man page

From ArchWiki
Jump to: navigation, search

man pages—abbreviation for "manual pages"—are the form of documentation that is available on almost all UNIX-like operating systems, including Arch Linux. The command used to display them is man.

In spite of their scope, man pages are designed to be self-contained documents, consequentially limiting themselves to referring to other man pages when discussing related subjects. This is in sharp contrast with the hyperlink-aware info files, GNU's attempt at replacing the traditional man page format.

man-db implements man on Arch Linux, and less is the default pager used with man.

Accessing man pages

To read a man page, simply enter:

$ man page_name

Manuals are sorted into several sections. For a full listing see the section entitled "Sections of the manual pages" in man-pages(7).

Man pages are usually referred to by their name, followed by their section number in parentheses. Often there are multiple man pages of the same name, such as man(1) and man(7). In this case, give man the section number followed by the name of the man page, for example:

$ man 5 passwd

to read the man page on /etc/passwd, rather than the passwd utility.

One-line descriptions of man pages can be displayed using the whatis command. For example, for a brief description of the man page sections about ls, type:

$ whatis ls
ls (1p)              - list directory contents
ls (1)               - list directory contents


Man pages all follow a fairly standard format, which helps in navigating them. See the section entitled "Sections within a manual page" in man-pages(7).

Searching manuals

Even though the man utility allows users to display man pages, and search their contents via less, a problem arises when one knows not the exact name of the desired manual page in the first place! Fortunately, the -k or --apropos options can be used to search the manual page descriptions for instances of a given keyword.

The research feature is provided by a dedicated cache. By default you may not have any cache built and all your searches will give you the nothing appropriate result. You can generate the cache or update it by running

# mandb

You should run it everytime a new manpage is installed.

Now you can begin your search. For example, to search for man pages related to "password":

$ man -k password


$ man --apropos password

This is equivalent to calling the apropos command:

$ apropos password

The given keyword is interpreted as a regular expression by default.

If you want to do a more in-depth search by matching the keywords found in the whole articles, you can use the -K option:

$ man -K password

Page width

The man page width is controlled by the MANWIDTH environment variable.

If the number of columns in the terminal is too small (e.g. the window width is narrow), the line breaks will be wrong. This can be very disturbing for reading. You can fix this by setting the MANWIDTH on man invocation. With Bash, that would be:

man() {
    local width=$(tput cols)
    [ $width -gt $MANWIDTH ] && width=$MANWIDTH
    env MANWIDTH=$width \
    man "$@"

Reading local man pages

Instead of the standard interface, using browsers such as lynx and Firefox to view man pages allows users to reap info pages' main benefit of hyperlinked text. Alternatives include the following:

  • KDE users can read man pages in
    • Konqueror using man:<name>.
    • KHelpCenter (khelpcenter) in "UNIX manual pages" or by running khelpcenter man:<name>.
  • xorg-xman provides a categorized look at man pages in X.
  • The GNOME Help Browser yelp can be used via yelp man:<name>.

Converting to browser-readable HTML


Install mdocmlAUR from AUR. To convert a page, for example free(1):

$ gunzip -c /usr/share/man/man1/free.1.gz | mandoc -Thtml -Ostyle=style.css 1> free.html

Now open the file called free.html in your favourite browser.


First, install man2html from the official repositories.

Now, convert a man page:

$ man free | man2html -compress -cgiurl man$section/$title.$section$subsection.html > ~/man/free.html

Another use for man2html is exporting to raw, printer-friendly text:

$ man free | man2html -bare > ~/free.txt

man -H

The GNU implementation of man in the Arch repositories also has the ability to do this on its own:

$ man -H free

This will read your BROWSER environment variable to determine the browser. You can override this by passing the binary to the -H option.


First install roffitAUR from AUR.

To convert a man page:

$ gunzip -c /usr/share/man/man1/free.1.gz | roffit > free.html

Converting to PDF

man pages have always been printable: they are written in troff, which is fundamentally a typesetting language. If you have ghostscript installed, converting a man page to PDF is actually very easy: man -t <manpage> | ps2pdf - <pdf>. This google image search should give you an idea of what the result looks like; it may not be to everybody's liking.

Caveats: Fonts are generally limited to Times at hardcoded sizes. There are no hyperlinks. Some man pages were specifically designed for terminal viewing, and won't look right in PS or PDF form.

Online Man Pages

There are several online databases of man pages, including:

Warning: Some distributions provide patched or outdated man pages that differ from those provided by Arch. Exercise caution when using online man pages.

Noteworthy manpages

Here follows a non-exhaustive list of noteworthy pages that might help you understand a lot of things more in-depth. Some of them might serve as a good reference (like the ASCII table).

More generally, have a look at category 7 pages:

$ man -s 7 -k ".*" 

Arch Linux specific pages:

See also