the copyright holder                       -->
<!-- ===============================================================     -->
<!-- ####ECOSDOCCOPYRIGHTEND####                                         -->
<!-- =============================================================== -->
<!-- #####DESCRIPTIONBEGIN####                                       -->
<!--                                                                 -->
<!-- Author(s):   bartv                                              -->
<!-- Contact(s):  bartv                                              -->
<!-- Date:        2000/02/06                                         -->
<!-- Version:     0.01                                               -->
<!--                                                                 -->
<!-- ####DESCRIPTIONEND####                                          -->
<!-- =============================================================== -->

<!-- }}} -->

<chapter id="package">
<title>Package Organization</title>

<!-- {{{ Introit                -->

<para>
For a package to be usable in the &eCos; component framework it must
conform to certain rules imposed by that framework. Packages must be
distributed in a form that is understood by the component repository
administration tool. There must be a top-level &CDL; script which
describes the package to the component framework. There are certain
limitations related to how a package gets built, so that the package
can still be used in a variety of host environments. In addition to
these rules, the component framework provides a number of guidelines.
Packages do not have to conform to the guidelines, but sticking to
them can simplify certain operations.
</para>
<para>
This chapter deals with the general organization of a package, for
example how to distinguish between private and exported header files.
<xref linkend="language"> describes the &CDL; language.
<xref linkend="build"> details the build process.
</para>

<!-- }}} -->
<!-- {{{ Repository             -->

<sect1 id="package.hierarchy">
<title>Packages and the Component Repository</title>

<para>
All &eCos; installations include a component repository. This is a
directory structure for all installed packages. The component
framework comes with an administration tool that allows new packages
or new versions of a package to be installed, old packages to be
removed, and so on. The component repository includes a simple
database, maintained by the administration tool, which contains
details of the various packages.
</para>

<informalfigure PgWide=1>
<mediaobject>
<imageobject>
<imagedata fileref="repo.png" Scalefit=1 Align="Center">
</imageobject>
</mediaobject>
</informalfigure>

<para>
Each package has its own little directory hierarchy within the
component repository. Keeping several packages in a single directory
is illegal. The error, infra and kernel packages all live at the
top-level of the repository. For other types of packages there are
some pre-defined directories: <filename
class="directory">compat</filename> is used for compatibility
packages, which implement other interfaces such as &uITRON; or POSIX
using native &eCos; calls; <filename class="directory">hal</filename>
is used for packages that port &eCos; to different architectures or
platforms, and this directory is further organized on a
per-architecture basis; <filename class="directory">io</filename> is
intended for device drivers; <filename
class="directory">language</filename> is used for language support
libraries, for example the C library. There are no strict rules
defining where new packages should get installed. Obviously if an
existing top-level directory such as <filename
class="directory">compat</filename> is applicable then the new package
should go in there. If a new category is desirable then it is possible
to create a new sub-directory in the component repository. For
example, an organization planning to release a number of &eCos;
packages may want them all to appear below a sub-directory
corresponding to the organization's name&nbsp;&mdash; in the hope that
the name will not change too often. It is possible to add new packages
directly to the top-level of the component repository, but this should
be avoided.
</para>

<para>
The <database>ecos.db</database> file holds the component repository
database and is managed by the administration tool. The various
configuration tools read in this file when they start-up to obtain
information about the various packages that have been installed. When
developing a new package it is necessary to add some information to
the file, as described in <xref linkend="language.database">. The
<filename class="directory">templates</filename> directory holds
various configuration templates.
</para>

<note>
<para>
Earlier releases of &eCos; came with two separate files,
<filename>targets</filename> and <filename>packages</filename>. The
<database>ecos.db</database> database replaces both of these.
</para>
</note>

<caution>
<para>
The current <database>ecos.db</database> database does not yet provide
all of the information needed by the component framework. Its format
is subject to change in future releases, and the file may be replaced
completely if necessary. There are a number of other likely future
developments related to the component repository and the database. The
way targets are described is subject to change. Sometimes it is
desirable for component writers to do their initial development in a
directory outside the component repository, but there is no specific
support in the framework for that yet.
</para>
</caution>

</sect1>

<!-- }}} -->
<!-- {{{ Versioning             -->

<sect1 id="package.versions">
<title>Package Versioning</title>
<para>
Below each package directory there can be one or more version
sub-directories, named after the versions. This is a requirement of
the component framework: it must be possible for users to install
multiple versions of a package and select which one to use for any
given application. This has a number of advantages to users: most
importantly it allows a single component repository to be shared
between multiple users and multiple projects, as required; also it
facilitates experiments, for example it is relatively easy to try out
the latest version of some package and see if it makes any difference.
There is a potential disadvantage in terms of disk space. However
since &eCos; packages generally consist of source code intended for
small embedded systems, and given typical modern disk sizes, keeping a
number of different versions of a package installed will usually be
acceptable. The administration tool can be used to remove versions
that are no longer required.
</para>

<informalfigure PgWide=1>
<mediaobject>
<imageobject>
<imagedata fileref="version.png" Scalefit=1 Align="Center">
</imageobject>
</mediaobject>
</informalfigure>

<para>
The version <filename>current</filename> is special. Typically it
corresponds to the very latest version of the sources, obtained by
anonymous &CVS;. These sources may change frequently, unlike full
releases which do not change (or only when patches are produced).
Component writers may also want to work on the
<filename>current</filename> version.
</para>
<para>
All other subdirectories of a package correspond to specific releases
of that package. The component framework allows users to select the
particular version of a package they want to use, but by default the
most recent one will be used. This requires some rules for ordering
version numbers, a difficult task because of the wide variety of ways
in which versions can be identified.
</para>

<orderedlist>
<listitem>
<para>
The version <filename>current</filename> is always considered to be
the most recent version.
</para
</listitem>

<listitem>
<para>
If the first character of both strings are either <literal>v</literal>
or <literal>V</literal>, these are skipped because it makes little
sense to enforce case sensitivity here. Potentially this could result
in ambiguity if there are two version directories
<literal>V1.0</literal> and <literal>v1.0</literal>, but this will
match the confusion experienced by any users of such a package.
However if two subsequent releases are called <literal>V1.0</literal>
and <literal>v1.1</literal>, e.g. because of a minor mix-up when
making the distribution file, then the case difference is ignored.
</para>
</listitem>

<listitem>
<para>
Next the two version strings are compared one character at a time.
If both strings are currently at a digit then a string to number
conversion takes place, and the resulting numbers are compared.
For example <literal>v10</literal> is a more recent release than
<literal>v2</literal>. If the two numbers are the same then processing
continues, so for <literal>v2b</literal> and <literal>v2c</literal>
the version comparison code would move on to <literal>b</literal> and
<literal>c</literal>. 
</para>
</listitem>

<listitem>
<para>
The characters dot <literal>.</literal>, hyphen <literal>-</literal>
and underscore <literal>_</literal> are treated as equivalent
separators, so if one release goes out as <literal>v1_1</literal> and
the next goes out as <literal>v1.2</literal> the separator has no
effect.
</para>
</listitem>

<listitem>
<para>
If neither string has yet terminated but the characters are different,
ASCII comparison is used. For example <literal>V1.1b</literal> is
more recent than <literal>v1.1alpha</literal>.
</para>
</listitem>

<listitem>
<para>
If one version string terminates before the other, the current
character determines which is the more recent. If the other string is
currently at a separator character, for example
<literal>v1.3.1</literal> and <literal>v1.3</literal>, then the former
is assumed to be a minor release and hence more recent than the
latter. If the other string is not at a separator character, for
example <literal>v1.3beta</literal>, then it is treated as an
experimental version of the <literal>v1.3</literal> release and hence
older. 
</para>
</listitem>

<listitem>
<para>
There is no special processing of dates, so with two versions
<literal>ss-20000316</literal> and <literal>ss-20001111</literal>
the numerical values <literal>20001111</literal> and
<literal>20000316</literal> determine the result: larger values are
more recent. It is suggested that the full year be used in such cases
rather than a shorthand like <literal>00</literal>, to avoid
Y2100 problems.
</para>
</listitem>

<listitem>
<para>
There is no limit on how many levels of versioning are used, so
there could in theory be a <literal>v3.1.4.1.5.9.2.7</literal> release
of a package. However this is unlikely to be of benefit to typical
users of a package.
</para>
</listitem>

</orderedlist>

<para>
The version comparison rules of the component framework may not be
suitable for every version numbering scheme in existence, but they
should cope with many common cases.
</para>

<caution>
<para>
There are some issues still to be resolved before it is possible to
combine the <filename>current</filename> sources available via
anonymous &CVS and full releases of &eCos; and additional packages in
a single component repository. The first problem relates to the
<database>ecos.db</database> database: if a new package is added via
the CVS repository then this requires a database update, but the
administration tool is bypassed. The second problem arises if an
organization chooses to place its component repository under source
code control using &CVS;, in which case different directories will
belong to different &CVS; servers. These issues will be addressed in a
future release.
</para>
</caution>

</sect1>

<!-- }}} -->
<!-- {{{ Package contents       -->

<!-- {{{ Introit                -->

<sect1 id="package.contents">
<title>Package Contents and Layout</title>
<para>
A typical package contains the following:
</para>
<orderedlist>
<listitem>
<para>
Some number of source files which will end up in a library. The
application code will be linked with this library to produce an
executable. Some source files may serve other purposes, for example to
provide a linker script.
</para>
</listitem>
<listitem>
<para>
Exported header files which define the interface provided by the
package. 
</para>
</listitem>
<listitem>
<para>
On-line documentation, for example reference pages for each exported
function. 
</para>
</listitem>
<listitem>
<para>
Some number of test cases, shipped in source format, allowing users to
check that the package is working as expected on their particular
hardware and in their specific configuration.
</para>
</listitem>
<listitem>
<para>
One or more &CDL; scripts describing the package to the configuration
system.
</para>
</listitem>
</orderedlist>
<para>
It is also conventional to have a per-package
<filename>ChangeLog</filename> file used to keep track of changes to
that package. This is especially valuable to end users of the package
who may not have convenient access to the source code control system
used to manage the master copy of the package, and hence cannot find
out easily what has changed. Often it can be very useful to the main
developers as well.
</para>
<para>
Any given packages need not contain all of these. It is compulsory to
have at least one &CDL; script describing the package, otherwise the
component framework would be unable to process it. Some packages may
not have any source code: it is possible to have a package that merely
defines a common interface which can then be implemented by several
other packages, especially in the context of device drivers; however
it is still common to have some code in such packages to avoid
replicating shareable code in all of the implementation packages.
Similarly it is possible to have a package with no exported header
files, just source code that implements an existing interface: for
example an ethernet device driver might just implement a standard
interface and not provide any additional functionality. Packages do
not need to come with any on-line documentation, although this may
affect how many people will want to use the package. Much the same
applies to per-package test cases.
</para>

<para>
The component framework has a recommended per-package directory layout
which splits the package contents on a functional basis:
</para>

<informalfigure PgWide=1>
<mediaobject>
<imageobject>
<imagedata fileref="package.png" Scalefit=1 Align="Center">
</imageobject>
</mediaobject>
</informalfigure>

<para>
For example, if a package has an <filename
class="directory">include</filename> sub-directory then the component
framework will assume that all header files in and below that
directory are exported header files and will do the right thing at
build time. Similarly if there is &doc; property indicating the
location of on-line documentation then the component framework will
first look in the <filename class="directory">doc</filename>
sub-directory.
</para>

<para>
This directory layout is just a guideline, it is not enforced by the
component framework. For simple packages it often makes more sense to
have all of the files in just one directory. For example a package
could just contain the files <filename>hello.cxx</filename>,
<filename>hello.h</filename>, <filename>hello.html</filename> and
<filename>hello.cdl</filename>. By default
<filename>hello.h</filename> will be treated as an exported header
file, although this can be overridden with the <link
linkend="ref.include-files">&include-files;</link> property. Assuming
there is a &doc; property referring to <filename>hello.html</filename>
and there is no <filename class="directory">doc</filename>
sub-directory then the tools will search for this file relative to the
package's top-level and everything will just work. Much the same
applies to <filename>hello.cxx</filename> and
<filename>hello.cdl</filename>. 
</para>

<tip>
<para>
Older versions of the &eCos; build system only supported packages that
followed the directory structure exactly. Hence certain core packages
such as <filename>error</filename> implement the full directory
structure, even though that is a particularly simple package and the
full directory structure is inappropriate. Component writers can
decide for themselves whether or not the directory structure
guidelines are appropriate for their package.
</para>
</tip>

<!-- }}} -->
<!-- {{{ Build process          -->

<sect2 id="package.build">
<title>Outline of the Build Process</title>
<para>
The full build process is described in <xref linkend="build">, but 
a summary is appropriate here. A build involves three directory
structures: 
</para>

<orderedlist>
<listitem>
<para>
The component repository. This is where all the package source code is
held, along with &CDL; scripts, documentation, and so on. For build
purposes a component repository is read-only. Application developers
will only modify the component repository when installing or removing
packages, via the administration tool. Component writers will
typically work on just one package in the component repository.
</para>
</listitem>
<listitem>
<para>
The build tree. Each configuration has its own build tree, which can
be regenerated at any time using the configuration's
<filename>ecos.ecc</filename> savefile. The build tree contains only
intermediate files, primarily object files. Once a build is complete
the build tree contains no information that is useful for application
development and can be wiped, although this would slow down any
rebuilds following changes to the configuration.
</para>
</listitem>
<listitem>
<para>
The install tree. This is populated during a build, and contains all
the files relevant to application development. There will be a
<filename class="directory">lib</filename> sub-directory which
typically contains <filename>libtarget.a</filename>, a linker script,
start-up code, and so on. There will also be an <filename
class="directory">include</filename> sub-directory containing all the
header files exported by the various packages. There will also be a
<filename class="directory">include/pkgconf</filename> sub-directory
containing various configuration header files with
<literal>#define's</literal> for the options. Typically the install
tree is created within the build tree, but this is not a requirement.
</para>
</listitem>
</orderedlist>

<para>
The build process involves the following steps:
</para>
<orderedlist>
<listitem>
<para>
Given a configuration, the component framework is responsible for
creating all the directories in the build and install trees. If these
trees already exist then the component framework is responsible for
any clean-ups that may be necessary, for example if a package has been
removed then all related files should be expunged from the build and
install trees. The configuration header files will be generated at
this time. Depending on the host environment, the component framework
will also generate makefiles or some other way of building the various
packages. Every time the configuration is modified this step needs to
be repeated, to ensure that all option consequences take effect. Care
is taken that this will not result in unnecessary rebuilds.
</para>
<note>
<para>
At present this step needs to be invoked manually. In a future version
the generated makefile may if desired perform this step automatically,
using a dependency on the <filename>ecos.ecc</filename> savefile.
</para>
</note>
</listitem>
<listitem>
<para>
The first step in an actual build is to make sure that the install
tree contains all exported header files. All compilations will use
the install tree's <filename class="directory">include</filename>
directory as one of the places to search for header files.
</para>
</listitem>
<listitem>
<para>
All source files relevant to the current configuration get compiled.
This involves a set of compiler flags initialized on a per-target
basis, with each package being able to modify these flags, and with
the ability for the user to override the flags as well. Care has to be
taken here to avoid inappropriate target-dependencies in packages that
are intended to be portable. The component framework has built-in
knowledge of how to handle C, C++ and assembler source files &mdash;
other languages may be added in future, as and when necessary. The
<link linkend="ref.compile">&compile;</link> property is used to
list the files that should get compiled. All object files end up in
the build tree.
</para>
</listitem>
<listitem>
<para>
Once all the object files have been built they are collected into a
library, typically <filename>libtarget.a</filename>, which can then be
linked with application code. The library is generated in the install
tree. 
</para>
</listitem>
<listitem>
<para>
The component framework provides support for custom build steps, using
the <link linkend="ref.make-object">&make-object;</link> and
<link linkend="ref.make">&make;</link> properties. The results of
these custom build steps can either be object files that should end up
in a library, or other files such as a linker script. It is possible
to control the order in which these custom build steps take place, for
example it is possible to run a particular build step before any of
the compilations happen.
</para>
</listitem>
</orderedlist>

</sect2>

<!-- }}} -->
<!-- {{{ Sources                -->

<sect2 id="package.source">
<title>Configurable Source Code</title>
<para>
All packages should be totally portable to all target hardware (with
the obvious exceptions of HAL and device driver packages). They should
also be totally bug-free, require the absolute minimum amount of code
and data space, be so efficient that cpu time usage is negligible, and
provide lots of configuration options so that application developers
have full control over the behavior. The configuration options are
optional only if a package can meet the requirements of every
potential application without any overheads. It is not the purpose of
this guide to explain how to achieve all of these requirements.
</para>
<para>
The &eCos; component framework does have some important implications
for the source code: compiler flag dependencies; package interfaces
vs. implementations; and how configuration options affect source code.
</para>

<sect3 id="package.source.flags">
<title>Compiler Flag Dependencies</title>
<para>
Wherever possible component writers should avoid dependencies on
particular compiler flags. Any such dependencies are likely to impact
portability. For example, if one package needs to be built in
big-endian mode and another package needs to be built in little-endian
mode then usually it will not be possible for application developers
to use both packages at the same time; in addition the application
developer is no longer given a choice in the matter. It is far better
for the package source code to adapt the endianness at compile-time,
or possibly at run-time although that will involve code-size
overheads.
</para>
<note>
<para>
A related issue is that the current support for handling compiler
flags in the component framework is still limited and incapable of
handling flags at a very fine-grain. The support is likely to be
enhanced in future versions of the framework, but there are
non-trivial problems to be resolved.
</para>
</note>
</sect3>

<sect3 id="package.source.interfaces">
<title>Package Interfaces and Implementations</title>
<para>
The component framework provides encapsulation at the package level. A
package <literal>A</literal> has no way of accessing the
implementation details of another package <literal>B</literal> at
compile-time. In particular, if there is a private header file
somewhere in a package's <filename class="directory">src</filename>
sub-directory then this header file is completely invisible to other
packages. Any attempts to cheat by using relative pathnames beginning
with <filename class="directory">../..</filename> are generally doomed
to failure because of the presence of package version directories.
There are two ways in which one package can affect another: by means
of the exported header files, which define a public interface; or via
the &CDL; scripts.
</para>
<para>
This encapsulation is a deliberate aspect of the overall &eCos;
component framework design. In most cases it does not cause any
problems for component writers. In some cases enforcing a clean
separation between interface and implementation details can improve
the code. Also it reduces problems when a package gets upgraded:
component writers are free to do pretty much anything on the
implementation side, including renaming every single source file; care
has to be taken only with the exported header files and with the &CDL;
data, because those have the potential of impacting other packages.
Application code is similarly unable to access package implementation
details, only the exported interface.
</para>
<para>
Very occasionally the inability of one package to see implementation
details of another does cause problems. One example occurs in HAL
packages, where it may be desirable for the architectural, variant and
platform HAL's to share some information that should not be visible to
other packages or to application code. This may be addressed in the
future by introducing the concept of <literal>friend</literal>
packages, just as a C++ class can have <literal>friend</literal>
functions and classes which are allowed special access to a class
internals. It is not yet clear whether such cases are sufficiently
frequent to warrant introducing such a facility.
</para>
</sect3>

<sect3 id="package.source.config">
<title>Source Code and Configuration Options</title>
<para>
Configurability usually involves source code that needs to implement
different behavior depending on the settings of configuration
options. It is possible to write packages where the only consequence
associated with various configuration options is to control what gets
built, but this approach is limited and does not allow for
fine-grained configurability. There are three main ways in which
options could affect source code at build time:
</para>

<orderedlist>
<listitem>
<para>
The component code can be passed through a suitable preprocessor,
either an existing one such as <application
class="software">m4</application> or a new one specially designed with
configurability in mind. The original sources would reside in the
component repository and the processed sources would reside in the
build tree. These processed sources can then be compiled in the usual
way.
</para>
<para>
This approach has two main advantages. First, it is independent from
the programming language used to code the components, provided
reasonable precautions are taken to avoid syntax clashes between
preprocessor statements and actual code. This would make it easier in
future to support languages other than C and C++. Second, configurable
code can make use of advanced preprocessing facilities such as loops
and recursion. The disadvantage is that component writers would have
to learn about a new preprocessor and embed appropriate directives in
the code. This makes it much more difficult to turn existing code into
components, and it involves extra training costs for the component
writers.
</para>
</listitem>
<listitem>
<para>
Compiler optimizations can be used to elide code that should not be
present, for example:
</para>
<programlisting width=72>
    &hellip;
    if (CYGHWR_NUMBER_UARTS &gt; 0) {
        &hellip;
     }
    &hellip;
</programlisting>
<para>
If the compiler knows that <varname>CYGHWR_NUMBER_UARTS</varname> is
the constant number 0 then it is a trivial operation to get rid of the
unnecessary code. The component framework still has to define this
symbol in a way that is acceptable to the compiler, typically by using
a <literal>const</literal> variable or a preprocessor symbol. In some
respects this is a clean approach to configurability, but it has
limitations. It cannot be used in the declarations of data structures
or classes, nor does it provide control over entire functions. In
addition it may not be immediately obvious that this code is affected
by configuration options, which may make it more difficult to
understand.
</para>
</listitem>
<listitem>
<para>
Existing language preprocessors can be used. In the case of C or C++
this would be the standard C preprocessor, and configurable code would
contain a number of <literal>#ifdef</literal> and
<literal>#if</literal> statements.
</para>
<programlisting width=72>
#if (CYGHWR_NUMBER_UARTS &gt; 0)
     &hellip;
#endif
</programlisting>
<para>
This approach has the big advantage that the C preprocessor is a
technology that is both well-understood and widely used. There are
also disadvantages: it is not directly applicable to components
written in other languages such as Java (although it is possible to
use the C preprocessor as a stand-alone program); the preprocessing
facilities are rather limited, for example there is no looping
facility; and some people consider the technology to be ugly. Of
course it may be possible to get around the second objection by
extending the preprocessor that is used by gcc and g++.
</para>
</listitem>
</orderedlist>

<para>
The current component framework generates configuration header files
with C preprocessor <literal>#define's</literal> for each option
(typically, there various properties which can be used to control
this). It is up to component writers to decide whether to use
preprocessor <literal>#ifdef</literal> statements or language
constructs such as <literal>if</literal>. At present there is no
support for languages which do not involve the C preprocessor,
although such support can be added in future when the need arises.
</para>

</sect3>


</sect2>

<!-- }}} -->
<!-- {{{ Headers                -->

<sect2 id="package.headers">
<title>Exported Header Files</title>

<para>
A package's exported header files should specify the interface
provided by that package, and avoid any implementation details.
However there may be performance or other reasons why implementation
details occasionally need to be present in the exported headers.
</para>

<note>
<para>
Not all programming languages have the concept of a header file. In
some cases the component framework would need extensions to support
packages written in such languages. 
</para>
</note>

<para>
Configurability has a number of effects on the way exported header
files should be written. There may be configuration options which
affect the interface of a package, not just the implementation. It is
necessary to worry about nested <literal>#include's</literal> and how
this affects package and application builds. A special case of this
relates to whether or not exported header files should
<literal>#include</literal> configuration headers. These configuration
headers are exported, but should only be <literal>#include'd</literal>
when necessary.
</para>

<sect3 id="package.headers.functions">
<title>Configurable Functionality</title>

<para>
Many configuration options affect only the implementation of a
package, not the interface. However some options will affect the
interface as well, which means that the options have to be tested in
the exported header files. Some implementation choices, for example
whether or not a particular function should be inlined, also need to
be tested in the header file because of language limitations.
</para>
<para>
Consider a configuration option
<varname>CYGFUN_KERNEL_MUTEX_TIMEDLOCK</varname> which controls
whether or not a function <function>cyg_mutex_timedlock</function> is
provided. The exported kernel header file <filename
class="headerfile">cyg/kernel/kapi.h</filename> could contain the
following:
</para>

<programlisting width=72>
#include &lt;pkgconf/kernel.h&gt;
&hellip;
#ifdef CYGFUN_KERNEL_MUTEX_TIMEDLOCK
extern bool cyg_mutex_timedlock(cyg_mutex_t*);
#endif
</programlisting>

<para>
This is a correct header file, in that it defines the exact interface
provided by the package at all times. However is has a number of
implications. First, the header file is now dependent on <filename
class="headerfile">pkgconf/kernel.h</filename>, so any changes to
kernel configuration options will cause <filename
class="headerfile">cyg/kernel/kapi.h</filename> to be out of date, and
any source files that use the kernel interface will need rebuilding.
This may affect sources in the kernel package, in other packages, and
in application source code. Second, if the application makes use of
this function somewhere but the application developer has
misconfigured the system and disabled this functionality anyway then
there will now be a compile-time error when building the application.
Note that other packages should not be affected, since they should
impose appropriate constraints on
<varname>CYGFUN_KERNEL_MUTEX_TIMEDLOCK</varname> if they use that
functionality (although of course some dependencies like this may get
missed by component developers).
</para>
<para>
An alternative approach would be:
</para>

<programlisting width=72>
extern bool cyg_mutex_timedlock(cyg_mutex_t*);
</programlisting>

<para>
Effectively the header file is now lying about the functionality
provided by the package. The first result is that there is no longer a
dependency on the kernel configuration header. The second result is
that an application file using the timed-lock function will now
compile, but the application will fail to link. At this stage the
application developer still has to intervene, change the
configuration, and rebuild the system. However no application
recompilations are necessary, just a relink.
</para>

<para>
Theoretically it would be possible for a tool to analyze linker errors
and suggest possible configuration changes that would resolve the
problem, reducing the burden on the application developer. No such
tool is planned in the short term.
</para>

<para>
It is up to component writers to decide which of these two approaches
should be preferred. Note that it is not always possible to avoid
<literal>#include'ing</literal> a configuration header file in an
exported one, for example an option may affect a data structure rather
than just the presence or absence of a function. Issues like this will
vary from package to package.
</para>

</sect3>

<sect3 id="package.headers.includes">
<title>Nested <literal>#include's</literal></title>
<para>
As a general rule, unnecessary <literal>#include's</literal> should be
avoided. A header file should <literal>#include</literal> only those
header files which are absolutely needed for it to define its
interface. Any additional <literal>#include's</literal> make it more
likely that package or application source files become dependent on
configuration header files and will get rebuilt unnecessarily when
there are minor configuration changes.
</para>
</sect3>

<sect3 id="package.headers.configincludes">
<title>Including Configuration Headers</title>
<para>
Exported header files should avoid <literal>#include'ing</literal>
configuration header files unless absolutely necessary, to avoid
unnecessary rebuilding of both application code and other packages
when there are minor configuration changes. A
<literal>#include</literal> is needed only when a configuration option
affects the exported interface, or when it affects some implementation
details which is controlled by the header file such as whether or not
a particular function gets inlined.
</para>
<para>
There are a couple of ways in which the problem of unnecessary
rebuilding could be addressed. The first would require more
intelligent handling of header file dependency handling by the tools
(especially the compiler) and the build system. This would require
changes to various non-eCos tools. An alternative approach would be to
support finer-grained configuration header files, for example there
could be a file <filename
class="headerfile">pkgconf/libc/inline.h</filename> controlling which
functions should be inlined. This could be achieved by some fairly
simple extensions to the component framework, but it makes it more
difficult to get the package header files and source code correct:
a C preprocessor <literal>#ifdef</literal> directive does not
distinguish between a symbol not being defined because the option is
disabled, or the symbol not being defined because the appropriate
configuration header file has not been <literal>#include'd</literal>.
It is likely that a cross-referencing tool would have to be developed
first to catch problems like this, before the component framework
could support finer-grained configuration headers.
</para>
</sect3>

</sect2>

<!-- }}} -->
<!-- {{{ Documentation          -->

<sect2 id="package.documentation">
<title>Package Documentation</title>
<para>
On-line package documentation should be in HTML format. The component
framework imposes no special limitations: component writers can decide
which version of the HTML specification should be followed; they can
also decide on how best to cope with the limitations of different
browsers. In general it is a good idea to keep things simple.
</para>
</sect2>

<!-- }}} -->
<!-- {{{ Tests                  -->

<sect2 id="package.tests">
<title>Test Cases</title>
<para>
Packages should normally come with one or more test cases. This allows
application developers to verify that a given package works correctly
on their particular hardware and in their particular configuration,
making it slightly more likely that they will attempt to find bugs in
their own code rather than automatically blaming the component
writers.
</para>
<para>
At the time of writing the application developer support for building
and running test cases via the component framework is under review and
likely to change. Currently each test case should consist of a single
C or C++ source file that can be compiled with the package's set of
compiler flags and linked like any application program. Each test case
should use the testing API defined by the infrastructure. A
magically-named calculated configuration option of the form
<varname>CYGPKG_&lt;PACKAGE-NAME&gt;_TESTS</varname> lists the test
cases.
</para>
</sect2>

<!-- }}} -->
<!-- {{{ Host tools             -->

<sect2 id="package.host">
<title>Host-side Support</title>
<para>
On occasion it would be useful for an &eCos; package to be shipped
with host-side support. This could take the form of an additional tool
needed to build that package. It could be an application intended to
communicate with the target-side package code and display monitoring
information. It could be a utility needed for running the package test
cases, especially in the case of device drivers. The component
framework does not yet provide any such support for host-side
software, and there are obvious issues related to portability to the
different machines that can be used for hosts. This issue may get
addressed in some future release. In some cases custom build steps can
be subverted to do things on the host side rather than the target
side, but this is not recommended.
</para>
</sect2>

<!-- }}} -->

</sect1>

<!-- }}} -->
<!-- {{{ Package Distributions  -->

<sect1 id="package.distrib">
<title>Making a Package Distribution</title>

<para>
Developers of new &eCos; packages are advised to distribute their
packages in the form of &eCos; package distribution files. Packages
distributed in this format may be added to existing &eCos; component
repositories in a robust manner using the Package Administration Tool.
This chapter describes the format of package distribution files and
details how to prepare an eCos package for distribution in this format.
</para>

<sect2 id="package.distrib.format">
<title>The &eCos; package distribution file format</title>

<para>
eCos package distribution files are gzipped GNU tar archives which
contain both the source code for one or more &eCos; packages and a
data file containing package information to be added to the component
repository database. The distribution files are subject to the
following rules:
</para>
<orderedlist numeration="Loweralpha">
<listitem>
<para>
The data file must be named <filename>pkgadd.db</filename> and must be
located in the root of the tar archive. It must contain data in a
format suitable for appending to the eCos repository database
(<database>ecos.db</database>). <xref linkend="language.database">
describes this data format. Note that a database consistency check is
performed by the &eCos; Administration Tool when
<filename>pkgadd.db</filename> has been appended to the database. Any
new target entries which refer to unknown packages will be removed at
this stage.
</para>
</listitem>
<listitem>
<para>
The package source code must be placed in one or more <filename
class="directory">&lt;package-path&gt;/&lt;version&gt;</filename>
directories in the tar archive, where each &lt;package-path&gt;
directory path is specified as the directory attribute of one of the
packages entries in <filename>pkgadd.db</filename>.
</para>
</listitem>
<listitem>
<para>
An optional license agreement file named
<filename>pkgadd.txt</filename> may be placed in the root of the tar
archive. It should contain text with a maximum line length of 79
characters. If this file exists, the contents will be presented to the
user during installation of the package. The &eCos; Package
Administration Tool will then prompt the user with the question
<prompt>"Do you accept all the terms of the preceding license
agreement?"</prompt>. The user must respond
<userinput>"yes"</userinput> to this prompt in order to proceed with
the installation.
</para>
</listitem>
<listitem>
<para>
Optional template files may be placed in one or more <filename
class="directory">templates/&lt;template_name&gt;</filename>
directories in the tar archive. Note that such template files would be
appropriate only where the packages to be distributed have a complex
dependency relationship with other packages. Typically, a third party
package can be simply added to an eCos configuration based on an
existing core template and the provision of new templates would not be
appropriate. <xref linkend="advanced.templates"> contains more
information on templates.
</para>
</listitem>
<listitem>
<para>
The distribution file must be given a <filename>.epk</filename> (not
<filename>.tar.gz</filename>) file extension. The
<filename>.epk</filename> file extension serves to distinguish &eCos;
package distributions files from generic gzipped GNU tar archives. It
also discourages users from attempting to extract the package from the
archive manually. The file browsing dialog of the &eCos; Package
Administration Tool lists only those files which have a
<filename>.epk</filename> extension.
</para>
</listitem>
<listitem>
<para>
No other files should be present in the archive.
</para>
</listitem>
<listitem>
<para>
Files in the tar archive may use <literal>LF</literal> or
<literal>CRLF</literal> line endings interchangably. The &eCos;
Administration Tool ensures that the installed files are given the
appropriate host-specific line endings.
</para>
</listitem>
<listitem>
<para>
Binary files may be placed in the archive, but the distribution of
object code is not recommended. All binary files must be given a
<literal>.bin</literal> suffix in addition to any file extension they
may already have. For example, the GIF image file
<filename>myfile.gif</filename> must be named
<filename>myfile.gif.bin</filename> in the archive. The
<filename>.bin</filename> suffix is removed during file extraction and
is used to inhibit the manipulation of line endings by the &eCos;
Administration Tool.
</para>
</listitem>
</orderedlist>
</sect2>

<sect2 id="package.distrib.prepare">
<title>Preparing eCos packages for distribution</title>

<para>
Development of new &eCos; packages or new versions of existing &eCos;
packages will take place in the context of an existing &eCos;
component repository. This section details the steps involved in
extracting new packages from a repository and generating a
corresponding &eCos; package distribution file for distribution of the
packages to other &eCos; users. The steps required are as follows:
</para>

<orderedlist numeration="Loweralpha">
<listitem>
<para>
Create a temporary directory <filename
class="directory">$PKGTMP</filename> for manipulation of the package
distribution file contents and copy the source files of the new
packages into this directory, preserving the relative path to the
package. In the case of a new package at <filename
class="directory">mypkg/current</filename> in the repository:
</para>
<screen width=72>
    $ mkdir -p $PKGTMP/mypkg
    $ cp -p -R $ECOS_REPOSITORY/mypkg/current $PKGTMP/mypkg
</screen>
<para>
Where more than one package is to be distributed in a single package
distribution file, copy each package in the above manner. Note that
multiple packages distributed in a single package distribution file
cannot be installed separately. Where such flexibility is required,
distribution of each new package in separate package distribution files
is recommended.
</para>
</listitem>
<listitem>
<para>
Copy any template files associated with the distributed packages into
the temporary directory, preserving the relative path to the template.
For example:
</para>
<screen width=72>
    $ mkdir -p $PKGTMP/templates
    $ cp -p -R $ECOS_REPOSITORY/templates/mytemplate $PKGTMP/templates
</screen>
</listitem>
<listitem>
<para>
Remove any files from the temporary directory hierarchy which you do
not want to distribute with the packages (eg object files, <filename
class="directory">CVS</filename> directories).
</para>
</listitem>
<listitem>
<para>
Add a <filename>.bin</filename> suffix to the name of any binary
files. For example, if the packages contains GIF image files (*.gif)
for documentation purposes, such files must be renamed to *.gif.bin as
follows:
</para>
<screen width=72>
   $ find $PKGTMP -type f -name '*.gif' -exec mv {} {}.bin ';'
</screen>
<para>
The <filename>.bin</filename> suffix is removed during file extraction
and is used to inhibit the manipulation of line endings by the eCos
Package Administration Tool.
</para>
</listitem>
<listitem>
<para>
Extract the package records for the new packages from the package
database file at <database>$ECOS_REPOSITORY/ecos.db</database> and
create a new file containing these records at
<filename>$PKGTMP/pkgadd.db</filename> (in the root of the temporary
directory hierarchy). Any target records which reference the
distributed packages must also be provided in pkgadd.db.
</para>
</listitem>
<listitem>
<para>
Rename the version directories under <filename
class="directory">$PKGTMP</filename> (typically <filename
class="directory">current</filename> during development) to reflect
the versions of the packages you are distributing. For example,
version 1.0 of a package may use the version directory name <filename
class="directory">v1_0</filename>:
</para>
<screen width=72>
    $ cd $PKGTMP/mypkg
    $ mv current v1_0
</screen>
<para>
<xref linkend="package.versions"> describes the version naming
conventions.
</para>
</listitem>
<listitem>
<para>
Rename any template files under <filename
class="directory">$PKGTMP</filename> (typically
<filename>current.ect</filename> during development) to reflect the
version of the template you are distributing. For example, version 1.0
of a template may use the filename <filename>v1_0.ect</filename>:
</para>
<screen width=72>
    $ cd $PKGTMP/templates/mytemplate
    $ mv current.ect v1_0.ect
</screen>
<para>
It is also important to edit the contents of the template file, changing
the version of each referenced package to match that of the packages you
are distributing. This step will eliminate version warnings during the
subsequent loading of the template.
</para>
</listitem>
<listitem>
<para>
Optionally create a licence agreement file at
<filename>$PKGTMP/pkgadd.txt</filename> containing the licensing terms
under which you are distributing the new packages. Limit each line in
this file to a maximum of 79 characters.
</para>
</listitem>
<listitem>
<para>
Create a GNU tar archive of the temporary directory hierarchy. By
convention, this archive would have a name of the form
<filename>&lt;package_name&gt;-&lt;version&gt;</filename>:
</para>
<screen width=72>
    $ cd $PKGTMP
    $ tar cf mypkg-1.0.tar *
</screen>
<para>
Note that non-GNU version of tar may create archive files which exhibit
subtle incompatibilities with GNU tar. For this reason, always use GNU
tar to create the archive file.
</para>
</listitem>
<listitem>
<para>
Compress the archive using gzip and give the resulting file a
<filename>.epk</filename> file extension:
</para>
<screen width=72>
    $ gzip mypkg-1.0.tar
    $ mv mypkg-1.0.tar.gz mypkg-1.0.epk
</screen>
<para>
The resulting eCos package distribution file (*.epk) is in a compressed
format and may be distributed without further compression.
</para>
</listitem>
</orderedlist>

</sect2>

</sect1>

<!-- }}} -->

</chapter>