Tcl Library Source Code

Bounty program for improvements to Tcl and certain Tcl packages.

[ Main Table Of Contents | Table Of Contents | Keyword Index | Categories | Modules | Applications ]


rcs - RCS low level utilities

Table Of Contents


package require Tcl 8.4
package require rcs ?0.1?

::rcs::text2dict text
::rcs::dict2text dict
::rcs::file2dict filename
::rcs::dict2file filename dict
::rcs::decodeRcsPatch text
::rcs::encodeRcsPatch pcmds
::rcs::applyRcsPatch text pcmds


The Revision Control System, short RCS, is a set of applications and related data formats which allow a system to persist the history of changes to a text. It, and its relative SCCS are the basis for many other such systems, like CVS, etc.

This package does not implement RCS.

It only provides a number of low level commands which should be useful in the implementation of any revision management system, namely:

  1. The conversion of texts into and out of a data structures which allow the easy modification of such text by patches, i.e. sequences of instructions for the transformation of one text into an other.

  2. And the conversion of one particular format for patches, the so-called RCS patches, into and out of data structures which allow their easy application to texts.



A text dictionary is a dictionary whose keys are integer numbers and text strings as the associated values. The keys represent the line numbers of a text and the values the text of that line. Note that one text can have many representations as a dictionary, as the index values only have to be properly ordered for reconstruction, their exact values do not matter. Similarly the strings may actually span multiple physical lines.

The text

Hello World,
how are you ?
Fine, and you ?

for example can be represented by

{{1 {Hello World,}} {2 {how are you ?}} {3 {Fine, and you ?}}}


{{5 {Hello World,}} {8 {how are you ?}} {9 {Fine, and you ?}}}


{{-1 {Hello World,
how are you ?}} {4 {Fine, and you ?}}}

The first dictionary is the canonical representation of the text, with line numbers starting at 1, increasing in steps of 1 and without gaps, and each value representing exactly one physical line.

All the commands creating dictionaries from text will return the canonical representation of their input text. The commands taking a dictionary and returning text will generally accept all representations, canonical or not.

The result of applying a patch to a text dictionary will in general cause the dictionary to become non-canonical.


A patch is in general a series of instructions how to transform an input text T into a different text T', and also encoded in text form as well.

The text format for patches understood by this package is a very simple one, known under the names RCS patch or diff -n format.

Patches in this format contain only two different commands, for the deletion of old text, and addition of new text. The replacement of some text by a different text is handled as combination of a deletion following by an addition.

The format is line oriented, with each line containing either a command or text data associated with the preceding command. The first line of a RCS patch is always a command line.

The commands are:

Note that the line indices start always refer to the text which is transformed as it is in its original state, without taking the precending changes into account.

Note also that the instruction have to be applied in the order they occur in the patch, or in a manner which produces the same result as in-order application.

This is the format of results returned by the command ::rcs::decodeRcsPatch and accepted by the commands ::rcs::encodeRcsPatch and ::rcs::appplyRcsPatch resp. Note however that the decoder will strip no-op commands, and the encoder will not generate no-ops, making them not fully complementary at the textual level, only at the functional level.

And example of a RCS patch is

d1 2
d4 1
a4 2
The named is the mother of all things.

a11 3
They both may be called deep and profound.
Deeper and more profound,
The door of all subtleties!


Patch command lists (sort: PCL's) are the data structures generated by patch decoder command and accepted by the patch encoder and applicator commands. They represent RCS patches in the form of Tcl data structures.

A PCL is a list where each element represents a single patch instruction, either an addition, or a deletion. The elements are lists themselves, where the first item specifies the command and the remainder represent the arguments of the command.

This is the format returned by the patch decoder command and accepted as input by the patch encoder and applicator commands.

An example for a patch command is shown below, it represents the example RCS patch found in section RCS PATCH FORMAT.

{{d 1 2} {d 4 1} {a 4 {The named is the mother of all things.

}} {a 11 {They both may be called deep and profound.
Deeper and more profound,
The door of all subtleties!}}}

Bugs, Ideas, Feedback

This document, and the package it describes, will undoubtedly contain bugs and other problems. Please report such in the category rcs of the Tcllib Trackers. Please also report any ideas for enhancements you may have for either package and/or documentation.

When proposing code changes, please provide unified diffs, i.e the output of diff -u.

Note further that attachments are strongly preferred over inlined patches. Attachments can be made by going to the Edit form of the ticket immediately after its creation, and then using the left-most button in the secondary navigation bar.


struct, textutil


CVS, RCS, RCS patch, SCCS, diff -n format, patching, text conversion, text differences


Text processing


Copyright © 2005, Andreas Kupries
Copyright © 2005, Colin McCormack