---
name: "dviasm"
section: "1"
platform: "linux"
package: "texlive-doc"
version: "2026.1-1"
release: "Arch Linux 2026-11-01"
platforms: ["linux"]
url: "https://new.carta.tech/dviasm"
---

# dviasm(1)

a TeX utility program for editing DVI files directly

## NAME

DVIasm - a TeX utility program for editing DVI files directly

## SYNOPSIS

**dviasm** \[\<*options*\>\] \<*dvi\_file*\>\|\<*dvi\_dump\_file*\>

## DESCRIPTION

**DVIasm** is a TeX utility program which is designed for editing DeVice-Independent (DVI) files directly. It consists of a single Python script, dviasm.py, in a human readable text format. It runs on any platform in which Python 3 is installed.

- Features of **DVIasm** includes:

  -

  1\) Disassemble a DVI file (or XeTeX XDV file) into a human-readable text file, the contents of which are easy to modify.

-

  2\) Assembles the output text file back to the binary format.

- There are several alternatives for dumping and editing DVI files:

  -

  1\) **DVItype** developed by Donald E. Knuth supports one-way conversion from DVI to a text format. There are also some derivatives: **pDVItype** for pTeX, **upDVItype** for upTeX, **ODVItype** for Omega.

-

  2\) The programs **dv2dt** and **dt2dv** can be used in pairs to allow two-way conversion between DVI and the DVI Text Language (DTL).

-

  3\) The program **dvispc** (part of DVIOUT previewer) also has an ability to convert between DVI and text.

Among those, **DVIasm** is designed to allow users to edit DVI files easily, for example by unifying \`\`right1''-\`\`right4'' to a single command \`\`right'' which can be used regardless of the amount of move.

## OPTIONS

- **--version**

  Show program's version number and exit.

- **-h**, **--help**

  Show this help message and exit.

- **-u** \<*STR*\>, **--unit=**\<*STR*\>

  Set unit \[default: \`pt'\]. Allowed values are: \`sp', \`pt', \`bp', \`mm', \`cm' and \`in'.

- **-o** \<*FILE*\>, **--output=**\<*FILE*\>

  Set filename for output instead of stdout.

- **-e** \<*STR*\>, **--encoding=**\<*STR*\>

  Set encoding for input/output of dumped text \[default: \`utf8'\]. Allowed values are: \`ascii', \`latin1', \`utf8', \`sjis' and \`eucjp'. When used with the option **-p** (or **--ptex**), allowed values are only \`utf8', \`sjis' and \`eucjp'.

  \# NOTE: This feature does not support dumping through stdout.

- **-x** \<*STR*\>, **--xxx-encoding=**\<*STR*\>

  Set encoding for interpreting \`xxx:' strings \[default: \`none'\]. Allowed values are: \`none', \`utf8', \`sjis' and \`eucjp'.

  This option allows users to specify the correct encoding for dumping/compiling the contents of \\special.

  - Typical usages are as follows:

    -

    1\) By default, the contents of \\special are dumped/compiled as byte-to-byte escape sequences in \`\\x..' format. This would suffice when you don't need to read/edit the contents of \\special.

  -

    3\) For XeTeX and upTeX users, \`**-x utf8**' would be useful since these engines use UTF-8 to encode strings inside \\special commands.

  -

    2\) For pTeX users with ISO-2022-JP-encoded DVI, it would be helpful to select the appropriate option \`**-x eucjp**' (for Unix) or \`**-x sjis**' (for Windows) to be consistent with the internal Kanji encoding used by pTeX engine.

- **-t** \<*INT*\>, **--tabsize=**\<*INT*\>

  Set tab size for push/pop \[default: 2\].

- **-p**, **--ptex**

  Dump and compile ISO-2022-JP-encoded DVI for Japanese pTeX. This option is required to decode or encode Japanese characters being typeset (using \`set2' or \`put2') properly.

- **-s** \<*STR*\>, **--subfont=**\<*STR*\>

  Specify the list of fonts with UCS2 subfont scheme in the comma-separated format. If \<*STR*\> is empty, the pre-defined internal subfont list is disabled.

  This option would be useful for example when using CJK package.

## EXAMPLE

Documentation of **DVIasm** by the original author Jin-Hwan Cho is published as:\
      https://tug.org/TUGboat/Articles/tb28-2/tb89cho.pdf (TUGboat)\
      http://ajt.ktug.kr/assets/2008/5/1/0201cho.pdf (Asian Journal of TeX)

## AUTHORS

The first version was written in Python 2 by Jin-Hwan Cho in 2007. Khaled Hosny started to support dumping/compiling XeTeX XDV format in 2014. Arthur Reutenauer ported the script to Python 3 in 2019. Current maintainer of **DVIasm** is Hironobu Yamashita, since 2020.

## BUG REPORT

Please visit the GitHub repository \<https://github.com/aminophen/dviasm\>.

## SEE ALSO

[**dvitype**(1)](/dvitype/1.md), [**dv2dt**(1)](/dv2dt/1.md), [**dt2dv**(1)](/dt2dv/1.md), [**dvispc**(1)](/dvispc/1.md).