* Updates.

This commit is contained in:
Eelco Dolstra 2003-11-26 12:30:16 +00:00
parent 2a4bac5459
commit f6a30ab264
5 changed files with 169 additions and 336 deletions

View file

@ -12,7 +12,7 @@
]> ]>
<book> <book>
<title>Nix: The Manual</title> <title>Nix: A System for Software Deployment</title>
<bookinfo> <bookinfo>
<author> <author>

View file

@ -1,34 +1,43 @@
<appendix> <appendix>
<title>Bugs</title> <title>Bugs / To-Do</title>
<itemizedlist> <itemizedlist>
<listitem> <listitem>
<para> <para>
Nix should automatically recover the Berkeley DB database. Nix should automatically remove Berkeley DB logfiles.
</para> </para>
</listitem> </listitem>
<listitem> <listitem>
<para> <para>
Nix should automatically remove Berkeley DB logfiles. Unify the concepts of successors and substitutes into a general notion
of <emphasis>equivalent expressions</emphasis>. Expressions are
equivalent if they have the same target paths with the same
identifiers. However, even though they are functionally equivalent,
they may differ stronly with respect to their <emphasis>performance
characteristics</emphasis>. For example, realising a slice is more
efficient that realising the derivation from which that slice was
produced. On the other hand, distributing sources may be more
efficient (storage- or bandwidth-wise) than distributing binaries. So
we need to be able to attach weigths or priorities or performance
annotations to expressions; Nix can then choose the most efficient
expression dependent on the context.
</para> </para>
</listitem> </listitem>
<listitem> <listitem>
<para> <para>
Unify the concepts of successors and substitutes into a general notion <emphasis>Build management.</emphasis> In principle it is already
of <emphasis>equivalent expressions</emphasis>. Expressions are possible to do build management using Nix (by writing builders that
equivalent if they have the same target paths with the same perform appropriate build steps), but the Nix expression language is
identifiers. However, even though they are functionally equivalent, not yet powerful enough to make this pleasant (?). The language should
they may differ stronly with respect to their <emphasis>performance be extended with features from the <ulink
characteristics</emphasis>. For example, realising a slice is more url='http://www.cs.uu.nl/~eelco/maak/'>Maak build manager</ulink>.
efficient that realising the derivation from which that slice was Another interesting idea is to write a <command>make</command>
produced. On the other hand, distributing sources may be more implementation that uses Nix as a back-end to support <ulink
efficient (storage- or bandwidth-wise) than distributing binaries. So url='http://www.research.att.com/~bs/bs_faq.html#legacy'>legacy</ulink>
we need to be able to attach weigths or priorities or performance build files.
annotations to expressions; Nix can then choose the most efficient
expression dependent on the context.
</para> </para>
</listitem> </listitem>

View file

@ -1,74 +1,114 @@
<chapter> <chapter>
<title>Installation</title> <title>Installation</title>
<sect1>
<title>Prerequisites</title>
<para>
Nix uses Sleepycat's Berkeley DB and CWI's ATerm library. However, these
are fetched automatically as part of the build process.
</para>
<para>
Other than that, you need a good C++ compiler. GCC 2.95 does not appear
to work; please use GCC 3.x.
</para>
</sect1>
<sect1> <sect1>
<title>Obtaining Nix</title> <title>Obtaining Nix</title>
<para> <para>
Nix can be obtained from its <ulink The easiest way to obtain Nix is to download a <ulink
url='http://losser.st-lab.cs.uu.nl:12080/repos/trace/nix/trunk'>Subversion url='http://www.cs.uu.nl/groups/ST/Trace/Nix'>source
distribution.</ulink>
</para>
<para>
Alternatively, the most recent sources of Nix can be obtained from its
<ulink url='https://svn.cs.uu.nl:12443/repos/trace/nix/trunk'>Subversion
repository</ulink>. For example, the following command will check out repository</ulink>. For example, the following command will check out
the latest revision into a directory called <filename>nix</filename>: the latest revision into a directory called <filename>nix</filename>:
</para> </para>
<screen> <screen>
$ svn checkout http://losser.st-lab.cs.uu.nl:12080/repos/trace/nix/trunk nix</screen> $ svn checkout https://svn.cs.uu.nl:12443/repos/trace/nix/trunk nix</screen>
<para> <para>
Likewise, specific releases can be obtained from the <ulink Likewise, specific releases can be obtained from the <ulink
url='http://losser.st-lab.cs.uu.nl:12080/repos/trace/nix/tags'>tags url='https://svn.cs.uu.nl:12443/repos/trace/nix/tags'>tags
directory</ulink> of the repository. If you don't have Subversion, you directory</ulink> of the repository. If you don't have Subversion, you
can download a <ulink can also download an automatically generated <ulink
url='http://losser.st-lab.cs.uu.nl:12080/dist/trace/'>compressed url='https://svn.cs.uu.nl:12443/dist/trace/'>compressed
tar-file</ulink> of the latest revision of the repository. tar-file</ulink> of the head revision of the trunk.
</para> </para>
</sect1> </sect1>
<sect1>
<title>Prerequisites</title>
<para>
A fairly recent version of GCC/G++ is required. Version 2.95 and higher
should work.
</para>
<para>
To rebuild this manual and the man-pages you need the
<command>xmllint</command> and <command>xsltproc</command>, which are
part of the <literal>libxml2</literal> and <literal>libxslt</literal>
packages, respectively. You also need the <ulink
url='http://docbook.sourceforge.net/projects/xsl/'>DocBook XSL
stylesheets</ulink> and optionally the <ulink
url='http://www.oasis-open.org/docbook/xml/4.2/docbook-xml-4.2.zip'>
DocBook XML 4.2 DTD</ulink>. Note that these are only required if you
modify the manual sources or when you are building from the Subversion
repository.
</para>
<para>
Nix uses Sleepycat's Berkeley DB, CWI's ATerm library, and SDF parser
library. These are included in the Nix source distribution. If you
build from the Subversion repository, you must download them yourself and
place them in the <filename>externals/</filename> directory. See
<filename>externals/Makefile.am</filename> for the precise URLs of these
packages.
</para>
</sect1>
<sect1> <sect1>
<title>Building Nix</title> <title>Building Nix</title>
<para> <para>
To build Nix, do the following: After unpacking or checking out the Nix sources, issue the following
commands:
</para> </para>
<screen> <screen>
$ autoreconf -i
$ ./configure <replaceable>options...</replaceable> $ ./configure <replaceable>options...</replaceable>
$ make $ make
$ make install</screen> $ make install</screen>
<para> <para>
Currently, the only useful switch for <command>configure</command> is When building from the Subversion repository, these should be preceded by
<option>--prefix=<replaceable>prefix</replaceable></option> to specify the command:
where Nix is to be installed. The default installation directory is </para>
<screen>
$ autoreconf -i</screen>
<para>
The installation path can be specified by passing the
<option>--prefix=<replaceable>prefix</replaceable></option> to
<command>configure</command>. The default installation directory is
<filename>/nix</filename>. You can change this to any location you like. <filename>/nix</filename>. You can change this to any location you like.
You should ensure that you have write permission to the installation You must have write permission to the <replaceable>prefix</replaceable>
prefix. path.
</para> </para>
<warning> <warning>
<para> <para>
It is advisable <emphasis>not</emphasis> to change the installation It is advisable <emphasis>not</emphasis> to change the installation
prefix, since doing so will in all likelihood make it impossible to use prefix from its default, since doing so will in all likelihood make it
derivates built on other systems. impossible to use derivations built on other systems.
</para> </para>
</warning> </warning>
<para>
If you want to rebuilt the documentation, pass the full path to the
DocBook XML catalog file (<filename>docbook.cat</filename>) and to the
DocBook XSL stylesheets using the
<option>--with-docbook-catalog=<replaceable>path</replaceable></option>
and <option>--with-docbook-xsl=<replaceable>path</replaceable></option>
options.
</para>
</sect1> </sect1>
</chapter> </chapter>

View file

@ -1,289 +1,94 @@
<chapter> <chapter>
<title>Introduction</title> <title>Introduction</title>
<sect1> <para>
<title>The problem space</title> Nix is a system for software deployment. It supports the
creation and distribution of software packages, as well as the installation
and subsequent management of these on target machines (i.e., it is also a
package manager).
</para>
<para> <para>
Nix is a system for controlling the automatic creation and distribution Nix solves some large problems that exist in most current deployment and
of data, such as computer programs and other software artifacts. This is package management systems. <emphasis>Dependency determination</emphasis>
a very general problem, and there are many applications that fall under is a big one: the correct installation of a software component requires
this description. that all dependencies of that component (i.e., other components used by it)
</para> are also installed. Most systems have no way to verify that the specified
dependencies of a component are actually sufficient.
</para>
<sect2> <para>
<title>Build management</title> Another big problem is the lack of support for concurrent availability of
multiple <emphasis>variants</emphasis> of a component. It must be possible
to have several versions of a component installed at the same time, or
several instances of the same version built with different parameters.
Unfortunately, components are in general not properly isolated from each
other. For instance, upgrading a component that is a dependency for some
other component might break the latter.
</para>
<para> <para>
Build management tools are used to perform <emphasis>software Nix solves these problems by building and storing packages in paths that
builds</emphasis>, that is, the construction of derived products are infeasible to predict in advance. For example, the artifacts of a
(<emphasis>derivates)</emphasis>) such as executable programs from package <literal>X</literal> might be stored in
source code. A commonly used build tool is Make, which is a standard <filename>/nix/store/d58a0606ed616820de291d594602665d-X</filename>, rather
tool on Unix systems. These tools have to deal with several issues: than in, say, <filename>/usr/lib</filename>. The path component
<itemizedlist> <filename>d58a...</filename> is actually a cryptographic hash of all the
inputs (i.e., sources, requisites, and build flags) used in building
<literal>X</literal>, and as such is very fragile: any change to the inputs
will change the hash. Therefore it is not sensible to
<emphasis>hard-code</emphasis> such a path into the build scripts of a
package <literal>Y</literal> that uses <literal>X</literal> (as does happen
with <quote>fixed</quote> paths such as <filename>/usr/lib</filename>).
Rather, the build script of package <literal>Y</literal> is parameterised
with the actual location of <literal>X</literal>, which is supplied by the
Nix system.
</para>
<listitem> <para>
<para> As stated above, the path name of a file system object contain a
<emphasis>Efficiency</emphasis>. Since building large systems cryptographic hash of all inputs involved in building it. A change to any
can take a substantial amount of time, it is desirable that build of the inputs will cause the hash to change--and by extension, the path
steps that have been performed in the past are not repeated name. These inputs include both sources (variation in time) and
unnecessarily, i.e., if a new build differs from a previous build configuration options (variation in space). Therefore variants of the same
only with respect to certain sources, then only the build steps package don't clash---they can co-exist peacefully within the same file
that (directly or indirectly) <emphasis>depend</emphasis> on system.
those sources should be redone. </para>
</para>
</listitem>
<listitem> <para>
<para> Other features:
<emphasis>Correctness</emphasis> is this context means that the </para>
derivates produced by a build are always consistent with the
sources, that is, they are equal to what we would get if we were
to build the derivates from those sources. This requirement is
trivially met when we do a full, unconditional build, but is far
from trivial under the requirement of efficiency, since it is not
easy to determine which derivates are affected by a change to a
source.
</para>
</listitem>
<listitem> <para>
<para> <emphasis>Transparent source/binary deployment.</emphasis>
<emphasis>Variability</emphasis> is the property that a software </para>
system can be built in a (potentially large) number of variants.
Variation exists both in <emphasis>time</emphasis>---the
evolution of different versions of an artifact---and in
<emphasis>space</emphasis>---the artifact might have
configuration options that lead to variants that differ in the
features they support (for example, a system might be built with
or without debugging information).
</para>
<para> <para>
Build managers historically have had good support for variation <emphasis>Unambiguous identification of configuration.</emphasis>
in time (rebuilding the system in an intelligent way when sources </para>
change is one of the primary reasons to use a build manager), but
not always for variation in space. For example,
<command>make</command> will not automatically ensure that
variant builds are properly isolated from each other (they will
in fact overwrite each other unless special precautions are
taken).
</para>
</listitem>
<listitem> <para>
<para> <emphasis>Automatic storage management.</emphasis>
<emphasis>High-level system modelling language</emphasis>. The </para>
language in which one describes what and how derivates are to be
produced should have sufficient abstraction facilities to make it
easy to specify the derivation of even very large systems. Also,
the language should be <emphasis>modular</emphasis> to enable
components from possible different sources to be easily combined.
</para>
</listitem>
</itemizedlist> <para>
</para> <emphasis>Atomic upgrades and rollbacks.</emphasis>
</para>
</sect2> <para>
<emphasis>Support for many simultaneous configurations.</emphasis>
</para>
<sect2> <para>
<title>Package management</title> <emphasis>Portability.</emphasis> Nix is quite portable. Contrary to
build systems like those in, e.g., Vesta and ClearCase, it does not rely on
<para> operating system extensions.
After software has been built, is must also be </para>
<emphasis>deployed</emphasis> in the intended target environment, e.g.,
the user's workstation. Examples include the Red Hat package manager
(RPM), Microsoft's MSI, and so on. Here also we have several issues to
contend with:
<itemizedlist>
<listitem>
<para>
The <emphasis>creation</emphasis> of packages from some formal
description of what artifacts should be distributed in the
package.
</para>
</listitem>
<listitem>
<para>
The <emphasis>deployment</emphasis> of packages, that is, the
mechanism by which we get them onto the intended target
environment. This can be as simple as copying a file, but
complexity comes from the wide range of possible installation
media (such as a network install), and the scalability of the
process (if a program must be installed on a thousand systems, we
do not want to visit each system and perform some manual steps to
install the program on that system; that is, the complexity for
the system administrator should be constant, not linear).
</para>
</listitem>
</itemizedlist>
</para>
</sect2>
</sect1>
<!--######################################################################-->
<sect1>
<title>What Nix provides</title>
<para>
Here is a summary of Nix's main features:
</para>
<itemizedlist>
<listitem>
<para>
<emphasis>Reliable dependencies.</emphasis> Builds of file system
objects depend on other file system object, such as source files,
tools, and so on. We would like to ensure that a build does not
refer to any objects that have not been declared as inputs for that
build. This is important for several reasons. First, if any of the
inputs change, we need to rebuild the things that depend on them to
maintain consistency between sources and derivates. Second, when we
<emphasis>deploy</emphasis> file system objects (that is, copy them
to a different system), we want to be certain that we copy everything
that we need.
</para>
<para>
Nix ensures this by building and storing file system objects in paths
that are infeasible to predict in advance. For example, the
artifacts of a package <literal>X</literal> might be stored in
<filename>/nix/store/d58a0606ed616820de291d594602665d-X</filename>,
rather than in, say, <filename>/usr/lib</filename>. The path
component <filename>d58a...</filename> is actually a cryptographic
hash of all the inputs (i.e., sources, requisites, and build flags)
used in building <literal>X</literal>, and as such is very fragile:
any change to the inputs will change the hash. Therefore it is not
sensible to <emphasis>hard-code</emphasis> such a path into the build
scripts of a package <literal>Y</literal> that uses
<literal>X</literal> (as does happen with <quote>fixed</quote> paths
such as <filename>/usr/lib</filename>). Rather, the build script of
package <literal>Y</literal> is parameterised with the actual
location of <literal>X</literal>, which is supplied by the Nix
system.
</para>
</listitem>
<listitem>
<para>
<emphasis>Support for variability.</emphasis>
</para>
<para>
As stated above, the path name of a file system object contain a
cryptographic hash of all inputs involved in building it. A change to
any of the inputs will cause the hash to change--and by extension,
the path name. These inputs include both sources (variation in time)
and configuration options (variation in space). Therefore variants
of the same package don't clash---they can co-exist peacefully within
the same file system. So thanks to Nix's mechanism for reliably
dealing with dependencies, we obtain management of variants for free
(or, to quote Simon Peyton-Jone, it's not free, but it has already
been paid for).
</para>
</listitem>
<listitem>
<para>
<emphasis>Transparent source/binary deployment.</emphasis>
</para>
</listitem>
<listitem>
<para>
<emphasis>Easy configuration duplication.</emphasis>
</para>
</listitem>
<listitem>
<para>
<emphasis>Automatic storage management.</emphasis>
</para>
</listitem>
<listitem>
<para>
<emphasis>Atomic upgrades and rollbacks.</emphasis>
</para>
</listitem>
<listitem>
<para>
<emphasis>Support for many simultaneous configurations.</emphasis>
</para>
</listitem>
<listitem>
<para>
<emphasis>Portability.</emphasis> Nix is quite portable. Contrary
to build systems like those in, e.g., Vesta and ClearCase [sic?], it
does not rely on operating system extensions.
</para>
</listitem>
</itemizedlist>
<para>
Here is what Nix doesn't yet provide, but will:
</para>
<itemizedlist>
<listitem>
<para>
<emphasis>Build management.</emphasis> In principle it is already
possible to do build management using Fix (by writing builders that
perform appropriate build steps), but the Fix language is not yet
powerful enough to make this pleasant. The <ulink
url='http://www.cs.uu.nl/~eelco/maak/'>Maak build manager</ulink>
should be retargeted to produce Nix expressions, or alternatively,
extend Fix with Maak's semantics and concrete syntax (since Fix needs
a concrete syntax anyway). Another interesting idea is to write a
<command>make</command> implementation that uses Nix as a back-end to
support <ulink
url='http://www.research.att.com/~bs/bs_faq.html#legacy'>legacy</ulink>
build files.
</para>
</listitem>
</itemizedlist>
</sect1>
<!--######################################################################-->
<sect1>
<title>The Nix system</title>
<para>
...
</para>
<para>
Existing tools in this field generally both a underlying model (such as
the derivation graph of build tools, or the versioning scheme that
determines when two packages are <quote>compatible</quote> in a package
management system) and a formalism that allows ...
</para>
<para>
Following the principle of separation of mechanism and policy, the Nix
system separates the <emphasis>low-level aspect</emphasis> of file system
object management form the <emphasis>high-level aspect</emphasis> of the
...
</para>
</sect1>
</chapter> </chapter>
<!-- <!--
local variables: local variables:
sgml-parent-document: ("book.xml" "chapter") sgml-parent-document: ("book.xml" "chapter")

View file

@ -1,27 +1,6 @@
<appendix> <appendix>
<title>Troubleshooting</title> <title>Troubleshooting</title>
<sect1>
<title>Database hangs</title>
<para>
If Nix or Fix appear to hang immediately after they are started, Nix's
database is probably <quote>wedged</quote>, i.e., some process died while
it held a lock on the database. The solution is to ensure that no other
processes are accessing the database and then run the following command:
</para>
<screen>
$ db_recover -e -h <replaceable>prefix</replaceable>/var/nix/db</screen>
<para>
Here, <replaceable>prefix</replaceable> should be replaced by Nix's
installation prefix.
</para>
</sect1>
<sect1> <sect1>
<title>Database logfile removal</title> <title>Database logfile removal</title>