A filter is used by the LP print service each time it has to print a type of file that the printer cannot interpret. Creating a new print filter is not easy, and it usually requires extensive experimentation. The process of defining a new print filter consists of two steps:
Writing a print filter program
Creating a print filter definition
A print filter can be as simple or as complex as needed. Filters contain input types, output types, and complex options that provide a language to process command-line arguments within the filter.
If you have nonPostScript printers, you have to create and add print filters as required. First, you need to understand what print filters are and the requirements that must be met by a filter program.
The LP print service provides filter programs in the /usr/lib/lp/postscript directory. These filters cover most PostScript printing situations, where
the destination printer requires the data to be in PostScript format. A print filter
program must be a binary executable file.
There are two types of print filters: fast filters and slow filters.
Fast filters do not require much processing time to prepare
a file for printing. They must have access to the printer when they run. To be capable
of detecting printer faults, a print filter must be a fast filter. Any filter that
uses the PRINTER keyword as a filter option must be installed as
a fast filter.
Slow filters require a great deal of processing time to prepare a file for printing. They do not require access to the printer when they run. Slow filters are run in the background so they do not tie up the printer, allowing other files that do not need slow filtering to be printed.
The LP print service uses print filters to convert files from one content type to another content type. You can specify the accepted file content types for each printer. The user specifies the file content type when submitting a print request. Then, the LP print service finds a printer that can print files of that content type. Because many applications can generate files for various printers, this method is often sufficient. However, some applications can generate files that cannot be printed on any available printers.
Each time the LP print service receives a request to print a type of file that is in a format that cannot be accepted directly by a printer, the LP print service tries to match the content type of the print request with the content type of the available (or specified) printer. If there is a match, the file can be sent directly to the printer without filtering. If no match is found, or if the content type specifies that a filter be used, the LP print service tries to match the content type of the file with the input content type of available filters. In addition, the LP print service tries to match the output type of the filter with the content type of the printer. When an appropriate filter is found, the print request is passed through the filter.
A print filter handles special printing modes and requests to print specific pages. A special printing mode is needed to print any characteristics of print requests that require a customized filter. Filters handle the following characteristics:
Printer type
Character pitch
Line pitch
Page length
Page width
Pages to print
Character set
Form name
Number of copies
The LP print service provides default settings for these characteristics. However, a print filter can handle some characteristics more efficiently. For example, some printers can handle multiple copies more efficiently than the LP print service. So, in this case, you can provide a filter for multiple-copy page control.
Each printer has its own way of detecting printer faults and transmitting fault signals to the LP print service. The LP print service only checks for hang-ups (loss of carrier) and excessive delays in printing.
Some printers provide good fault coverage and can send a message that describes the reason for a fault. Other printers indicate a fault by using signals other than the signals for loss of carrier signal or shut off of data flow. A filter is required to interpret this additional printer fault information.
A filter can also put a print request on hold, wait for a printer fault to clear, and then resume printing. With this capability, the print request that was interrupted does not need to be reprinted in its entirety. Only a filter that knows the control sequences used by a printer can determine where to break a file into pages. Consequently, only such a filter can find the place in the file where printing should start after a fault is cleared.
When a print filter generates messages, those messages are handled by the LP print service. If alerts are enabled, the LP print service sends alerts to the system administrator. For further information, see Setting Up Printer Fault Alerts.
A print filter can be simple or complex, but it has to meet the following requirements:
The filter should get the contents of a file from its standard input and send the converted file to the standard output.
A program cannot be used as a filter if it references external files. You might be tempted to use a program such astroff, nroff, or a similar word processing program as a filter. The LP print service does not recognize references to other files, known as include files, from a filter program. Because troff and nroff programs allow include files, they can fail when used as filters. If the program needs other files to complete its processing. The program t should not be used as a filter.
The filter should not depend on files that normally would not be accessible to a user. If a filter fails when run directly by a user, it will fail when run by the LP print service.
A slow filter can send messages about errors in the file to standard error. a fast filter should not. Error messages from a slow filter are collected and sent to the user who submitted the print request.
If a slow filter dies because it received a signal, the print request is stopped and the user who submitted the request is notified. Likewise, if a slow filter exits with a nonzero exit code, the print request is stopped and the user is notified. The exit codes from fast filters are treated differently.
If you want the filter to detect printer faults, it should also meet the following requirements:
If possible, the filter should wait for a fault to be cleared before exiting. The filter should also continue to print at the top of the page where printing stopped after the fault is cleared. If you do not want use the continuation feature, the LP print service stops the filter before alerting the administrator.
The filter should send printer fault messages to its standard error as soon as the fault is recognized. The filter does not have to exit, but can wait for the fault to be cleared.
The filter should not send messages about errors in the file to standard error. These messages should be included in the standard output, where they can be read by the user.
The filter should exit with a zero exit code if the file is finished printing (even if errors in the file have prevented it from being printed correctly).
The filter should exit with a nonzero exit code, only if a printer fault has prevented the filter from finishing a print request.
When added to the filter table, the filter must be added as a fast filter.
A print filter definition tells the LP print service about the filter, what
print filter program to run, what kind of conversion the print filter does, and so
on. A set of filter descriptor files are provided in the /etc/lp/fd directory.
These files describe the characteristics of the filters (for example, fast filter
or slow filter), and point to the filter programs (for example, /usr/lib/lp/postscript/postdaisy).
When defining a new print filter, you must create a print filter definition. A print filter definition contains the following information used by the LP print service:
Name of the filter program to run
Input types the filter program accepts
Output types the filter program produces
Printer types to which the filter program can send jobs
Names of specific printers to which the filter program can send jobs
Filter types (either fast or slow)
Options
You can type the characteristics as direct input to the lpfilter command.
You also can create a file that specifies the filter's characteristics, and use the
file name as input to the lpfilter command arguments. Such a file
is called a filter descriptor file and should be located in the /etc/lp/fd directory. These files are not the filters themselves. Rather,
these files point to the filters.
Whether you store the information in a file, or type the information directly on the command line, use the following format:
Command:command-pathname [options]Input types:input-type-listOutput types:output-type-listPrinter types:printer-type-listPrinters:printer-listFilter type: fast or slow Options:template-list
If you provide more than one definition (that is, more than one line)
for any filter characteristic other than Options, only the second
definition is used by the print service.
The information can be arranged in any order, and not all the information is required. When you do not specify values, the values shown in the following table are assigned by default. These values are not very useful, which is why you should specify values.
Table 6.7. Default Values for lpfilter Command
|
Item |
Default Value |
|---|---|
|
Input types |
|
|
Output type |
|
|
Printer types |
|
|
Printers |
|
|
Filter type |
|
Use the full path of the filter program. If there are any fixed options that the program always needs, include them here.
Input types is a list of file content types that the print filter can process. The LP print service does limit the number of input types, but most filters can accept only one input type. Several file types can be similar enough, such that the filter can deal with them. You can use whatever names you prefer, with a maximum of 14 alphanumeric characters and dashes. Do not use underscores as part of the input type name.
The LP print service uses these names to match a filter to a file type, so follow a consistent naming convention. For example, if more than one filter can accept the same input type, use the same name for that input type when you specify it for each filter. Inform your users of these names so that they know how to identify the file type when submitting a file for printing.
Output types is list of file types that the filter can produce as output. For each input type, the filter produces a single output type. The output type can vary, however, from job to job. The name of the output type is restricted to 14 alphanumeric characters and dashes.
The output type names should either match the types of available (local or remote) printers, or match the input types handled by other filters. The LP print service groups filters in a shell pipeline if it finds that several passes by different filters are needed to convert a file. You will unlikely need this level of sophistication, but the LP print service allows it. Try to find a set of filters that takes as input types all the different files the users might want printed, and that converts those files directly into file types the printer can handle.
Printer types is a list of the types of printers into which the print filter can convert files. For most printers and filters, you can leave this part of the filter definition blank, because this list is identical to the list of output types. However, this list can be different. For example, you could have a printer with a single printer type for purposes of initialization. However, that printer can recognize several different file content types. Essentially, this printer has an internal filter that converts the various file types into a filter type that it can handle. Thus, a filter might produce one of several output types that match the file types that the printer can handle. The print filter should be marked as working with that printer type.
As another example, you might have two different models of printers that are listed as accepting the same file types. Due to slight differences in manufacture, however, one printer deviates in the results it produces. You label the printers as being of different printer types, say A and B, where B is the printer that deviates. You create a filter that adjusts files to account for the deviation produced by printers of type B. Because this filter is needed only for those printer types, you would list this filter as working only on type B printers.
A print filter normally can work with all printers that accept its output, so you can usually skip this part of the filter definition.
You might, however, have some printers that are inappropriate for the output that the filter produces. For example, you might want to dedicate one printer for fast turnaround, only sending files that require no filtering to that printer. Other printers of identical type can be used for files that need extensive filtering before they can be printed.
The LP print service recognizes fast filters and slow filters, as described in Types of Filters.
Slow filters that are invoked by printing modes, by using the lp
y command, must be run on the system from which the print request originated.
The LP print service cannot pass values for modes to print servers. It can, however,
match a file content type (specified after the -T option of the lp command) to a content type on a print server. Therefore, if you want
to activate special modes on a print server, you must specify content types that permit
the LP print service to match input types and output types.
Options specify how different types of information are converted into command-line arguments to the filter command. This information can include specifications from a user (with the print request), the printer definition, and the specifications implemented by any filters that are used to process the request.
There are 13 sources of information for defining print filter options, each of which is represented by a keyword. Each option is defined in a template. A template is a statement in a filter definition that defines an option to be passed to the filter command, based on the value of one of the filter characteristics
The options specified in a filter definition can include none, all, or any subset of the 13 keywords. In addition, a single keyword can be defined more than once, if multiple definitions are required for a complete filter definition. The following table contains descriptions of the 13 keywords available for defining options in a print filter definition.
Table 6.8. Keywords for Print Filter Options
|
Keyword |
Patterns |
Example | |
|---|---|---|---|
|
Content type (input) |
|
|
|
|
Content type (output) |
|
|
|
|
Printer type |
|
|
|
|
Printer name |
|
|
|
|
Character pitch |
|
|
|
|
Line pitch |
|
|
|
|
Page length |
|
|
|
|
Page width |
|
|
|
|
Pages to print |
|
|
|
|
Character set |
|
|
|
|
Form name |
|
|
|
|
Number of copies |
|
|
|
|
|
|
|
A print filter definition can include more than one template. Multiple templates are entered on a single line and separated with commas, or they are entered on separate lines, preceded by the Options: prefix.
The format of a template is as follows: keyword
pattern = replacement
The keyword identifies the type of option being registered for a particular characteristic of the filter.
The pattern is a specific option for the keyword.
The replacement is what happens when the keyword
has the noted value.
For an example of how an option is defined for a particular filter, suppose you want to have the print service scheduler assign print requests to filters following this criteria:
If the type of OUTPUT to be produced by the filter
is impress, then pass the -I option to the filter.
If the type of OUTPUT to be produced by the filter
is postscript, then pass the -P option to the
filter.
To specify these criteria, provide the following templates as options to the lpfilter command:
Options: OUTPUT impress=-I, OUTPUT postscript=-P
If the Options line becomes too long, put each template on a separate line, as follows:
Options: OUTPUT impress=-I Options: OUTPUT postscript=-P
In both templates, the keyword is defined as OUTPUT. In the first template, the pattern is impress,
and the value of the replacement is
-I. In the
second template, the value of pattern is postscript, and the value of replacement is
P.
To determine which values to supply for each type of template (that is, for
the pattern and replacement arguments
for each keyword), consider the following:
The values for the INPUT templates come from the
file content type that needs to be converted by the filter.
The values for the OUTPUT templates come from the
output type that has to be produced by the filter.
The value for the TERM template is the printer
type.
The value for the PRINTER template is the name
of the printer that will print the final output.
The values for the CPI, LPI, LENGTH, and WIDTH templates come from the user's print
request, the form being used, or the default values for the printer.
The value for the PAGES template is a list of pages
that should be printed. Typically, the value consists of a list of page ranges separated
by commas. Each page range consists of a pair of numbers separated by a dash, or a
single number. (For example, 1–5,6,8,10 indicates pages 1 through 5, plus pages
6, 8, and 10.) However, whatever value was given in the
P option
to a print request is passed unchanged.
The value for the CHARSET template is the name
of the character set to be used.
The value for the FORM template is the name of
the form requested by the lp
f command (the command
used to submit a print request).
The value of the COPIES template is the number of copies of the file to print. If the filter uses this template, the LP print service reduces to one the number of copies of the filtered file it prints. This “single copy” includes the multiple copies that are produced by the filter.
The value of the MODES template comes from the lp
y command. Because a user can specify several
y options, the MODES template might have several values.
The values are applied in the left-to-right order given by the user.
The replacement option shows how the value of a template
should be given to the filter program. This part typically a literal option, sometimes
with the placeholder asterisk (*) included to show where the value goes. The pattern and replacement options also can
use the regular expression syntax of the ed command for more complex
conversion of user input options into filter options. All regular expression syntax
of ed is supported, including the \( ... \) and \n constructions. These constructions can be used to extract portions of
the pattern option for copying into the replacement option, and the & option, which can be used
to copy the entire pattern option into the replacement option. For more information, see the
ed
(
1
)
man page.
If a comma or an equal sign (=) is included in a pattern or a replacement option, precede
it with a backslash (\). A backslash in front of any of these characters
is removed when the pattern or replacement option is used.
How to Create a New Print Filter
Log in as superuser, lp, or
assume an equivalent role on the print server.
Create a print filter program.
For
information on print filter programs, see Writing a Print Filter Program. By convention, filter programs for PostScript printers
are located in the /usr/lib/lp/postscript directory. You should
put programs you create under the /usr/lib/lp directory in a
directory of your choosing.
Create a print filter definition.
For information on print filter definitions, see Creating a Print Filter Definition. You should save the print filter definition
in a text file. By convention, filter definitions are located in the /etc/lp/fd directory and are identified with the .fd suffix.
Add the print filter to a print server.
For instructions, see How to Add a Print Filter.
Example 6.5. Creating a New Print Filter
This example shows a print filter definition that is used to convert N37 or Nlp to simple.
Input types: N37, Nlp, simple Output types: simple Command: /usr/bin/col Options: MODES expand = -x Options: INPUT simple = -p -f
In this example, the print filter program is named col. Once
you add the new print filter to a print server, a user's print requests are handled
as follows:
When a user types the following command:
$ lp -y expand report.doc
The print filter program is run with the following arguments to convert the file:
/usr/bin/col -x -p -f
When a user types the following command:
$ lp -T N37 -y expand report.doc
The print filter program is run with the following arguments to convert the file:
/usr/bin/col -x
Example 6.6. Creating a New Print Filter by Converting One Input Type to Another Output Type
This example shows a print filter definition that is used to convert from troff to a PostScript file.
Input types: troff Output types: postscript Printer types: PS Filter type: slow Command: /usr/lib/lp/postscript/dpost Options: LENGTH * = -l* Options: MODES port = -pp, MODES land = -pl Options: MODES group \=\([1-9]\) = -n\l
In this example, the filter program is named dpost. The
filter program takes one input type, troff, produces a postscript output, and works with any printer of type PS (PostScript).
Users need to give just the abbreviation port or land when they ask for the paper orientation to be in either portrait mode or
landscape mode. Because these options are not intrinsic to the LP print service, users
must specify them by using the lp -y command.
After you add the new print filter to a print server, print requests will be handled as follows:
When a user types the following command to submit a troff file type for printing on a PostScript printer (type PS),
with requests for landscape orientation and a page length of 60 lines:
$ lp -T troff -o length=60 -y land -d luna ch1.doc
The print filter program dpost is run with the following arguments to convert the file:
/usr/lib/lp/postscript/dpost -l60 -pl
When a user enters the following command:
$ lp -T troff -y group=4 -d luna ch1.doc
The print filter program dpost is run with the following arguments to convert the file:
/usr/lib/lp/postscript/dpost -n4