Consider this scenario: you read about this excellent Linux command that is new to you. You quickly try it out, and discover that the command is not found. Before you can install the command, you must first find the name of the package that provides it.

Below, we outline the instructions to search for a package in Debian/Ubuntu. My next post will do the same for Fedora/Red Hat.

Debian offers multiple command-line tools to search for a package, and which one to use depends on how wide you want to search.

Do you want to search in the installed package base only, or in the entire release repository? Do you want to match the keywords against the file path only, or against the package description as well?

dpkg

Given a file path pattern, the dpkg -S command searches among installed packages only. The pattern is used to match the file path, but not the package description.

$ dpkg -S iostat
sysstat: /usr/bin/nfsiostat
sysstat: /usr/share/man/man1/iostat.1.gz
sysstat: /usr/share/man/man1/nfsiostat.1.gz
linux-headers-3.2.0-4-common: /usr/src/linux-headers-3.2.0-4-common/include/linux/nfs_iostat.h
sysstat: /usr/bin/iostat
sysstat: /usr/share/man/man1/cifsiostat.1.gz
sysstat: /usr/bin/cifsiostat
nfs-common: /usr/sbin/nfsiostatg
nfs-common: /usr/share/man/man8/nfsiostat.8.gz

Note that the above search returns more results than we want. This is because the query automatically adds the wildcard character "*" before and after your search keyword iostat. Hence it matches /usr/bin/nfsiostat as well as /usr/share/man/man1/iostat.1.gz.

If you are sure about the command name, enter it with the following wildcard pattern to avoid the extra output:

$ dpkg -S */iostat$*
sysstat: /usr/bin/iostat

The above output concludes that the sysstat package contains the iostat command.

If the result is path not found matching pattern, it simply means that the command is not in the installed package base. Use apt-file or apt-cache to expand your search.

apt-file

To install apt-file:

$ sudo apt-get update
$ sudo apt-get install apt-file

apt-file builds its own database to index files provided by packages. Before your first search - and periodically after that - you must update the database as follows:

 
$ apt-file update
Downloading complete file http://deb.vanvps.com/debian/dists/wheezy/main/Contents-amd64.gz
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 20.4M  100 20.4M    0     0   372k      0  0:00:56  0:00:56 --:--:--  373k
Ignoring source without Contents File:
  http://security.debian.org/dists/wheezy/updates/main/Contents-amd64.gz
Downloading Index http://deb.vanvps.com/debian/dists/wheezy-updates/main/Contents-amd64.diff/Index:
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0  1057    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
Index is up-to-date.

The apt-file update command reads the /etc/apt/sources.list file to identify the repositories used by your system, and the corresponding mirror sites for downloading packages. For instance, the following line is in sources.list:

deb http://deb.vanvps.com/debian/ wheezy main 

Then, it downloads, from the chosen mirror site, the architecture-specific (amd64) Contents file (http://deb.vanvps.com/debian/dists/wheezy/main/Contents-amd64.gz). The Contents file lists all files in all packages of the selected release (wheezy) and archive area (main).

Now that the database is up-to-date, you are ready to do the actual search:

$ apt-file search -x /pdfcrop$
texlive-extra-utils: /usr/bin/pdfcrop

With the -x parameter, the search string is treated as a regular expression. Therefore, the "$" character is interpreted as "end of line" instead of literally as the dollar sign.

apt-cache

$ sudo apt-get update
$ apt-cache search 2ping
2ping - Ping utility to determine directional packet loss
texlive-extra-utils - TeX Live: TeX auxiliary programs

The apt-cache search command searches the apt cache database for the given pattern (2ping). Before the search, update the apt cache database with apt-get update. The update command reads the /etc/apt/sources.list file, and updates package information for the repositories specified in the file.

By default, apt-cache matches the input pattern (2ping) against both package names and package descriptions stored in the apt cache. Note that the above search returns the texlive-extra-utils package because a match is found in the package description. To view the package description, run the apt-cache show command:

$ apt-cache show texlive-extra-utils

If the --names-only parameter is specified to apt-cache search, only the package name is matched (and the description is not).

$ apt-cache search --names-only 2ping
2ping - Ping utility to determine directional packet loss

The next article discusses how you accomplish the same task in Fedora, and any RPM-based distribution.

A continental breakfast -- but which continent ...?

On this, the final morning of INTA 2014, this Kat let himself in for a completely new experience -- moderating a two-hour Breakfast Table Topic for an expected audience of ten trade mark experts. Table Topics have been around for a while now, and they have certainly grown in popularity. They are also something of a cash cow for INTA, if this Kat's arithmetic is reliable.  Anyway, the topic for this particular table was one of this Kat's long-time favourites, Coexistence Agreements, and the participants who were able to attend (there being a high attrition rate on account of urgent client demands, problems navigating the conference centre and the occasional hangover) were able to offer contributions based on their personal experience of clients, businesses in general, courts and trade mark registries in their respective jurisdictions -- in alphabetical order Israel, Poland, Switzerland, Taiwan, Turkey and the United States.

How did the Table Topic session go?  The two hours positively flew by.  This Kat's role as moderator was quite a difficult one, as his friends and colleagues will appreciate, since it involved him having to refrain from talking. Instead, he had to make himself a catalyst [surely katalyst, murmurs Merpel] for the crystallisation of the thoughts and opinions of others. Despite the struggle to stay relatively silent, or maybe because of it, this moderating moggie learned a great deal more about coexistence of brands than he expected to.  Issues such as the relationship between coexistence agreements and competition law, the different approaches to perpetual contracts taken in different legal codes, the various means of keeping trade mark registries sweet when objections are taken to agreements that appear to condone consumer confusion, the need to educate clients in the monitoring and post-agreement phase were among those which leapt vividly into life.

This Kat was however left with a couple of disappointments.  First, despite the obvious relevance of this topic to brand owners, none signed up for this discussion, leaving the field exclusively to private practitioners. The addition of even one trade mark portfolio holder would have given a more multi-dimensional flavour to our experience.  Secondly, the absence of some of the registrants caused gaps around the table that those who did attend had to overcome. Be that as it may, a good time was had, or appeared to be had, by all; business cards were duly exchanged and compliments mutually distributed.  It's great to feel that you've actually done a good day's work by the time you get up from your breakfast table.

Read More»