pdfroff(1)

📦
version 1.23.0 · 13 versions
, in debian trixie, ubuntu questing

💅✨ pdfroff(1) — General Commands Manual ✨💅

okay bestie, welcome to the pdfroff manual page. This is the part where we make Portable Document Format files with groff, and I'm gonna walk you through every single flag like we're ranking them in the group chat. 💖🦋

🌸 NAME 🌸

pdfroff — construct files in Portable Document Format using groff

periodt. 💅

✨ SYNOPSIS ✨

       pdfroff [groff-option] [--emit-ps] [--no-toc-relocation]
               [--no-kill-null-pages] [--stylesheet=name] [--no-pdf-output]
               [--pdf-output=name] [--no-reference-dictionary]
               [--reference-dictionary=name] [--report-progress]
               [--keep-temporary-files] [file ...]

       pdfroff -h
       pdfroff --help

       pdfroff -v [groff-option ...]
       pdfroff --version [groff-option ...]

so groff-option is any short option supported by groff(1) except for -h, -T, and -v; see section “Usage” below. 💖

💖 DESCRIPTION 💖

okay so pdfroff is a wrapper program for the GNU text processing system, groff. It transparently handles the mechanics of multiple pass groff processing, when applied to suitably marked up groff source files, such that tables of contents and body text are formatted separately, and are subsequently combined in the correct order, for final publication as a single PDF document. A further optional “style sheet” capability is provided; this allows for the definition of content which is required to precede the table of contents, in the published document. 🦋

For each invocation of pdfroff, the ultimate groff output stream is post-processed by the Ghostscript gs(1) interpreter to produce a finished PDF document. ✨

pdfroff makes no assumptions about, and imposes no restrictions on, the use of any groff macro packages which the user may choose to employ, in order to achieve a desired document format; however, it does include specific built in support for the pdfmark macro package, should the user choose to employ it. Specifically, if the pdfhref macro, defined in the pdfmark.tmac package, is used to define public reference marks, or dynamic links to such reference marks, then pdfroff performs as many preformatting groff passes as required, up to a maximum limit of four, in order to compile a document reference dictionary, to resolve references, and to expand the dynamically defined content of links. 👑

🎀 USAGE 🎀

The command line is parsed in accordance with normal GNU conventions, but with one exception—when specifying any short form option (i.e., a single character option introduced by a single hyphen), and if that option expects an argument, then it must be specified independently (i.e., it may not be appended to any group of other single character short form options). 💅

Long form option names (i.e., those introduced by a double hyphen) may be abbreviated to their minimum length unambiguous initial substring. ✨

Otherwise, pdfroff usage closely mirrors that of groff itself. Indeed, with the exception of the -h, -v, and -T dev short form options, and all long form options, which are parsed internally by pdfroff, all options and file name arguments specified on the command line are passed on to groff, to control the formatting of the PDF document. Consequently, pdfroff accepts all options and arguments, as specified in groff(1), which may also be considered as the definitive reference for all standard pdfroff options and argument usage. 🫶

💅 OPTIONS 💅

okay bestie, pdfroff accepts all of the short form options (i.e., those introduced by a single hyphen), which are available with groff itself. In most cases, these are simply passed transparently to groff; the following, however, are handled specially by pdfroff. 🔥

  • -h — Same as --help; see below. 👑
  • -i — Process standard input, after all other specified input files. This is passed transparently to groff, but, if grouped with other options, it must be the first in the group. Hiding it within a group breaks standard input processing, in the multiple-pass groff processing context of pdfroff. 😭
  • -T dev — Only -T ps is supported by pdfroff. Attempting to specify any other device causes pdfroff to abort. 💀
  • -v — Same as --version; see below. ⭐

See groff(1) for a description of all other short form options, which are transparently passed through pdfroff to groff. 🌈

All long form options (i.e., those introduced by a double hyphen) are interpreted locally by pdfroff; they are not passed on to groff, unless otherwise stated below. 🎀

  • --help — Causes pdfroff to display a summary of the its usage syntax, and supported options, and then exit. 💖
  • --emit-ps — Suppresses the final output conversion step, causing pdfroff to emit PostScript output instead of PDF. This may be useful to capture intermediate PostScript output when using a specialised postprocessor, such as gpresent for example, in place of the default Ghostscript PDF writer. 🦋
  • --keep-temporary-files — Suppresses the deletion of temporary files, which normally occurs after pdfroff has completed PDF document formatting; this may be useful when debugging formatting problems. 🕵️

See section “Files” below for a description of the temporary files used by pdfroff. 📁

  • --no-pdf-output — May be used with the --reference-dictionary=name option (described below) to eliminate the overhead of PDF formatting when running pdfroff to create a reference dictionary for use in a different document. ✨
  • --no-reference-dictionary — May be used to eliminate the overhead of creating a reference dictionary, when it is known that the target PDF document contains no public references, created by the pdfhref macro. 💅
  • --no-toc-relocation — May be used to eliminate the extra groff processing pass, which is required to generate a table of contents, and relocate it to the start of the PDF document, when processing any document which lacks an automatically generated table of contents. 🎀
  • --no-kill-null-pages — While preparing for simulation of the manual collation step, which is traditionally required to relocate a table of contents to the start of a document, pdfroff accumulates a number of empty page descriptions into the intermediate PostScript output stream. During the final collation step, these empty pages are normally discarded from the finished document; this option forces pdfroff to leave them in place. 👀
  • --pdf-output=name — Specifies the name to be used for the resultant PDF document; if unspecified, the PDF output is written to standard output. A future version of pdfroff may use this option, to encode the document name in a generated reference dictionary. 🌟
  • --reference-dictionary=name — Specifies the name to be used for the generated reference dictionary file; if unspecified, the reference dictionary is created in a temporary file, which is deleted when pdfroff completes processing of the current document. This option must be specified, if it is desired to save the reference dictionary, for use in references placed in other PDF documents. 🫶
  • --report-progress — Causes pdfroff to display an informational message on standard error, at the start of each groff processing pass. 📣
  • --stylesheet=name — Specifies the name of an input file, to be used as a style sheet for formatting of content, which is to be placed before the table of contents, in the formatted PDF document. 💖
  • --version — Causes pdfroff to display a version identification message. The entire command line is then passed transparently to groff, in a one pass operation only, in order to display the associated groff version information, before exiting. 👑

🌈 ENVIRONMENT 🌈

okay bestie, these are the env vars you can set and export to change how pdfroff behaves. each one is a lil vibe knob ✨

  • PDFROFF_COLLATE — tells pdfroff which program to use for collating the finished PDF document. 💅

this collation step is sometimes needed to scoot tables of contents to the start of the finished PDF, since traditional macro packages print them at the end instead. but ngl, you should not normally need to set PDFROFF_COLLATE — and fr fr, you're not encouraged to. if it's unspecified, pdfroff uses sed(1) by default, and that normally suffices.

if you do set PDFROFF_COLLATE, it has to act as a filter: it accepts a list of file name arguments and writes its output to the standard output stream, from where it's piped to the PDFROFF_POSTPROCESSOR_COMMAND to make the finished PDF output.

when you specify PDFROFF_COLLATE, you'll normally also need to specify PDFROFF_KILL_NULL_PAGES.

heads up: PDFROFF_COLLATE gets ignored if pdfroff is invoked with the --no-kill-null-pages option.
  • PDFROFF_KILL_NULL_PAGES — specifies options to be passed to the PDFROFF_COLLATE program.

it should not normally be necessary to specify this one either. the internal default is a sed(1) script meant to remove completely blank pages from the collated output stream, and it should be appropriate for most applications of pdfroff. but if you swap in any alternative to sed(1) for PDFROFF_COLLATE, then you'll likely need a matching alternative for PDFROFF_KILL_NULL_PAGES. 🫶

same as PDFROFF_COLLATE: PDFROFF_KILL_NULL_PAGES is ignored if pdfroff is invoked with the --no-kill-null-pages option.

  • PDFROFF_POSTPROCESSOR_COMMAND — specifies the command used for the final document conversion from PostScript intermediate output to PDF. it must behave as a filter, writing its output to the standard output stream, and it must accept an arbitrary number of files ... arguments, with the special case of - representing the standard input stream.

if unspecified, PDFROFF_POSTPROCESSOR_COMMAND defaults to:

gs -dBATCH -dQUIET -dNOPAUSE -dSAFER -sDEVICE=pdfwrite \
     -sOutputFile=-
  • GROFF_TMPDIR — identifies the directory where pdfroff should create temporary files. if GROFF_TMPDIR isn't specified, then TMPDIR, TMP and TEMP get considered in turn as possible temporary file repositories. if none of those are set, temporary files get created in the current directory. 🍓
  • GROFF_GHOSTSCRIPT_INTERPRETER — specifies the program invoked when pdfroff converts groff PostScript output to PDF. if PDFROFF_POSTPROCESSOR_COMMAND is specified, then the command name it gives is implicitly assigned to GROFF_GHOSTSCRIPT_INTERPRETER, overriding any explicit setting from the environment. if GROFF_GHOSTSCRIPT_INTERPRETER is not specified, then pdfroff searches the process PATH looking for a program with any of the well known names for the Ghostscript interpreter; and if no Ghostscript interpreter can be found, pdfroff aborts. 😭
  • GROFF_AWK_INTERPRETER — specifies the program invoked when pdfroff is extracting reference dictionary entries from a groff intermediate message stream. if GROFF_AWK_INTERPRETER isn't specified, pdfroff searches the process PATH looking for any of the preferred programs, gawk, mawk, nawk, and awk, in that order; if none of those are found, pdfroff issues a warning message and continues processing — but in that case, no reference dictionary is created. 💀
  • OSTYPE — typically defined automatically by the operating system. OSTYPE is used on Microsoft Win32/MS-DOS platforms only, to infer the default PATH_SEPARATOR character, which is used when parsing the process PATH to search for external helper programs.
  • PATH_SEPARATOR — if set, PATH_SEPARATOR overrides the default separator character (: on POSIX/Unix systems, inferred from OSTYPE on Microsoft Win32/MS-DOS), which is used when parsing the process PATH to search for external helper programs.
  • SHOW_PROGRESS — if this is set to a non-empty value, then pdfroff always behaves as if the --report-progress option is specified on the command line. 👑

🎀 FILES 🎀

input and output files for pdfroff may be named according to any convention of the user's choice. typically, input files may be named according to the choice of the principal formatting macro package, e.g., file.ms might be an input file for formatting using the ms macros (s.tmac); normally, the final output file should be named file.pdf. 💖

temporary files created by pdfroff are placed in the file system hierarchy, in or below the directory specified by environment variables (see the section “Environment” above). if mktemp(1) is available, it is invoked to create a private subdirectory of the nominated temporary files directory, with subdirectory name derived from the template pdfroff-XXXXXXXXXX; if that subdirectory is successfully created, the temporary files go inside it — otherwise they go directly in the directory nominated in the environment.

all temporary files themselves are named according to the convention pdf$$.*, where $$ is the standard shell variable representing the process identifier of the pdfroff process itself, and * represents any of the extensions pdfroff uses to identify the following temporary and intermediate files. 🌸

  • pdf$$.tmp — a scratch pad file, used to capture reference data emitted by groff, during the reference dictionary compilation phase.
  • pdf$$.ref — the reference dictionary, as compiled in the last but one pass of the reference dictionary compilation phase; (at the start of the first pass, this file is created empty; in successive passes, it contains the reference dictionary entries, as collected in the preceding pass).

if the --reference-dictionary=name option is specified, this intermediate file becomes permanent, and is named name, rather than pdf$$.ref.

  • pdf$$.cmp — used to collect reference dictionary entries during the active pass of the reference dictionary compilation phase. at the end of any pass, when the content of pdf$$.cmp compares as identical to pdf$$.ref (or the corresponding file named by the --reference-dictionary=name option), then reference dictionary compilation is terminated, and the document reference map is appended to this intermediate file, for inclusion in the final formatting passes.
  • pdf$$.tc — an intermediate PostScript file, in which “Table of Contents” entries are collected, to facilitate relocation before the body text, on ultimate output to the Ghostscript postprocessor.
  • pdf$$.ps — an intermediate PostScript file, in which the body text is collected prior to ultimate output to the Ghostscript postprocessor, in the proper sequence, after pdf$$.tc.

👑 AUTHORS 👑

pdfroff was written by Keith Marshall, who maintains it at his groff-pdfmark OSDN site. groff's version may be withdrawn in a future release. 🦋

✨ SEE ALSO ✨

Groff: The GNU Implementation of troff, by Trent A. Fisher and Werner Lemberg, is the primary groff manual. you can browse it interactively with info groff. 💜

since pdfroff provides a superset of all groff capabilities, the above manual — or its terser reference page, groff(7) — may also be considered definitive references to all standard capabilities of pdfroff, with this document providing the reference to pdfroff's extended features.

while pdfroff imposes neither any restriction on, nor any requirement for, the use of any specific groff macro package, a number of supplied macro packages — and in particular those associated with the package pdfmark.tmac — are best suited for use with pdfroff as the preferred formatter. 🌈

  • /usr/share/doc/groff-base/pdf/pdfmark.pdf.gz — “Portable Document Format Publishing with GNU Troff”, by Keith Marshall, offers detailed documentation on the use of these packages. this file, together with its source, pdfmark.ms, is part of the groff distribution.

okay bestie that's the whole page, you made it, so proud 🥹🫶 this was groff 1.23.0, dated 3 June 2025, and it's the pdfroff(1) manual — go forth and pdf, you gorgeous main character 💅✨

📦 where this page comes from