- file history

filename... option... -Help -VERSion

The program is used to keep track of the successive versions of a file. Using this program, you can remember all of your changes to a file, and get back any one of the old versions. The uses of this ability are:

1.

You can make a series of tentative edits to the file, and if necessary back up to the last "good" edit.

2.

You can delete old subroutines and code from your file which are obsolete, but still be able to get them back in the future in case a need for them arises.

3.

You can compare two versions of the file to see how you fixed some old problem, so that you can check up on the correctness of the fix at a later date.

4.

You get a record of your remarks for each version, so that you can quickly know what bugs were fixed, and what features were implemented.

5.

The date the file was last edited can be automatically stored in the file.

The program manipulates modules. A module is simply any text file that you are interested in keeping versions of. For example, a source file doit.c is a module, and so is a documentation file howto.doc. The module name includes the suffix of the file (as in the above examples). However, pathnames are not part of a module name, so that /usr/dbell/bar.c cannot be a legal module name. A module name is limited to 12 characters since the program needs two extra characters for its own purpose.

Keyword Substitution

The following options are understood:

-Path pathname

Modules are stored in a directory, called the module storage directory. The default directory is FHIST, and therefore is located relative to your current directory. This is convenient when you are in a directory containing many modules, and you want a local storage directory to contain just those modules. If you use the -p option, then you can locate the storage directory anywhere you choose. This is useful if you choose to have a common storage directory for all of your files, independent of where they actually are used.

The files inside of the storage directory should not be changed by you. Doing so will probably corrupt your edit history, causing errors when you extract old revisions. For your information, though, each module is stored as two files in the directory. The one with the .s suffix is a copy of the newest version of the module, with one extra line at the beginning. The one with the .e suffix is the edit history of the module, and contains the information needed to extract previous revisions of the module. Thus if the edit history is ever corrupted, you will at least have the most recent version of the module.

-MaKe_Path

This option may be used to request that the path directory be created automatically if it does not yet exist. This works for both the directory specified by the -Path option, and for the default. Intermediate directories will also be created if necessary.

-BINary

This option may be used to specify that the file is binary, that it may contain NUL characters. It is essential that you have consistent presence or absence of the -BINary option for each file when combined with the -CReate, -Update, -Conditional_Update and -Extract options. Failure to do so will produce inconsistent results. Note: this is different behaviour to the fcomp(1) option of the same name. Note: the -BINary option does not imply the -No-Keywords option.

-CReate

To use the program for the first time, you need to create your storage directory. Therefore, cd to the directory where you want it to be, which is probably the directory containing the modules you want to save the revisions of. Then create the directory FHIST (or some other name if you don't want to use the default name).

To start using a module under , you must first use the -CReate option. This creates the initial edit for that module in the storage directory, with the contents of the specified module as the initial edit. Thus, if you have a source file prog.c, then the command:

 prog.c -create

The -CReate option may be combined with the -Update or -Conditional_Update options to create the file if required.

-Update

To save another revision of the module, you use the -Update option. This updates the files in the storage directory to include the latest changes. Remarks are again asked for so that you can document why you made this edit. Thus, to continue the example, after editing prog.c, the command:

 prog.c -u

will save the changes as a new edit. This command compares the newest version of the module to the previous version, saves the differences in the .e file, and copies the new source to the .s file. At this point, you can once again delete the prog.c file if desired, and later get back either of the two versions of the program.

The program handles quota or disk full problems during a create or update operation without damage occurring to the edit history files. If an edit cannot be completed because of such problems, the edit is backed out completely, and you will get an error message about the disk problem. There is no need for any error recovery in this case, other than retrying the update when more disk space is available. The program also disables signals during the critical file operations, so you do not have to worry about damaging the edit history files because of attempts to quit out of the program.

The -CReate option may be combined with the -Update or -Conditional_Update options to create the file if required.

-Input filename

In either the -CReate or -Update options, the file containing the new version of the module defaults to the same name as the module. In the example, the module prog.c was created and updated from the data in the file prog.c. When you wish the data to come from some other file, you can use the -Input option, which specifies the input file to use for the data. For example, if you wanted to update prog.c, but from a filename called newprog.c, then the command:

 prog.c -u -i newprog.c

would save a new revision of module prog.c, but with the data that was in the file newprog.c. In this case, the file prog.c does not have to exist, and isn't referenced even if it did exist. Again, once the update is complete, you could delete the newprog.c file if desired and then later you can retrieve its contents.

-Remarks

Remarks can be read from a file instead of from the terminal. The -Remarks option can be used to specify a file name containing the remarks. If there is no file name following the -Remarks option, then no remarks at all are used. The command:

 prog.c -u -r

would create a new revision of prog.c without asking for or saving any remarks about the edit.

-Extract [ edit ]

To retrieve a previous revision of a module, you specify the name of the module and use the -Extract option to specify the edit number you want retrieved. Edit numbers are assigned sequentially starting with 1. Thus the initial version of the module has edit number 1, the first revision has edit number 2, and so on until the latest revision. If the -Extract option is not used, or if no edit number is supplied for it, then the latest edit number is extracted. Therefore, this is the default action if no options at all are specified.

Edit numbers can also be zero, negative, or be a name with an optional offset. The number zero represents the latest edit number, and negative numbers indicate edit numbers backwards from the latest edit number. Edit names represent edit numbers whose name had been set by using the -Name option. For example, if edit number 10 was associated with the name foo, then the edit name foo represents 10, foo-4 represents edit number6, and foo+2 represents edit number 12. The special reserved names oldest and newest refer to the oldest and newest versions of the module in the edit history.

As an example of retrievals, assume that you have saved ten versions of the module prog.c. The following commands will then extract the versions of the file with the specified edit numbers:

prog.c

version 10 (the latest)

prog.c -e 9

version 9 (the version just prior)

prog.c -e oldest

version 1 (the oldest version)

prog.c -e -2

version 8 (latest version - 2)

The output filename is again defaulted to the module name. So when the module prog.c is extracted, the specified version of the module is written to the prog.c file.

In order to prevent accidental overwriting of a file, the program will by default ask you if overwriting is permitted if that would occur. A common mistake is to edit prog.c, and then try to update the module, but forget to specify the -u option. Then the program would try to extract the newest version of the module, and thus overwrite the file with the new changes. Asking the question allows you to notice your mistake, and prevent the overwriting.

-Output filename

You can change the output filename using the -Output option. Thus, the command:

 prog.c -o newprog.c

will extract the latest version of the module prog.c, and put it into the file newprog.c. Once again, the file "prog.c" is ignored, whether or not it existed.

-Force_Write

This option will force overwriting of the file, thus never asking you if overwriting is permitted. This is often useful in shell scripts, or when you are sure that you want to overwrite any existing file.

-No_Write

This option is the no-overwrite option, and will cause any existing files to not be overwritten, again without asking you. This is useful if you already have some of the modules in your directory, and you want to extract the rest of the modules without overwriting the ones you already have. Specifying both -Fore_Write and -No_Write is an error.

-Terminal [ edit ]

This option is used to output an extracted module to the standard output, instead of writing it to a file. This is useful in order to view the beginning of a version of the file. This can be interrupted if you do not want to see the whole file.

-Modify number

When extracting a file, the program looks for and updates special character sequences in the first few lines of the file. These special sequences are used for documentation purposes, such as describing the edit number the file is from. For speed of extraction and updating, these sequences are usually limited to the first 25 lines of the file, since the program then does not have to examine the entire file. The -Modify option can be used to change the number of lines to be modified from the default value of 25. Specifying zero totally disables the special character sequences, whereas specifying a very large number will cause the sequences to be checked for each line of the file (and thus slow the program down).

Each special sequence is of the form [# keyword value, keyword value, ..., keyword value #] , where each keyword describes an item, and each value is the value for the preceding keyword. The keywords can be in upper or lower case, or both. The single space following the [#, following each comma, and preceding the #] must be present. If the sequence is wrong, an unknown keyword is used, the line is longer than 200 characters, or more than four keywords are used, then the whole line will not be changed. The current keywords which can be used are the following:

edit

The edit number

date

The date that the edit was created

user

The user name of the user who created the edit

module

The module name

In order to use this special character sequence, you simply insert it into your module inside of a comment (within the first few lines). When this is done, the value parts of the sequence can be null. For example, if you want to put a special sequence into a program called delete.c, then you could edit the first few lines as follows:

/*
 * Delete - program to delete files
 * [# Edit, Date #]
 */

/*
 * Delete - program to delete files
 * [# Edit 23, Date 8-Aug-89 #]
 */

When updating a module, it is never necessary to edit these sequences, as any old values will be removed and replaced with the new ones. Also, when using the -d or -du options (described below), lines with these sequences compare as if the values were null, and thus will not cause spurious differences.

During an update, the special character sequences are read and any edit value found is compared against the current edit number of the module. If they differ, then the update fails. This provides an interlock check for the case of two users extracting the same version of a file, editing it, and then both updating it without knowledge of each other. In this case, the second user would fail, and then he can merge his edits with the previous user's edit and then retry the update. This checking is disabled if there is no special character sequence containing the edit keyword, the edit number value is null, or if the -Forced_Update option is used to indicate that the check is not needed.

-No_Keywords

This option may be used to disable the use of the keyword special character sequences described above. Text containing keyword sequences is treated as plain text. Note: the -No_Keywords option does not imply the -BINary option.

-Name string

This option is used to associate a name for the newest version of a module. It can be given along with the -CReate, -Update, or -Difference_Update options, to specify a name for the new version of the module. It can also be given by itself in order to specify a name for the newest version of a module. Each edit number can have many names associated with it, so this will not remove any previously defined name for the edit. This option is useful to correlate many modules together. For example, when a new version of a program is ready to be released, you could give each module of the program the same name release1. Then in the future, you can recreate the sources making up that release by extracting the edits with the name release1 for every module. Edit names cannot begin with a digit, and cannot contain plus or minus signs. These rules prevent ambiguous parsing of edit numbers for the -Extract, -Terminal, -ALL, and -List options.

-List [ edit1 [ edit2 ]]

This option prints a list of edits for the module, giving the user name, date, user remarks, and names specified for the edits. If no edit number is supplied, then all edits are printed in reverse order. If a single edit number is supplied, then only that edit number is printed. If two edit numbers are supplied, then all edits in the specified range are printed. The output from this option defaults to the terminal. You can use the -Output option to save the results to a file.

-Difference [ edit1 [ edit2 ]]

This option is used to display the differences between two versions of a module, or a file and a version of a module. There are three modes for this action, depending on how many edit numbers are supplied. These modes are illustrated by the following examples:

foo.c -d

Compare latest version against file "foo.c"

foo.c -d 3

Compare version 3 against file "foo.c"

foo.c -d 3 4

Compare version 3 against version 4

This option accepts the -Input option to specify the file to be compared. When using the -Difference option, the output defaults to the terminal. Therefore, you must use -Output if you wish the differences saved to a file. Using -Quick with -Difference will only output a quick summary of the changes, instead of the detailed changes. This summary only supplies the number of lines inserted, deleted, and unchanged between the files. Using -What with -Difference will display all of both files, showing in detail what the differences are using change bars.

The -Difference option may need to write one or two temporary files in order to extract old versions of a module to be compared. These files have names like T$n_nnn . They are deleted again just before differences are output, so that stopping the output before it is complete will not leave these files around. The temporary files are usually written to the current directory. If this is not reasonable because of permission or quota problems, then you can specify the directory for writing the temporary files into. This is done by defining the TMPDIR environment variable to be the path of the directory.

-Difference_Update

This option combines the effects of the -Difference and -Update options. It displays the differences between a file and the latest version of a module. If there are any differences, it then proceeds to perform an update of the module with that file, asking for remarks as usual. This option is very useful when used with wildcarded module names. Then you can update just those modules which were changed by an edit session, and see the changes for each module before typing the appropriate remark for each module.

You may specify both of the -Difference and -Update options, or you may use this option. The results are identical.

-Conditional_Update

This option conditionally updates a module. That is, it will only do an update if there are any differences between a file and the latest version of a module. This is convenient when related changes are made to many modules in a directory, and one command using wildcards can update just those modules that were changed.

The -CReate option may be combined with the -Update or -Conditional_Update options to create the file if required.

-CLean

This option is used to remove files which match the newest versions of modules. If a file exists which matches the newest version of a module, then the file is deleted, otherwise it is kept. This option is used to clean up a work directory after building a new version of a product. This option is especially useful when used with the -ALL option. It will also accept the -Input option to specify a directory containing the files to be cleaned.

-CHeck

This option is used to find out if a file does not match the latest version of a module. If so, a message is given. If the file does match, no output occurs. This option is thus useful to determine which files have been modified and in need of updating. The -ALL option is defaulted for this option, since it is usually used for all modules. For example,

 -CHeck

will report on all files which are different than the latest modules. If -Quick is specified, then the output will consist of the module names with no other output. This is useful for the backquote operator in shell scripts for referencing the modules which are out of date. The -CHeck option will also accept the -Input option.

-PRune edit

This option is used to permanently remove early edits from an edit history. This is useful if you wish to cut down on the amount of disk space taken by an edit history file, or when you want to start another release of a file, and want a copy of the edit history file for that new release. The option takes an edit number to preserve, and all edits in the edit history file before that edit are deleted, and can no longer be referenced. For example, to keep only the current edit plus the previous 10 edits of the module file, you could use the command:

 file -prune -10

Since the -PRune option is unrecoverable (unless backup files are available), the program asks the user to verify that the prune is really wanted. The -Forced_Update option can be used to bypass this verification.

-ALL

This option can be used with any of the action options. It means perform the operation for all modules in the module storage directory. Alternatively, you can specify multiple module names on the command line, and the actions will be performed with those modules. You cannot specify both -ALL and module names.

When using multiple modules or the -ALL option, the -Input and -Output options have a slightly different meaning. In these cases, the -Input and -Output arguments are a directory name which contains filenames with the same name as the module names. If the argument is not a directory, then an error is given. This feature is useful for example, to extract all the modules and place them into some remote directory, as in:

 -all -e -o tempdir

You should be careful when specifying numeric edit numbers for multiple modules. Most probably, a particular edit number is not appropriate for multiple modules, since changes corresponding to a particular edit number are not usually related. Using named edits avoids these problems. As an example, if you wanted to extract every module which had an edit that was named rev3, then you could use the command:

 -all -e rev3

Some other useful examples of commands which use multiple modules are:

 *.c -create
 -check -all
 -cu -all

-Verbose

This option can be specified with any other action, and outputs status information about the progress of the action. This is useful for debugging of problems, or just for amusement when the system is slow or a large file is being processed. It accepts a numeric argument to indicate the verbosity for output. The levels are as follows:

0

No output at all (except for errors).

1

Single-line output describing action (default).

2

Detailed status as action proceeds.

-Help

Give some help on how to use the program.

-VERSion

Show what version of is running.

All options may be abbreviated; the abbreviation is documented as the upper case letters, all lower case letters and underscores (_) are optional. You must use consecutive sequences of optional letters.

All options are case insensitive, you may type them in upper case or lower case or a combination of both, case is not important.

For example: the arguments "-help, "-HELP" and "-h" are all interpreted to mean the -Help option. The argument "-hlp" will not be understood, because consecutive optional characters were not supplied.

Options and other command line arguments may be mixed arbitrarily on the command line.

The GNU long option names are understood. Since all option names for are long, this means ignoring the extra leading '-'. The "--option=value" convention is also understood.

As a convenience, if a pathname begins with a period and a environment variable exists with that name, then the value of the environment variable will be used as the actual pathname. For example, if a environment variable of .FOO has the value this.is.a.long.name, then the command

 -o .FOO

 -o this.is.a.long.name

 -o ./.FOO

In general, can handle all text files you throw at it, even international text with unusual encodings. However, is unable to cope elegantly with files which contain the NUL character.

The fcomp(1) program simply prints a warning, and continues, you need to know that it converts NUL characters into an 0x80 value before performing the comparison.

The fmerge(1) program also converts the NUL character to an 0x80 value before merging, after a warning, and any output file will contain this value, rather than the original NUL character.

The fhist(1) program, however, generates a fatal error if any input file contains NUL characters. This is intended to protect your source files for unintentional corruption. Use --BINary for files which absolutely must contain NUL characters.

The program will exit with a status of 1 on any error. The program will only exit with a status of 0 if there are no errors.

This program is based on the algorithm in An O(N) Difference Algorithm and Its Variations, Eugene W. Myers, TR 85-6, 10-April-1985, Department of Computer Science, The University of Arizona, Tuscon, Arizona 85721. See also: A File Comparison Program, Webb Miller and Eugene W. Myers, Software Practice and Experience, Volume 15, No. 11, November 1985.

version

Copyright Peter Miller; All rights reserved.

This program is derived from a work

Copyright 1990 David I. Bell.

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.

tab(;); l r l. Peter Miller;Web:;http://www.canb.auug.org.au/~millerp /\/\*;E-Mail:;millerp@canb.auug.org.au David I. Bell;Web:;http://www.canb.auug.org.au/~dbell ;E-Mail:;dbell@canb.auug.org.au