dspinellis/unix-history-make: Code and data to create a git repository representing the Unix source code history

Unix History Repository

The history and evolution of the Unix operating system is made available as a revision management repository, covering the period from its inception in 1970 as a 2.5 thousand line kernel and 48 commands, to 2025 as a widely-used 41 million line system. The 2 GB repository contains about 850 thousand commits and more than eight thousand merges. The repository employs Git system for its storage and is hosted on GitHub. It has been created by synthesizing with custom software 22 snapshots of systems developed at Bell Labs, the University of California at Berkeley, and the 386BSD team, and the modern repository of the open source FreeBSD system. In total, about 2700 individual contributors are identified, the early ones through primary research. The data set c…

Unix History Repository

You can read more details about the contents, creation, and uses of this repository through this link.

Two repositories are associated with the project:

unix-history-repo is a repository representing a reconstructed version of the Unix history, based on the currently available data. This repository will be often automatically regenerated from scratch, so this is not a place to make contributions. To ensure replicability its users are encouraged to fork it or archive it.
unix-history-make is a repository containing code and metadata used to build the above repository. Contributions to this repository are welcomed.

Project status

The project has achieved its major goal with the establishment of a continuous timeline from 1970 until today. The repository contains:

snapshots of PDP-7, V1, V2, V3, V4, V5, V6, and V7 Research Edition,
Unix/32V,
all available BSD releases,
the CSRG SCCS history,
two releases of 386BSD,
the 386BSD patchkit,
the current FreeBSD repository, which now contains its full history.

The files appear to be added in the repository in chronological order according to their modification time, and large parts of the source code have been attributed to their actual authors. Commands like git blame and git log produce the expected results.

The repository contains a number of two-way merges.

3 BSD is merged from Unix/32V and Research Edition 6
Various BSD releases are merged from the development branch and a time point of BSD-SCCS

Blame is apportioned appropriately.

Tags and Branches

The following tags or branch names mark specific releases, listed in rough chronological order.

Epoch
Research-PDP7
Research-V1–6
BSD-1
BSD-2
Research-V7
Bell-32V
BSD-3, 4, 4_1_snap, 4_1c_2, 4_2, 4_3, 4_3_Reno, 4_3_Net_1, 4_3_Tahoe, 4_3_Net_2, 4_4, 4_4_Lite1, 4_4_Lite2 SCCS-END,
386BSD-0.0, 0.1, patchkit
All release and releng branches found in the FreeBSD Git repository

A detailed description of the major tags is available in the file releases.md.

More tags and branches are available.

The -Snapshot-Development branches denote commits that have been synthesized from a time-ordered sequence of a snapshot’s files.
The -VCS-Development tags denote the point along an imported version control history branch where a particular release occurred.

Cool things you can do

If you have a broadband network connection and about 2 GB of free disk space, you can download the repository and run Git commands that go back decades. Run

git clone https://github.com/dspinellis/unix-history-repo
git checkout BSD-Release

to get a local copy of the Unix history repository.

View log across releases

Running

git log --reverse --date-order

will give you commits such as the following

commit 64d7600ea5210a9125bd1a06e5d184ef7547d23d
Author: Ken Thompson <ken@research.uucp>
Date:   Tue Jun 20 05:00:00 1972 -0500

Research V1 development
Work on file u5.s

Co-Authored-By: Dennis Ritchie <dmr@research.uucp>
Synthesized-from: v1/sys
[...]
commit 4030f8318890a026e065bc8926cebefb71e9d353
Author: Ken Thompson <ken@research.uucp>
Date:   Thu Aug 30 19:30:25 1973 -0500

Research V3 development
Work on file sys/ken/slp.c

Synthesized-from: v3
[...]
commit c4999ec655319a01e84d9460d84df824006f9e2d
Author: Dennis Ritchie <dmr@research.uucp>
Date:   Thu Aug 30 19:33:01 1973 -0500

Research V3 development
Work on file sys/dmr/kl.c

Synthesized-from: v3
[...]
commit 355c543c6840fa5f37d8daf2e2eaa735ea6daa4a
Author: Brian W. Kernighan <bwk@research.uucp>
Date:   Tue May 13 19:43:47 1975 -0500

Research V6 development
Work on file usr/source/rat/r.g

Synthesized-from: v6
[...]
commit 0ce027f7fb2cf19b7e92d74d3f09eb02e8fea50e
Author: S. R. Bourne <srb@research.uucp>
Date:   Fri Jan 12 02:17:45 1979 -0500

Research V7 development
Work on file usr/src/cmd/sh/blok.c

Synthesized-from: v7
[...]
Author: Eric Schmidt <schmidt@ucbvax.Berkeley.EDU>
Date:   Sat Jan 5 22:49:18 1980 -0800

BSD 3 development

Work on file usr/src/cmd/net/sub.c

View changes to a specific file

Run

git checkout Research-Release
git log --follow --simplify-merges usr/src/cmd/c/c00.c

to see dates on which the C compiler was modified.

Annotate lines in a specific file by their version

Run

git blame -C -C usr/sys/sys/pipe.c

to see how the Unix pipe functionality evolved over the years.

3cc1108b usr/sys/ken/pipe.c     (Ken Thompson 1974-11-26 18:13:21 -0500  53) 	rf->f_flag = FREAD|FPIPE;
3cc1108b usr/sys/ken/pipe.c     (Ken Thompson 1974-11-26 18:13:21 -0500  54) 	rf->f_inode = ip;
3cc1108b usr/sys/ken/pipe.c     (Ken Thompson 1974-11-26 18:13:21 -0500  55) 	ip->i_count = 2;
[...]
1f183be2 usr/sys/sys/pipe.c     (Ken Thompson 1979-01-10 15:19:35 -0500 122) 	register struct inode *ip;
1f183be2 usr/sys/sys/pipe.c     (Ken Thompson 1979-01-10 15:19:35 -0500 123)
1f183be2 usr/sys/sys/pipe.c     (Ken Thompson 1979-01-10 15:19:35 -0500 124) 	ip = fp->f_inode;
1f183be2 usr/sys/sys/pipe.c     (Ken Thompson 1979-01-10 15:19:35 -0500 125) 	c = u.u_count;
1f183be2 usr/sys/sys/pipe.c     (Ken Thompson 1979-01-10 15:19:35 -0500 126)
1f183be2 usr/sys/sys/pipe.c     (Ken Thompson 1979-01-10 15:19:35 -0500 127) loop:
1f183be2 usr/sys/sys/pipe.c     (Ken Thompson 1979-01-10 15:19:35 -0500 128)
1f183be2 usr/sys/sys/pipe.c     (Ken Thompson 1979-01-10 15:19:35 -0500 129) 	/*
9a9f6b22 usr/src/sys/sys/pipe.c (Bill Joy     1980-01-05 05:51:18 -0800 130) 	 * If error or all done, return.
9a9f6b22 usr/src/sys/sys/pipe.c (Bill Joy     1980-01-05 05:51:18 -0800 131) 	 */
9a9f6b22 usr/src/sys/sys/pipe.c (Bill Joy     1980-01-05 05:51:18 -0800 132)
9a9f6b22 usr/src/sys/sys/pipe.c (Bill Joy     1980-01-05 05:51:18 -0800 133) 	if (u.u_error)
9a9f6b22 usr/src/sys/sys/pipe.c (Bill Joy     1980-01-05 05:51:18 -0800 134) 		return;
6d632e85 usr/sys/ken/pipe.c     (Ken Thompson 1975-07-17 10:33:37 -0500 135) 	plock(ip);
6d632e85 usr/sys/ken/pipe.c     (Ken Thompson 1975-07-17 10:33:37 -0500 136) 	if(c == 0) {
6d632e85 usr/sys/ken/pipe.c     (Ken Thompson 1975-07-17 10:33:37 -0500 137) 		prele(ip);
6d632e85 usr/sys/ken/pipe.c     (Ken Thompson 1975-07-17 10:33:37 -0500 138) 		u.u_count = 0;
6d632e85 usr/sys/ken/pipe.c     (Ken Thompson 1975-07-17 10:33:37 -0500 139) 		return;
6d632e85 usr/sys/ken/pipe.c     (Ken Thompson 1975-07-17 10:33:37 -0500 140) 	}

Marvel at the oldest code in a modern Unix system

Run

git blame -C -C -M -M FreeBSD-releng/15.0 -- lib/libc/gen/timezone.c | grep Ritchie

to see code written by Dennis Ritchie in 1979 still part of the 2025 FreeBSD 15.0 C library.

dbf40ea3ff692 usr/src/libc/gen/timezone.c                     (Dennis Ritchie    1979-01-10 14:58:57 -0500  73) static struct zone {
dbf40ea3ff692 usr/src/libc/gen/timezone.c                     (Dennis Ritchie    1979-01-10 14:58:57 -0500  74)       int     offset;
dbf40ea3ff692 usr/src/libc/gen/timezone.c                     (Dennis Ritchie    1979-01-10 14:58:57 -0500  75)       char    *stdzone;
dbf40ea3ff692 usr/src/libc/gen/timezone.c                     (Dennis Ritchie    1979-01-10 14:58:57 -0500  76)       char    *dlzone;
dbf40ea3ff692 usr/src/libc/gen/timezone.c                     (Dennis Ritchie    1979-01-10 14:58:57 -0500  77) } zonetab[] = {
dbf40ea3ff692 usr/src/libc/gen/timezone.c                     (Dennis Ritchie    1979-01-10 14:58:57 -0500 111)                       if (dst && zp->dlzone)
dbf40ea3ff692 usr/src/libc/gen/timezone.c                     (Dennis Ritchie    1979-01-10 14:58:57 -0500 112)                               return(zp->dlzone);
dbf40ea3ff692 usr/src/libc/gen/timezone.c                     (Dennis Ritchie    1979-01-10 14:58:57 -0500 113)                       if (!dst && zp->stdzone)
dbf40ea3ff692 usr/src/libc/gen/timezone.c                     (Dennis Ritchie    1979-01-10 14:58:57 -0500 114)                               return(zp->stdzone);
dbf40ea3ff692 usr/src/libc/gen/timezone.c                     (Dennis Ritchie    1979-01-10 14:58:57 -0500 115)

How you can help

You can help if you were there at the time, or if you can locate a source that contains information that is currently missing.

If your current GitHub account is not linked to your past contributions, (you can search them through this page), associate your past email with your current account through your GitHub account settings. (Contact me for instructions on how to add email addresses to which you no longer have access.)
Look for errors and omissions in the files that map file paths to authors.
Look for parts of the system that have not yet been attributed in these files and propose suitable attributions. Keep in mind that attributions for parts that were developed in one place and modified elsewhere (e.g. developed at Bell Labs and modified at Berkeley) should be for the person who did the modification, not the original author.
Look for authors whose identifier starts with x- in the author id to name map files for Bell Labs, and Berkeley, and provide or confirm their actual login identifier. (The one used is a guess.)
Contribute a path regular expression to contributor map file (see v7.map) for 4.2BSD, 4.3BSD, 4.3BSD-Reno, 4.3BSD-Tahoe, 4.3BSD-Alpha, and Net2.
Import further branches, such as 2BSD, NetBSD, OpenBSD, and Plan 9 from Bell Labs.

Re-creating the historical repository from scratch

The -make repository is provided to share and document the creation process, rather than as a bullet-proof way to get consistent and repeatable results. For instance, importing the snapshots on a system that is case-insensitive (NTFS, HFS Plus with default settings) will result in a few files getting lost.

Prerequisites

Git
Perl
The Perl modules VCS::SCCS and Git::FastExport (Install with sudo GIT_ALLOW_PROTOCOL=file cpan VCS::SCCS Git::FastExport.)
If compiling patch under GNU/Linux the library libbsd (e.g. the libbsd-dev package).
Sudo (and authorization to use it to mount ISO images)

Repository creation

The -repo repository can be created with the following commands.

make
./import.sh

Adding a single source

If you want to add a new source without running the full import process, you can do the following.

Prepare the source’s maps and data
cd to the repo directory
Checkout the repo at the point where the new source will branch out
Run a Perl command such as the following.

perl ../import-dir.pl [-v] -m Research-V7 -c ../author-path/Bell-32V \
-n ../bell.au -r Research-V7 -i ../ignore/Bell-32V \
$ARCHIVE/32v Bell 32V -0500 | gfi

Acknowledgements

The following people helped with Bell Labs login identifiers.
Brian W. Kernighan
Doug McIlroy
Arnold D. Robbins
The following people helped with *BSD login identifiers.
Clem Cole
Era Eriksson
Mary Ann Horton
Warner Losh
Kirk McKusick
Jeremy C. Reed
Ingo Schwarze
Anatole Shaw
The BSD SCCS import code is based on work by
H. Merijn Brand
Jonathan Gray

Unix History Repository

Unix History Repository

Project status

Tags and Branches

Cool things you can do

View log across releases

View changes to a specific file

Annotate lines in a specific file by their version

Marvel at the oldest code in a modern Unix system

How you can help

Re-creating the historical repository from scratch

Prerequisites

Repository creation

Adding a single source

Further reading

Acknowledgements

Similar Posts