Errata

.Xsession tried if .xsession does not exist (page 101)

madduck — 2005-07-25T10:21:00Z

Item 3 in section 3.3.4 states: "If the file [.xsession] is present, ./.Xsession is tried."
This should be: "If the file [.xsession] is not present, ./.Xsession is tried."

Missing labels in archive structure graph (page 105)

madduck — 2005-11-17T14:57:42Z

The two folders marked '-' in figure 4.1 should have been labeled 'pool' and 'main' (from top to bottom).

Missing folder level in figure (page 105)

madduck — 2006-02-16T10:06:36Z

Unfortunately, Figure 4.1 in section 4.1 is missing a level of folders. Below /debian/dists/sarge, each section (main, contrib, non-free) gets a folder, which then hold the binary-* and source folders each.

The corrected diagram looks like this (click to enlarge):

Source package index is named Sources, not Packages (page 107)

madduck — 2006-03-16T09:37:58Z

In the very end of section 4.1.2, I claim that index files are named Packages in all cases. This is only true for binary packages. Source packages are indexed in a file called Sources. Thus, the index of all source packages of the contrib section of the sid archive is in the file /dists/sid/contrib/source/Sources.

Unfortunately, the error also occurs in figure 4.1 on page 105. The following is an updated figure (which also corrects this error):

Should be overlooks, not oversees (page 110)

madduck — 2005-07-18T21:29:53Z

The second paragraph pf 4.3.1 says that a maintainer may occasionally "oversee" a detail. This should have been "overlooks."

Wrong source directory for copy operation (page 133)

madduck — 2005-11-17T13:46:07Z

The first command executed on the shell in the example code snippet in section 5.2.3 should copy the file /var/cache/apt/archives/postfix_2.1.5-1_i386.deb, not /var/spool/apt/archives/postfix_2.1.5-1_i386.deb.

Dpkg does report an error when asked to configure a configured package (page 141)

madduck — 2005-11-17T13:45:13Z

At the end of section 5.3.2, I claim that dpkg does not report an error when asked to configure an already configured package. In fact, dpkg does report an error:

~# dpkg --configure postfix                                                                                                                                [300]
dpkg: error processing postfix (--configure):
 package postfix is already installed and configured
Errors were encountered while processing:
 postfix

Improper redirection of dpkg output (page 147)

madduck — 2005-11-26T18:08:17Z

The last listing on page 147 as well as the first listing on page 148 include a redirection of the form 1>&2. This redirection is an error; nothing should be redirected in both cases.

Missing argument to cut command (page 148)

madduck — 2006-07-30T15:09:36Z

At the bottom of the page, the cut command is missing the field delimiter argument. The correct line reads:

dpkg-query --show --showformat='${Status} ${Package}\n' \
  | grep ^deinstall | cut -f4 -d' ' | xargs dpkg -P

dpkg-query command line should specify postfix, not mc (page 148)

madduck — 2005-09-01T22:33:02Z

The example displaying the use of the --showformat option to dpkg-query should obviously be:

# dpkg-query --show --showformat='${Package}\t${Status}\n' postfix
postfix install ok installed

In the book, it unfortunately specifies mc as the last argument in the command line. I had previously used mc as an example, but switched almost all examples over to postfix as a tribute to my favourite mail server, and its author, Wietse Venema.

Should be dpkg --status, not --show (page 149)

madduck — 2005-12-03T15:53:51Z

The paragraph after the first code snippet on page 149 in section 5.3.4 states that dpkg --show is more or less the same as dpkg --info. Instead of the --show, this should read --status, as in the code snippet.

Name of status is config-files (page 152)

madduck — 2005-09-24T20:55:42Z

The fourth item in the enumeration on page 152 introduces the conf-files dpkg state. The actual name of the state is config-files instead.

Removed state is called deinstall, not remove (page 152)

madduck — 2006-06-25T10:03:17Z

The name of the state of a package that has been removed is called deinstall, not remove.

Name of file set in database is .list, not .files (page 152)

madduck — 2006-07-30T15:10:43Z

In the third point of the list of steps dpkg takes during package removal, the name of the file storing the list of installed files for each package should be postfix.list, not postfix.files.

The same applies for the third point in the list on the next page.

dpkg does not automatically complete installation after dependencies are met (page 157)

madduck — 2005-09-01T22:47:17Z

In chapter 5.3.7, I state:

Since dpkg remembers the administrators request to install postfix, it will retry and automatically complete the installation as soon as the dependencies are fulfilled.

This is not correct. dpkg sets the desired state of the postfix package to "install", but it does not automatically configure it when netbase is installed.

Wrong APT source line for non-US archive (page 164)

madduck — 2006-01-26T15:05:28Z

In the first listing in section 5.4.1, the line for the non-US archive should be:

deb http://nonus.debian.org/debian-non-US sarge/non-US main

This is to say: the name non-US is also part of the distribution.

Components select non-free software, not non-US (page 166)

madduck — 2006-01-26T15:08:05Z

The item on components on the top of page 166 states that components "also specify when non-US software is to be used."

This should say "non-free software" instead. Components cannot be used to select non-US software.

Wrong argument to apt-cdrom (page 167)

madduck — 2006-07-30T15:17:59Z

The argument to apt-cdrom should be add, not --cdrom. Therefore, the first example on page 167 in section should read:

# apt-cdrom add /media/cdrom
Using CD-ROM mount point /media/cdrom/
[...]

Default-Release setting in wrong namespace (page 169)

madduck — 2006-07-30T15:17:55Z

The apt.conf examples on page 169, which illustrate the two different syntax styles, list Default-Release under the wrong namespace. Instead of APT::Get, it should just be APT:

APT::Default-Release "sarge";
[...]

and:

APT {
  [...]
  Default-Release "sarge";
  [...]
}

Unclear wording on Pre-Depends relation (page 210)

madduck — 2006-01-30T10:11:34Z

The description of the Pre-Depends relation as part of the discussion of the Depends relation in section 5.7.3 is unclear. A better explanation is:

Another relation, Pre-Depends, provides a tighter dependency and is only used in very special cases;
packages listed here need to be fully configured before the dependending package can be unpacked or configured. If a previous version of the pre-dependent package has once been configured on the system, the depending package can be unpacked (but not configured) even if the current version of the pre-dependent package has not been configured yet.

Sarge also does not support tilde in versions (page 218)

madduck — 2007-05-10T06:52:00Z

In the end of the forth paragraph, I claim that the woody tools do not support the tilde (~) character in the version number string. The sarge tools also do not support this character. Full support for the tilde character in version numbers will come with the release of Debian etch.

Update: while sarge's dpkg and even APT support the tilde character in version strings, the majority of package- and packaging-related tools in the sarge archive do not. Therefore, the tilde is not fully supported in sarge, but it is in the etch release.

Bad versioning scheme for pre-release versions (page 218)

madduck — 2006-02-01T11:51:44Z

The versioning scheme for pre-release upstream versions proposed in section 5.7.5 is not desirable: encoding the pre-release version in the Debian revision prevents a new orig.tar.gz file from being uploaded when the final release is uploaded.

The following rewrite should make the situation clear. It replaces the second half of the middle paragraphon page 218, starting at "Within the Debian archive, it is customary...":

One scheme could be to release 1.0-rc1 as 1.0-0+1.0rc1+1, followed by 1.0-0+1.0rc1+2, and then 1.0-1 when the final gets released. However, this approach has an inherent problem: for each upstream version, an orig.tar.gz file exists in the Debian archive; Debian revisions are released as diff.gz files against the orig.tar.gz file (see chapter 9.2.12). In the proposed scheme, the orig.tar.gz file will contain the 1.0-rc1 upstream release, and when the final 1.0 release is to be packaged, the diff.gz file for the 1.0-1 Debian release will have to encompass all upstream changes between 1.0-rc1 and 1.0 as well, which is undesirable.

Thus, a better approach is to encode the pre-release version as part of the upstream vesion. This can be done in one of two ways: using the previous upstream version as base (let's assume 0.90), or an obviously fake upstream version that sorts before the final release: 0.90+1.0rc1-1 or 0.999+1.0rc1-1. Now, upon release of 1.0, a new orig.tar.gz file can be uploaded to replace the previous one. With that in mind, it should be easy to infer the upstream version number from the Debian version, even in complex cases.

Source machine name should be 'remote' (page 229)

madduck — 2005-11-17T13:54:08Z

In section 5.8.6, on page 229, the source machine name of the scp call in the code snippet should obviously be remote, not rempte.

Changes are not expected in the stable release (page 24)

madduck — 2005-11-21T11:26:02Z

Section 1.4.2 talks about anticipated changes to the Debian project and system. As should become apparent when reading section 4.3.3, such changes are not to be expected in the stable release of the operating system, which is one of its main selling points. Thus, if you only track stable and have no interest in the surrounding project or new developments, you need not be concerned with these changes.

I keep a list of changes online to make it easier for you to track developments since the book's release.

Definition of source package type flipped (page 233)

madduck — 2005-08-02T06:46:26Z

In the first paragraph of the subsection titled "Obtaining source package" talks about the two types of source packages in the Debian archive: packages requiring modifications prior to their inclusion in Debian, and packages included directly. The text calls the former the class of "native packages" and terms the latter as "normal packages." It should be the other way around.

Long option for dch -v is --newversion (page 237)

madduck — 2005-09-30T09:23:46Z

In the section "Logging the changes" in chapter 5.9, the following invocation of dch is used:

dch --version=$VERSION+0.local.1 -- Made some local changes

The option --version displays the version information of the tool, and does not pass its argument to dch the way I had intended it. The correct long option for -v is --newversion, and thus the command should be:

dch --newversion=$VERSION+0.local.1 -- Made some local changes

Deborphan may need to be run iteratively (page 248)

madduck — 2006-07-30T14:06:57Z

Deborphaner is not iterative (or recursive, depending on how you look at it). In one run it may suggest to remove a set of packages, but it fails to identify packages that would be obsolete if the user removed the packages suggested in the first run.

Wrong port and confusing assumption (page 250)

madduck — 2006-01-09T19:57:42Z

In the subsection on apt-proxy of section 5.11.7, the sources.list file does not include apt-proxy's default port 9999. Instead, it should read:

~# cat <<EOF > /etc/apt/sources.list
http://arakis:9999/debian sarge main
http://arakis:9999/security sarge/updates main
http://arakis:9999/pdo/~madduck/packages/stage ./

In addition, the paragraph just before the snippet says that the machine running apt-proxy must be resolvable from the machine named arakis. Rather, the machine running apt-proxy must be resolvable as arakis from the machine hosting the sources.list file.

Wrong regular expressions in refresh_pattern (page 251)

madduck — 2005-10-06T11:16:16Z

The mysteries of LaTeX have caused the fifth and sixth line of the squid.conf file snippet to be falsely printed. In the fifth line, instead of the underlined parenthesis and 't' and 'g' characters, escaped periods should have appeared. In the sixth line, the wrapping was messed up:

refresh_pattern \.(u?deb|dsc|changes|(orig\.tar|diff)\.gz)$ 14400 20% 2592000
refresh_pattern ((Packages|Sources)(\.(gz|bz2))?|Release(\.gpg)?)$ 14300 20% 14400

Wrong hostname for the Debian snapshot server (page 266)

madduck — 2006-01-10T14:27:36Z

The third paragraph of section 5.12.3 refers to snapshot.debian.org. The correct hostname is snapshot.debian.net.

Code snippet involuntarily repeated (page 278)

madduck — 2005-12-03T15:23:54Z

The first example on page 278 in section 6.1.4 contains a repeated instruction. You can ignore all code after the second command, update-alternatives --display editor, and its output.

Command should be adduser, not aduser (page 281)

madduck — 2005-07-25T10:23:07Z

The command referenced in the third row of table 6.1 should be adduser, not aduser.

Faulty way of spawning sshd in single-user mode (page 299)

madduck — 2006-10-22T15:25:26Z

In chapter 6.3.1, I suggest to add sshd to the /etc/inittab file to be able to login to a remote system even in single-user mode:

# sed -i -e "/^~~/ish:S:respawn:sshd -Do 'AllowUsers=root'" /etc/inittab

Unfortunately, there are two errors with this: first, sshd must be called with the complete path (to guard against trojans), and init must be told to reload the configuration file after the change.

The following corrects both problems:

# sed -i -e "/^~~/ish:S:respawn:/usr/sbin/sshd -Do 'AllowUsers=root'" /etc/inittab
# telinit q

Wrong section numbering (page 307)

madduck — 2005-06-21T11:25:04Z

Section 6.3.3 "Backups" should have been 6.4 and not a child of 6.3 "System initialisation and automatic processes".

Incomplete paragraph about sysklogd daily cron file (page 306)

madduck — 2005-07-27T17:26:17Z

The item describing the sysklogd cron.daily file somehow got mangled hard:

the parenthesis before the word "mainly" in the first line should be closed after the filename, and before the word "written."
the paragraph ends in an incomplete sentence which should have read: "Note that permissions of other log files are not corrected if the administrator altered them."

Confusing year: rex was released in December 1996 (page 32)

madduck — 2005-07-12T13:41:57Z

The fourth paragraph of the subsection entitled "The early days" in "A history lesson" states that Debian rex was released in December "of the same year". This refers to the release of buzz in June 1996, not the release of the Debian botch in December 1995. rex was released in December 1996.

QPL situation not correctly described (page 33)

madduck — 2005-12-13T16:01:10Z

In subsection "Into the next milennium" in section 2.1, I state that Trolltech, the manufacturer of the Qt library, "agreed to license the library under the GNU Public License (GPL) for non-commercial use."

Instead, it should read: "[Trolltech] agreed to license they X11/*nix library under the GNU Public License (GPL) thus making it compatible with the Debian Free Software Guidelines."

Elections had been completed by the time of release (page 33)

madduck — 2005-06-24T10:43:28Z

The last sentence of the "Into the next millenium" subsection should not be there. It was overlooked during the time we were waiting for sarge to release.

Mappings are applied iteratively (page 332)

madduck — 2005-07-26T08:51:52Z

When the second paragraph in the subsection "Interface mappings" in section 6.7.1 states that only the first matching mapping is applied, it is only partially correct. Mappings are applied iteratively until no further mappings match.

Missing braces in udev configuration (page 335)

madduck — 2005-07-26T08:44:22Z

The /etc/udev/rules.d/local-netifaces file in the last example in the subsection "Renaming network devices" in section 6.7.1 fell prey to LaTeX. Instead of SYSFSaddress, it should say SYSFS{address}. The example thus becomes:

~# grep lan /etc/udev/rules.d/local-netifaces
KERNEL="eth*", SYSFS{address}="00:00:de:ad:be:ef", NAME="lan"

Anthony signed the key, not James (page 373)

madduck — 2005-08-08T20:39:53Z

The first text paragraph on page 373 in section 7.5.1 talks about the archive key, which was signed by one of the FTP masters, Anthony Towns. Later in the paragraph, Anthony suddenly becomes James. It should just say Anthony all along

Rewrite of section 2.2.2 (page 39)

madduck — 2006-01-18T15:33:53Z

I am not an expert on legalese and have made quite a few errors in section 2.2.2. Javier Candeira of Barrapunto.com was kind enough to rewrite pages 39 and half of 40 for me to help me get it right.

I have no desire to keep his work (with few modifications from my side) from you:

\subsection{Free beer and free speech} \clabel{freedom}

\index{Debian!freedom|(} \index{free software}

One of the continuously recurring themes in Debian is the topic of Free
Software. Efforts to ensure the freedom of software distributed with Debian
enjoy a similarly high priority as users of the operating system. The \ac{FSF}
identifies four kinds of
freedom\footnote{\link{http://www.fsf.org/philosophy/free-sw.html}}, namely
the freedom to

\begin{mditemize}

\item run the program, for any purpose.

\item study how the program works, and adapt it to your needs.

\item redistribute copies so you can help your neighbour.

\item to improve the program, and release your improvements to the public,
so that the whole community benefits.

\end{mditemize}

For the purpose of the following discussion, these types of freedom
characterise what in this book we will call Free Software, as opposed to
software that is merely free:

\begin{mditemize}

\item Software which is free with a lower-case `f' is that which may be
copied and used without payment. This is what we call ``free as in free
beer'', and here the adjective is connoting freeness, not freedom. This
software can be said to be ``\foreign{gratis}'' (although it is more comon
in French, Spanish and Italian, this is also an English word meaning
``without charge or recompense'' or ``at no cost'').

%TODO use free for beer and Free for speech. consistently.

\item Software which is Free with an upper-case `F' is that which can be
freely used, copied, studied, modified, and redistributed by
the user. This applies to the freedom the user obtains along with the
software. This software is often said to be ``\foreign{libre}'' (which is
the French and Spanish adjective alongside the noun ``freedom'') or ``free
as in free speech.''

\end{mditemize}

These two meanings of the word ``free'' are a cause of much confusion. Many
programmes are available at no charge, but the software may only be used,
not reverse engineered or distributed in modified form, or has a time limit
after which it can't be used at all. This software may be free as in free
beer, but is not Free as in free speech, and is in fact as propietary as the
box-packaged software one pays for in computer shops.

Some companies licence software source code to paying users, but forbid
redistribution in non-binary form, or any redistribution outside the company
that has paid for it. This software is not only non-free (it requires
payment) but non-Free also.

The same confusion can arise on the other end: software which one pays for
(hence it is not free as in beer), but that is Free as in speech by virtue
of licences that grant its users all four permissions listed above.

According to this distinction between these two meanings of the same word,
we have two words for the opposite of ``free''. Non-free software that
requires payment is said to be
``commercial'', whereas non-Free software that
does not give users the four basic freedoms is said to be ``proprietary''.

Thus, there can be commercial distribution of proprietary software
(shrink-wrapped games sold at stores are an example); free distribution of
propietary software (downloadable freeware applications for which no source
code or licence is given); commercial distribution of Free software (the
shrink-wrapped boxes of certain GNU/Linux distributions such as RedHat, for
instance); and finally free distribution of Free software, as in the case of
the Debian System that one can download from the Internet.

Note that in the case of Free Software it is acceptable to charge for its
distribution (which is a freedom after all). However, paid copies are governed
by the same rules as their free-of-charge counterparts and may be used and
redistributed, in a verbatim or modified form, according to the terms of its
licence.

\subsection{Misconceptions about free software}

Despite popular belief, Free software is not:

\begin{mditemize}

\item free from all and any rules or regulations, including copyright.

\item in the Public Domain

\item at odds with Open Source software.

\end{mditemize}

Free software is not totally free of constraints: it is governed by the
laws of copyright (and their analogues in other countries). Free software
authors can use two mechanisms to put copyright law in the service of users'
freedom: the software licence and the copyright statement.

The software licence aims to give users the flexibility to put the software
to productive use according to the will of the software author. The main
difference between Free and proprietary software is that, while propietary
software licences seek to restrict the liberties that users may take with the
code, Free software licences aim to enhance those liberties.

The copyright statement serves to protect the rights of the author, who
chooses the licence and is free to modify it (within the terms of the
licence)\footnote{Obviously, if an author releases
a piece of software under a free licence and then later chooses to
commercialise the product in a propietary manner, any code previously
available continues to remain available under the free licence.}.

Only software in the \emph{public domain} is completely unrestricted and
does not have an owner who may change the licence. Some people mistakenly call
Free software ``Public Domain software'', without realising that the Public
Domain is a precise legal term for works (books and images, not only code)
which are no longer protected under copyright law, and can be freely used and
copied by anyone, without need for a permission or licence.

\emph{Open Source} software and Free software are also often confused with
each other. The \ac{OSI}\footnote{\link{http://www.opensource.org}} started
with the stated good will to end with the confusion between ``free as in
beer''
and ``Free as in free speech'', and intended to have the words ``Open Source''
certified as a trademark which would signal software which was Free as in free
speech. Obviously, they have all but ended with the confusion, for nowadays
many people think that Open Source software is merely software for which there
is available source, but that needs not fulfill other conditions that make it
Free. This is an error and, unfortunately, a common one\footnote{\ldots which
I also made in the first edition of this book.}.

It is true that the supporters of the term ``Free software'' (with members of
the \ac{FSF} and Debian being among the most vocal) emphasise the philosophy
and idealistic value behind the freedom of software, while the followers of
open source software proclaim the improved development cycle and cheaper
costs. But it is also true that the \ac{OSD} states that Open Source licences
must give users the four freedoms demanded by the \ac{FSF}, and thus Open
Source software is also Free, and \foreign{vice versa}. While there might be
an Open Source licence which is not FSF-Free or not approved by Debian, this
is a matter of interpretation, not of principles, as the \ac{OSD} and the
\ac{DFSG} match almost word for word (the \ac{OSD} is based on the \ac{DFSG}).

It can be said that ``Open Source Software'' and ``Free Software'' are
functional synonyms, and for that reason it is becoming common to find the
acronyms ``FLOSS'' (Free, Libre and Open Source Software) or ``F/OSS'' (Free
/ Open Source Software) used by authors who want to stay out of the Free/Open
Source debate.

Qmail is not Open Source (page 40)

madduck — 2005-12-13T17:35:22Z

In the end of section 2.2.2, I cite Qmail as an example of non-free open source software. Here, I have made the common mistake of equating availability of source with "open-source-ness". Qmail is non-free, and not Open Source by the Open Source Definition, although its sources are available.

Free software is not a subset of Open Source (page 40)

madduck — 2005-12-13T17:33:07Z

In chapter 2.2.2, Free Software is incorrectly described as a subset of Open Source: while it is true that there are there Open Source licenses which are not FSF-Free or not approved by Debian, that is a matter of interpretation, not of principles, as the Open Source Definition and the Debian Free Software Guidelines match almost word for word.

Option should be --upgradeable, not --upgraded (page 403)

madduck — 2005-12-03T15:57:19Z

In the third example on page 403 in section 8.2.2, the option to apt-show-versions should be --upgradeable (as in the example just before), not --upgraded.

MIRROR variable incomplete (page 408)

madduck — 2006-07-30T13:46:02Z

In section 8.3.1, the $MIRROR variable set in the example is incomplete: it lacks the path to the /debian directory. The assignment should read:

# MIRROR=http://ftp.debian.org/debian

Regardless, you should still use apt-spy to determine a closer mirror!

grub-update should be update-grub (page 413)

madduck — 2006-01-26T11:47:32Z

The (Debian) command to update grub's menu.lst file is update-grub, not grub-update, which erroneously appears several times on pages 413ff.

Wrong mount options for /dev/hda2 (page 417)

madduck — 2005-12-09T18:47:23Z

In subsection "Converting another Linux to Debian" in section 8.3.1, the /etc/fstab snippet on page 417 includes the following line:

/dev/hda2   /   ext2    default      0  1

Instead of "default", the mount options should be "defaults":

/dev/hda2   /   ext2    defaults     0  1

Wrong chapter reference (page 427)

madduck — 2005-12-09T18:47:00Z

The chapter reference in the last sentence of the item called "Partitioning" should point to appendix C.5, not chapter 19.

Incomplete sentence about use of dh_clean during installation (page 469)

madduck — 2005-08-08T08:52:50Z

The last paragraph of the subsection "Build dependencies and cleaning" abruptly ends with the word "Finally". The missing sentence was: "Finally, we invoke the software's installation method and tell it to install runtime files into a hierarchy rooted at debian/tmp."

Command is dh_install, not dg_install (page 471)

madduck — 2005-12-05T10:06:10Z

In the second sentence of the first item of the list on page 471 in section 9.2, the command should have been dh_install, not dg_install.

Should be -clean, not -free (page 479)

madduck — 2005-07-18T14:48:15Z

The third paragraph on this page talks about packages passing the package checkers as being lintian-free and linda-free. The correct terms are lintian-clean and linda-clean.

Malformed reference to mailing list chapter (page 49)

madduck — 2005-08-01T06:50:25Z

The first paragraph of the section titled "The developer collective" refers to a section on mailinglists in chapter 10. Unfortunately, a markup character was left out in the final script. Instead of "crefmailinglists," it should read "chapter 10.4.1."

Two small errors in preinst file (page 490)

madduck — 2006-01-30T22:34:11Z

The listing in section 9.4.1 contains two small errors related to the preinst file:

The cat instruction should be redirected to debian/foo.preinst, not debian/foo/preinst.
The version comparison within the preinst file should compare against 1.2.3-4, because presumably only versions before 1.2.3-4 used a working directory in /var/state.

dwarfs-debian-guide package out of date and removed (page 519)

madduck — 2005-06-21T16:40:22Z

The dwarfs-debian-guide package no longer exists in the sarge archive. It has been removed in December 2003 by reason of being out of date (according to removals.txt).

Missing tilde in survey URL (page 519)

madduck — 2005-06-21T11:25:05Z

The URL for Enrico Zini's survey is missing a tilde character. The correct URL should read: http://people.debian.org/~enrico/survey/survey.php.

The #debian IRC channel is a users' forum, not a developer forum (page 52)

madduck — 2005-06-25T20:20:59Z

Footnote 43 mentions "developer forums, such as the #debian IRC channel". Even though many developers hang out in the #debian channel, it is primarily a users' channel.

Standard mode allows to choose from a set of tags (page 540)

madduck — 2006-01-31T14:53:16Z

The item on reportbug's standard mode mentions a set of patches to be applied to the bug report. Instead of patches, this should read tags.

Gnoppix is based on Ubuntu, not vice versa (page 564)

madduck — 2006-01-31T15:05:07Z

The last sentence of A.2.3 states that Gnoppix is the basis for the Ubuntu Live CD. This is not accurate. Andreas Müller, the developer of Gnoppix, actually developed the Ubuntu Live CD first, and then reused this work for Gnoppix. Thus, Gnoppix is based on the Ubuntu Live CD, not vice versa.

Partition types are mixed up (page 577)

madduck — 2005-11-10T09:28:17Z

At the bottom of page 577, in section C.2.1., the suggestion should be for data partitions to be of type 0x83 and swap partitions to have the type 0x82, not the other way around.

Wrong URL to package tracking system (page 59)

madduck — 2005-08-03T09:00:15Z

The link given in footnote 60 in section 2.5.1 should be http://packages.qa.debian.org, not http://pts.qa.debian.org

PPP is point-to-point, not peer-to-peer (page 71)

madduck — 2005-11-26T18:16:43Z

In section 3.1.2, PPP is expanded to "Peer-to-Peer Protocol". It should be "Point-to-Point Protocol".

Figure 4.1 - Archive Tree

madduck — 2006-03-16T09:34:06Z