|
ae0628a…
|
drh
|
1 |
<title>Fossil Versus Git</title> |
|
ae0628a…
|
drh
|
2 |
|
|
ae0628a…
|
drh
|
3 |
<h2>1.0 Don't Stress!</h2> |
|
ae0628a…
|
drh
|
4 |
|
|
3609ff7…
|
wyoung
|
5 |
The feature sets of Fossil and [http://git-scm.com | Git] overlap in |
|
3609ff7…
|
wyoung
|
6 |
many ways. Both are |
|
3609ff7…
|
wyoung
|
7 |
[https://en.wikipedia.org/wiki/Distributed_version_control | distributed |
|
e299892…
|
wyoung
|
8 |
version control systems] which store a tree of check-in objects to a |
|
e299892…
|
wyoung
|
9 |
local repository clone. In both systems, the local clone starts out as a |
|
e299892…
|
wyoung
|
10 |
full copy of the remote parent. New content gets added to the local |
|
e299892…
|
wyoung
|
11 |
clone and then later optionally pushed up to the remote, and changes to |
|
e299892…
|
wyoung
|
12 |
the remote can be pulled down to the local clone at will. Both systems |
|
0996347…
|
wyoung
|
13 |
offer diffing, patching, branching, merging, cherry-picking, bisecting, |
|
3609ff7…
|
wyoung
|
14 |
private branches, a stash, etc. |
|
a5284fa…
|
drh
|
15 |
|
|
16cb9c0…
|
wyoung
|
16 |
Fossil has inbound and outbound Git conversion features, so if you start |
|
16cb9c0…
|
wyoung
|
17 |
out using one DVCS and later decide you like the other better, you can |
|
431245e…
|
wyoung
|
18 |
easily [./inout.wiki | move your version-controlled file content].¹ |
|
16cb9c0…
|
wyoung
|
19 |
|
|
e299892…
|
wyoung
|
20 |
In this document, we set all of that similarity and interoperability |
|
e299892…
|
wyoung
|
21 |
aside and focus on the important differences between the two, especially |
|
e299892…
|
wyoung
|
22 |
those that impact the user experience. |
|
a5284fa…
|
drh
|
23 |
|
|
a5284fa…
|
drh
|
24 |
Keep in mind that you are reading this on a Fossil website, and though |
|
35fe8bf…
|
wyoung
|
25 |
we try to be fair, the information here will inevitably |
|
35fe8bf…
|
wyoung
|
26 |
be biased in favor of Fossil purely because we spend most of our |
|
16cb9c0…
|
wyoung
|
27 |
time using Fossil, not Git. Ask around for second opinions from |
|
a5284fa…
|
drh
|
28 |
people who have used <em>both</em> Fossil and Git. |
|
7e907a1…
|
wyoung
|
29 |
|
|
7e907a1…
|
wyoung
|
30 |
If you want a more practical, less philosophical guide to moving from |
|
7e907a1…
|
wyoung
|
31 |
Git to Fossil, see our [./gitusers.md | Git to Fossil Translation Guide]. |
|
8452627…
|
wyoung
|
32 |
|
|
6ae97be…
|
drh
|
33 |
|
|
6ae97be…
|
drh
|
34 |
<h2>2.0 Differences Between Fossil And Git</h2> |
|
6ae97be…
|
drh
|
35 |
|
|
6ae97be…
|
drh
|
36 |
Differences between Fossil and Git are summarized by the following table, |
|
6ae97be…
|
drh
|
37 |
with further description in the text that follows. |
|
6ae97be…
|
drh
|
38 |
|
|
8a1ba49…
|
wyoung
|
39 |
<table style="width: fit-content"> |
|
8a1ba49…
|
wyoung
|
40 |
<tr><th>GIT</th><th>FOSSIL</th><th>more</th></tr> |
|
fd9e211…
|
wyoung
|
41 |
<tr> |
|
fd9e211…
|
wyoung
|
42 |
<td>File versioning only</td> |
|
8a1ba49…
|
wyoung
|
43 |
<td> |
|
8a1ba49…
|
wyoung
|
44 |
VCS, tickets, wiki, docs, notes, forum, chat, UI, |
|
8a1ba49…
|
wyoung
|
45 |
[https://en.wikipedia.org/wiki/Role-based_access_control|RBAC] |
|
8a1ba49…
|
wyoung
|
46 |
</td> |
|
fd9e211…
|
wyoung
|
47 |
<td><a href="#features">2.1 ↓</a></td> |
|
fd9e211…
|
wyoung
|
48 |
</tr> |
|
fd9e211…
|
wyoung
|
49 |
<tr> |
|
358d7d8…
|
drh
|
50 |
<td>A federation of many small programs</td> |
|
358d7d8…
|
drh
|
51 |
<td>One self-contained, stand-alone executable</td> |
|
41b3491…
|
drh
|
52 |
<td><a href="#selfcontained">2.2 ↓</a></td> |
|
fd9e211…
|
wyoung
|
53 |
</tr> |
|
fd9e211…
|
wyoung
|
54 |
<tr> |
|
358d7d8…
|
drh
|
55 |
<td>Custom key/value data store</td> |
|
358d7d8…
|
drh
|
56 |
<td>[https://sqlite.org/mostdeployed.html|The most used SQL database in the world]</td> |
|
fd9e211…
|
wyoung
|
57 |
<td><a href="#durable">2.3 ↓</a></td> |
|
fd9e211…
|
wyoung
|
58 |
</tr> |
|
fd9e211…
|
wyoung
|
59 |
<tr> |
|
358d7d8…
|
drh
|
60 |
<td>Runs natively on POSIX systems</td> |
|
358d7d8…
|
drh
|
61 |
<td>Runs natively on both POSIX and Windows</td> |
|
fd9e211…
|
wyoung
|
62 |
<td><a href="#portable">2.4 ↓</a></td> |
|
fd9e211…
|
wyoung
|
63 |
</tr> |
|
fd9e211…
|
wyoung
|
64 |
<tr> |
|
fd9e211…
|
wyoung
|
65 |
<td>Bazaar-style development</td> |
|
fd9e211…
|
wyoung
|
66 |
<td>Cathedral-style development</td> |
|
fd9e211…
|
wyoung
|
67 |
<td><a href="#devorg">2.5.1 ↓</a></td> |
|
fd9e211…
|
wyoung
|
68 |
</tr> |
|
fd9e211…
|
wyoung
|
69 |
<tr> |
|
fd9e211…
|
wyoung
|
70 |
<td>Designed for Linux kernel development</td> |
|
fd9e211…
|
wyoung
|
71 |
<td>Designed for SQLite development</td> |
|
fd9e211…
|
wyoung
|
72 |
<td><a href="#scale">2.5.2 ↓</a></td> |
|
fd9e211…
|
wyoung
|
73 |
</tr> |
|
fd9e211…
|
wyoung
|
74 |
<tr> |
|
fd9e211…
|
wyoung
|
75 |
<td>Focus on individual branches</td> |
|
fd9e211…
|
wyoung
|
76 |
<td>Focus on the entire tree of changes</td> |
|
b08f062…
|
drh
|
77 |
<td><a href="#branches">2.5.3 ↓</a></td> |
|
fd9e211…
|
wyoung
|
78 |
</tr> |
|
fd9e211…
|
wyoung
|
79 |
<tr> |
|
fd9e211…
|
wyoung
|
80 |
<td>One check-out per repository</td> |
|
fd9e211…
|
wyoung
|
81 |
<td>Many check-outs per repository</td> |
|
fd9e211…
|
wyoung
|
82 |
<td><a href="#checkouts">2.6 ↓</a></td> |
|
fd9e211…
|
wyoung
|
83 |
</tr> |
|
fd9e211…
|
wyoung
|
84 |
<tr> |
|
fd9e211…
|
wyoung
|
85 |
<td>Remembers what you should have done</td> |
|
fd9e211…
|
wyoung
|
86 |
<td>Remembers what you actually did</td> |
|
fd9e211…
|
wyoung
|
87 |
<td><a href="#history">2.7 ↓</a></td> |
|
fd9e211…
|
wyoung
|
88 |
</tr> |
|
fd9e211…
|
wyoung
|
89 |
<tr> |
|
fd9e211…
|
wyoung
|
90 |
<td>Commit first</td> |
|
fd9e211…
|
wyoung
|
91 |
<td>Test first</td> |
|
fd9e211…
|
wyoung
|
92 |
<td><a href="#testing">2.8 ↓</a></td> |
|
fd9e211…
|
wyoung
|
93 |
</tr> |
|
fd9e211…
|
wyoung
|
94 |
<tr> |
|
358d7d8…
|
drh
|
95 |
<td>SHA-1 or SHA-2</td> |
|
358d7d8…
|
drh
|
96 |
<td>SHA-1 and/or SHA-3, in the same repository</td> |
|
fd9e211…
|
wyoung
|
97 |
<td><a href="#hash">2.9 ↓</a></td> |
|
fd9e211…
|
wyoung
|
98 |
</tr> |
|
8a1ba49…
|
wyoung
|
99 |
</table> |
|
5c24bee…
|
wyoung
|
100 |
|
|
5c24bee…
|
wyoung
|
101 |
<h3 id="features">2.1 Featureful</h3> |
|
5566c85…
|
drh
|
102 |
|
|
5566c85…
|
drh
|
103 |
Git provides file versioning services only, whereas Fossil adds |
|
bd7c47c…
|
wyoung
|
104 |
an integrated [./wikitheory.wiki | wiki], |
|
ae0628a…
|
drh
|
105 |
[./bugtheory.wiki | ticketing & bug tracking], |
|
fd1282e…
|
drh
|
106 |
[./embeddeddoc.wiki | embedded documentation], |
|
f5ab488…
|
drh
|
107 |
[./event.wiki | technical notes], a [./forum.wiki | web forum], |
|
f5ab488…
|
drh
|
108 |
and a [./chat.md | chat service], |
|
c5a74af…
|
wyoung
|
109 |
all within a single nicely-designed [./customskin.md|skinnable] web |
|
c64f28d…
|
drh
|
110 |
[/help/ui|UI], |
|
779ddef…
|
wyoung
|
111 |
protected by [./caps/ | a fine-grained role-based |
|
bd7c47c…
|
wyoung
|
112 |
access control system]. |
|
fd99012…
|
wyoung
|
113 |
These additional capabilities are available for Git as 3rd-party |
|
fd99012…
|
wyoung
|
114 |
add-ons, but with Fossil they are integrated into |
|
c7afd68…
|
wyoung
|
115 |
the design, to the point that it approximates |
|
fd99012…
|
wyoung
|
116 |
"[https://github.com/ | GitHub]-in-a-box." |
|
fd99012…
|
wyoung
|
117 |
|
|
c7afd68…
|
wyoung
|
118 |
Even if you only want straight version control, Fossil has affordances |
|
c7afd68…
|
wyoung
|
119 |
not available in Git. |
|
c7afd68…
|
wyoung
|
120 |
|
|
c7afd68…
|
wyoung
|
121 |
For instance, Fossil can do operations over all local repo clones and |
|
c7afd68…
|
wyoung
|
122 |
check-out directories with a single command. You can say "<tt>fossil |
|
c7afd68…
|
wyoung
|
123 |
all sync</tt>" on a laptop prior to taking it off the network hosting |
|
c7afd68…
|
wyoung
|
124 |
those repos, as before going on a trip. It doesn't matter if those |
|
c7afd68…
|
wyoung
|
125 |
repos are private and restricted to your company network or public |
|
c7afd68…
|
wyoung
|
126 |
Internet-hosted repos, you get synced up with everything you need while |
|
c7afd68…
|
wyoung
|
127 |
off-network. |
|
c7afd68…
|
wyoung
|
128 |
|
|
c7afd68…
|
wyoung
|
129 |
You get the same capability with several other Fossil |
|
c7afd68…
|
wyoung
|
130 |
sub-commands as well, such as "<tt>fossil all changes</tt>" to get a list of files |
|
d357f81…
|
wyoung
|
131 |
that you forgot to commit prior to the end of your working day, across |
|
d357f81…
|
wyoung
|
132 |
all repos. |
|
d357f81…
|
wyoung
|
133 |
|
|
d357f81…
|
wyoung
|
134 |
Whenever Fossil is told to modify the local checkout in some destructive |
|
c64f28d…
|
drh
|
135 |
way ([/help/rm|fossil rm], [/help/update|fossil update], |
|
c64f28d…
|
drh
|
136 |
[/help/revert|fossil revert], etc.) Fossil remembers the prior state |
|
d357f81…
|
wyoung
|
137 |
and is able to return the check-out directory to that state with a |
|
c7afd68…
|
wyoung
|
138 |
<tt>fossil undo</tt> command. While you cannot undo a commit in Fossil |
|
c7afd68…
|
wyoung
|
139 |
— [#history | on purpose!] — as long as the change remains confined to |
|
d357f81…
|
wyoung
|
140 |
the local check-out directory only, Fossil makes undo |
|
d357f81…
|
wyoung
|
141 |
[https://git-scm.com/book/en/v2/Git-Basics-Undoing-Things|easier than in |
|
d357f81…
|
wyoung
|
142 |
Git]. |
|
d357f81…
|
wyoung
|
143 |
|
|
c7afd68…
|
wyoung
|
144 |
For developers who choose to self-host projects rather than rely on a |
|
c7afd68…
|
wyoung
|
145 |
3rd-party service such as GitHub, Fossil is much easier to set up: |
|
f146e21…
|
drh
|
146 |
the stand-alone Fossil executable together with a [./server/any/cgi.md|2-line CGI script] |
|
ca891ab…
|
drh
|
147 |
suffice to instantiate a full-featured developer website. To accomplish |
|
fcfaae3…
|
jan.nijtmans
|
148 |
the same using Git requires locating, installing, configuring, integrating, |
|
ca891ab…
|
drh
|
149 |
and managing a wide assortment of separate tools. Standing up a developer |
|
ca891ab…
|
drh
|
150 |
website using Fossil can be done in minutes, whereas doing the same using |
|
ca891ab…
|
drh
|
151 |
Git requires hours or days. |
|
ca891ab…
|
drh
|
152 |
|
|
5c24bee…
|
wyoung
|
153 |
Fossil is small, complete, and self-contained. If you clone |
|
5c24bee…
|
wyoung
|
154 |
[https://github.com/git/git|Git's self-hosting repository], you get just |
|
5c24bee…
|
wyoung
|
155 |
Git's source code. If you clone Fossil's self-hosting repository, you |
|
5c24bee…
|
wyoung
|
156 |
get the entire Fossil website — source code, documentation, ticket |
|
5c24bee…
|
wyoung
|
157 |
history, and so forth.² That means you get a copy of this very article |
|
5c24bee…
|
wyoung
|
158 |
and all of its historical versions, plus the same for all of the other |
|
5c24bee…
|
wyoung
|
159 |
public content on this site. |
|
d23a845…
|
wyoung
|
160 |
|
|
d23a845…
|
wyoung
|
161 |
|
|
41b3491…
|
drh
|
162 |
<h3 id="selfcontained" name="selfcontained">2.2 Self Contained</h3> |
|
d23a845…
|
wyoung
|
163 |
|
|
5c24bee…
|
wyoung
|
164 |
Git is actually a collection of many small tools, each doing one small |
|
5c24bee…
|
wyoung
|
165 |
part of the job, which can be recombined (by experts) to perform |
|
5c24bee…
|
wyoung
|
166 |
powerful operations. Git has a lot of complexity and many dependencies, |
|
5c24bee…
|
wyoung
|
167 |
so that most people end up installing it via some kind of package |
|
00af3db…
|
wyoung
|
168 |
manager, simply because the creation of complicated binary packages is |
|
00af3db…
|
wyoung
|
169 |
best delegated to people skilled in their creation. Normal Git users are |
|
00af3db…
|
wyoung
|
170 |
not expected to build Git from source and install it themselves. |
|
00af3db…
|
wyoung
|
171 |
|
|
c7afd68…
|
wyoung
|
172 |
Fossil is a single self-contained stand-alone executable which |
|
c7afd68…
|
wyoung
|
173 |
depends only on common platform libraries in its default configuration. |
|
c7afd68…
|
wyoung
|
174 |
To install one of [https://fossil-scm.org/home/uv/download.html | our |
|
c7afd68…
|
wyoung
|
175 |
precompiled binaries], unpack the executable from the archive and put it |
|
c7afd68…
|
wyoung
|
176 |
somewhere in your <tt>PATH</tt>. To uninstall it, delete the executable. |
|
c7afd68…
|
wyoung
|
177 |
|
|
c7afd68…
|
wyoung
|
178 |
This policy is particularly useful when running Fossil inside a |
|
c7afd68…
|
wyoung
|
179 |
restrictive container, anything from [./chroot.md | classic chroot |
|
c7afd68…
|
wyoung
|
180 |
jails] to modern [https://en.wikipedia.org/wiki/OS-level_virtualization |
|
c7afd68…
|
wyoung
|
181 |
| OS-level virtualization mechanisms] such as |
|
da545c9…
|
wyoung
|
182 |
[https://en.wikipedia.org/wiki/Docker_(software) | Docker]. |
|
b95abfd…
|
wyoung
|
183 |
Our [./containers.md | stock container image] is under 11 MB when |
|
b95abfd…
|
wyoung
|
184 |
uncompressed and running because it contains nothing but a single |
|
283b1a4…
|
wyoung
|
185 |
statically-linked binary. |
|
283b1a4…
|
wyoung
|
186 |
|
|
b95abfd…
|
wyoung
|
187 |
If you build a dynamically linked binary instead, a Linux |
|
b95abfd…
|
wyoung
|
188 |
x86_64 build of Fossil drops to under 5 MB on-disk, stripped. |
|
b95abfd…
|
wyoung
|
189 |
It will depend only on widespread |
|
283b1a4…
|
wyoung
|
190 |
platform libraries with stable ABIs such as glibc, zlib, and openssl. |
|
283b1a4…
|
wyoung
|
191 |
|
|
b95abfd…
|
wyoung
|
192 |
Much the same is true on Windows, where our precompiled static binaries |
|
b95abfd…
|
wyoung
|
193 |
are distributed inside a ZIP archive containing "<tt>fossil.exe</tt>" |
|
b95abfd…
|
wyoung
|
194 |
and not a thing else, with a size on-par that of the Linux container |
|
b95abfd…
|
wyoung
|
195 |
build. There is no "<tt>setup.exe</tt>" to run; just copy it into your |
|
b95abfd…
|
wyoung
|
196 |
<tt>%PATH%</tt>. |
|
358d7d8…
|
drh
|
197 |
|
|
358d7d8…
|
drh
|
198 |
Fossil is easy to build from sources. Just run |
|
358d7d8…
|
drh
|
199 |
"<tt>./configure && make</tt>" on POSIX systems and |
|
358d7d8…
|
drh
|
200 |
"<tt>nmake /f Makefile.msc</tt>" on Windows. |
|
358d7d8…
|
drh
|
201 |
|
|
b95abfd…
|
wyoung
|
202 |
Git, by contrast, takes 25 MiB on Ubuntu 26.04 across 940 files. |
|
b95abfd…
|
wyoung
|
203 |
That doesn't count platform facilities like the POSIX shell and script |
|
b95abfd…
|
wyoung
|
204 |
interpreters its full feature set requires. A fairer comparison to |
|
b95abfd…
|
wyoung
|
205 |
Fossil's single static binary container is the Docker Hardened Image for |
|
b95abfd…
|
wyoung
|
206 |
[https://hub.docker.com/hardened-images/catalog/dhi/git |Git 2.x on |
|
b95abfd…
|
wyoung
|
207 |
Alpine], which presently weighs in at 36.3 megs unpacked and 607 files, |
|
b95abfd…
|
wyoung
|
208 |
if you count all the <tt>busybox</tt> symlinks which would otherwise be |
|
b95abfd…
|
wyoung
|
209 |
separate binaries on a more traditional Linux system. Worst of all are |
|
b95abfd…
|
wyoung
|
210 |
the "Git for Windows" packages where they end up needing to ship a large |
|
b95abfd…
|
wyoung
|
211 |
yet nerfed Linux-like userland to support Git's many loosely coupled |
|
b95abfd…
|
wyoung
|
212 |
pieces, approaching a hundred megs and thousands of files. |
|
5c24bee…
|
wyoung
|
213 |
|
|
5c24bee…
|
wyoung
|
214 |
Some say that Git more closely adheres to the Unix philosophy, |
|
5c24bee…
|
wyoung
|
215 |
summarized as "many small tools, loosely joined," but we have many |
|
5c24bee…
|
wyoung
|
216 |
examples of other successful Unix software that violates that principle |
|
5c24bee…
|
wyoung
|
217 |
to good effect, from Apache to Python to ZFS. We can infer from that |
|
5c24bee…
|
wyoung
|
218 |
that this is not an absolute principle of good software design. |
|
5c24bee…
|
wyoung
|
219 |
Sometimes "many features, tightly-coupled" works better. What actually |
|
5c24bee…
|
wyoung
|
220 |
matters is effectiveness and efficiency. We believe Fossil achieves |
|
5c24bee…
|
wyoung
|
221 |
this. |
|
5c24bee…
|
wyoung
|
222 |
|
|
9dcb3de…
|
wyoung
|
223 |
The above size comparisons aren't apples-to-apples anyway. We've |
|
9dcb3de…
|
wyoung
|
224 |
compared the size of Fossil with all of its [#features | many built-in |
|
b95abfd…
|
wyoung
|
225 |
features] to a relatively bare-bones Git installation. You must add a |
|
b95abfd…
|
wyoung
|
226 |
lot of third-party software to Git to give it a Fossil-equivalent |
|
b95abfd…
|
wyoung
|
227 |
feature set. |
|
b95abfd…
|
wyoung
|
228 |
|
|
b95abfd…
|
wyoung
|
229 |
Consider [https://about.gitlab.com/|GitLab], which wraps Git in enough |
|
b95abfd…
|
wyoung
|
230 |
features to bring it to a rough (very rough) parity with Fossil. While |
|
b95abfd…
|
wyoung
|
231 |
it is certainly ahead in some areas — automation comes to mind — it must |
|
b95abfd…
|
wyoung
|
232 |
be noted that GitLab's own forums are on Discourse, not self-hosted as |
|
b95abfd…
|
wyoung
|
233 |
with Fossil. You may then argue that Discourse is superior to the Fossil |
|
b95abfd…
|
wyoung
|
234 |
forum feature, but this merely brings us back to another of Fossil's |
|
b95abfd…
|
wyoung
|
235 |
advantages: when the posts are part of the repo, you can migrate your |
|
b95abfd…
|
wyoung
|
236 |
entire project to another host merely by standing that repo back up on |
|
b95abfd…
|
wyoung
|
237 |
other hardware. Choosing to assemble your project hosting from multiple |
|
b95abfd…
|
wyoung
|
238 |
pieces requires each to have their own backups, their own management |
|
b95abfd…
|
wyoung
|
239 |
processes, and their own migration strategies. |
|
b95abfd…
|
wyoung
|
240 |
|
|
b95abfd…
|
wyoung
|
241 |
Even then, these disparate services do not cooperate at the same level |
|
b95abfd…
|
wyoung
|
242 |
as in Fossil, where the ability to have a forum post linking to a wiki |
|
b95abfd…
|
wyoung
|
243 |
article linking to a trouble ticket linking to a commit falls out of the |
|
b95abfd…
|
wyoung
|
244 |
model nearly for free. These are all internal links, mind, potentially |
|
b95abfd…
|
wyoung
|
245 |
using nothing but repository artifact hashes, all backed by SQLite's |
|
b95abfd…
|
wyoung
|
246 |
referential integrity. With something like GitLab, its "internal |
|
b95abfd…
|
wyoung
|
247 |
references" are tacked on after the fact, not Git repo hashes at all, |
|
b95abfd…
|
wyoung
|
248 |
and when it comes to external services like Discourse, history tells us |
|
b95abfd…
|
wyoung
|
249 |
you're storing up tech debt which will come due at some future point in |
|
b95abfd…
|
wyoung
|
250 |
the form of piles of broken links when one piece or the other needs |
|
b95abfd…
|
wyoung
|
251 |
changing out. |
|
b95abfd…
|
wyoung
|
252 |
|
|
b95abfd…
|
wyoung
|
253 |
Furthermore, GitLab is far more resource hungry even in its |
|
b95abfd…
|
wyoung
|
254 |
[https://docs.gitlab.com/omnibus/settings/memory_constrained_envs/ | |
|
b95abfd…
|
wyoung
|
255 |
minimal configuration], hence more costly to run than the equivalent |
|
b95abfd…
|
wyoung
|
256 |
Fossil setup. [https://hub.docker.com/r/gitlab/gitlab-ce/ | The |
|
b95abfd…
|
wyoung
|
257 |
official GitLab Community Edition container] currently clocks in at |
|
b95abfd…
|
wyoung
|
258 |
3.44 GiB, independent of add-ons like Discourse. You can expect it to |
|
b95abfd…
|
wyoung
|
259 |
cost around 8× as much to host it on a cloud service. Even pared down |
|
b95abfd…
|
wyoung
|
260 |
to the minimum, there remain the consequences from the difference in |
|
b95abfd…
|
wyoung
|
261 |
basic technology choices: Ruby and PostgreSQL vs C and SQLite. |
|
41b3491…
|
drh
|
262 |
|
|
41b3491…
|
drh
|
263 |
The Fossil project itself is [./selfhost.wiki|hosted on a small and |
|
41b3491…
|
drh
|
264 |
inexpensive VPS]. A bare-bones $5/month VPS or a |
|
41b3491…
|
drh
|
265 |
spare Raspberry Pi is sufficient to run a full-up project |
|
f5ab488…
|
drh
|
266 |
site, complete with tickets, wiki, chat, and forum, in addition to |
|
41b3491…
|
drh
|
267 |
being a code repository. |
|
41b3491…
|
drh
|
268 |
|
|
358d7d8…
|
drh
|
269 |
<h3 id="durable" name="database">2.3 Query Language</h3> |
|
fd99012…
|
wyoung
|
270 |
|
|
fd99012…
|
wyoung
|
271 |
The baseline data structures for Fossil and Git are the same, modulo |
|
e299892…
|
wyoung
|
272 |
formatting details. Both systems manage a |
|
e299892…
|
wyoung
|
273 |
[https://en.wikipedia.org/wiki/Directed_acyclic_graph | directed acyclic |
|
e299892…
|
wyoung
|
274 |
graph] (DAG) of [https://en.wikipedia.org/wiki/Merkle_tree | Merkle |
|
1ddb400…
|
wyoung
|
275 |
tree] structured check-in objects. |
|
e299892…
|
wyoung
|
276 |
Check-ins are identified by a cryptographic hash of the check-in |
|
358d7d8…
|
drh
|
277 |
contents, and each check-in refers to its parent via the parent's hash. |
|
e299892…
|
wyoung
|
278 |
|
|
e299892…
|
wyoung
|
279 |
The difference is that Git stores its objects as individual files in the |
|
358d7d8…
|
drh
|
280 |
<tt>.git</tt> folder or compressed into bespoke key/value |
|
e299892…
|
wyoung
|
281 |
[https://git-scm.com/book/en/v2/Git-Internals-Packfiles|pack-files], |
|
e299892…
|
wyoung
|
282 |
whereas Fossil stores its objects in a [https://www.sqlite.org/|SQLite] |
|
358d7d8…
|
drh
|
283 |
database file which provides ACID transactions and a high-level query |
|
358d7d8…
|
drh
|
284 |
language. |
|
939136a…
|
wyoung
|
285 |
|
|
939136a…
|
wyoung
|
286 |
This difference is more than an implementation detail; it has important |
|
e299892…
|
wyoung
|
287 |
practical consequences. |
|
e299892…
|
wyoung
|
288 |
|
|
939136a…
|
wyoung
|
289 |
One notable example is that it is difficult to find the descendants |
|
358d7d8…
|
drh
|
290 |
of check-ins in Git. |
|
358d7d8…
|
drh
|
291 |
One can easily locate the ancestors of a particular Git check-in |
|
939136a…
|
wyoung
|
292 |
by following the pointers embedded in the check-in object, but |
|
939136a…
|
wyoung
|
293 |
going the other direction is difficult enough |
|
939136a…
|
wyoung
|
294 |
that neither native Git nor the big "forge" facilities |
|
939136a…
|
wyoung
|
295 |
like GitHub and GitLab provide this capability short of crawling the |
|
939136a…
|
wyoung
|
296 |
[https://www.git-scm.com/docs/git-log|commit log]. In Fossil, |
|
939136a…
|
wyoung
|
297 |
we can find descendants using a simple SQL query, which then allows |
|
939136a…
|
wyoung
|
298 |
us to see [/timeline?df=release&y=ci|all check-ins since the last release], |
|
939136a…
|
wyoung
|
299 |
as but one example. |
|
939136a…
|
wyoung
|
300 |
Git lets you see "what came before," but Fossil makes it just as |
|
939136a…
|
wyoung
|
301 |
easy to also see "what came after." |
|
fd99012…
|
wyoung
|
302 |
|
|
fd99012…
|
wyoung
|
303 |
Leaf check-ins in Git that lack a "ref" become "detached," making them |
|
cd989fe…
|
wyoung
|
304 |
difficult to locate and subject to garbage collection. This |
|
788db46…
|
stephan
|
305 |
[https://stackoverflow.com/q/3965676 | detached head |
|
358d7d8…
|
drh
|
306 |
state] problem has caused grief for |
|
939136a…
|
wyoung
|
307 |
[https://www.google.com/search?q=git+detached+head+state | |
|
939136a…
|
wyoung
|
308 |
untold millions of Git users]. With |
|
e299892…
|
wyoung
|
309 |
Fossil, detached heads are simply impossible because we can always find |
|
358d7d8…
|
drh
|
310 |
our way back into the Merkle tree using one or more of the relations |
|
358d7d8…
|
drh
|
311 |
in the SQL database. |
|
358d7d8…
|
drh
|
312 |
|
|
358d7d8…
|
drh
|
313 |
The SQL query capabilities of Fossil make it easier to track the |
|
358d7d8…
|
drh
|
314 |
changes for one particular file within a project. For example, |
|
358d7d8…
|
drh
|
315 |
you can easily find |
|
358d7d8…
|
drh
|
316 |
[/finfo/www/fossil-v-git.wiki|the complete edit history of this one document], |
|
358d7d8…
|
drh
|
317 |
or even |
|
358d7d8…
|
drh
|
318 |
[/finfo/www/fossil-v-git.wiki?ubg|the same history color-coded by committer], |
|
939136a…
|
wyoung
|
319 |
Both come down to simple SQL queries in Fossil, with procedural code |
|
358d7d8…
|
drh
|
320 |
only being used to format the result for display. |
|
358d7d8…
|
drh
|
321 |
The same result could be obtained from Git, but because the data is |
|
358d7d8…
|
drh
|
322 |
in a key/value store, much more procedural code has to be written to |
|
939136a…
|
wyoung
|
323 |
walk the data and compute the result. |
|
358d7d8…
|
drh
|
324 |
|
|
358d7d8…
|
drh
|
325 |
The ease of querying Fossil data using SQL means that status or |
|
358d7d8…
|
drh
|
326 |
history information about the project under management is easier |
|
939136a…
|
wyoung
|
327 |
to obtain, hence more likely to happen, giving its developers better |
|
939136a…
|
wyoung
|
328 |
situational awareness. |
|
358d7d8…
|
drh
|
329 |
Fossil reports tend to be more detailed and useful. |
|
e28c25e…
|
wyoung
|
330 |
Compare [/timeline?c=6df7a853ec16865b|this Fossil timeline] |
|
e28c25e…
|
wyoung
|
331 |
to |
|
9f135f2…
|
wyoung
|
332 |
[https://github.com/drhsqlite/fossil-mirror/commits/master?after=f720c106d297ca1f61bccb30c5c191b88a626d01+34 | |
|
939136a…
|
wyoung
|
333 |
its closest equivalent in the GitHub mirror]. Judge for yourself: which of those |
|
358d7d8…
|
drh
|
334 |
reports is more useful to a developer trying to understand what happened? |
|
64a33a6…
|
wyoung
|
335 |
|
|
64a33a6…
|
wyoung
|
336 |
<h3 id="portable">2.4 Portable</h3> |
|
64a33a6…
|
wyoung
|
337 |
|
|
64a33a6…
|
wyoung
|
338 |
Fossil is largely written in ISO C, almost purely conforming to the |
|
64a33a6…
|
wyoung
|
339 |
original 1989 standard. We make very little use of |
|
4b574be…
|
wyoung
|
340 |
[https://en.wikipedia.org/wiki/C99|C99], and we do not knowingly make |
|
4b574be…
|
wyoung
|
341 |
any use of |
|
64a33a6…
|
wyoung
|
342 |
[https://en.wikipedia.org/wiki/C11_(C_standard_revision)|C11]. Fossil |
|
4b574be…
|
wyoung
|
343 |
does call POSIX and Windows APIs where necessary, but it's about |
|
64a33a6…
|
wyoung
|
344 |
as portable as you can ask given that ISO C doesn't define all of the |
|
64a33a6…
|
wyoung
|
345 |
facilities Fossil needs to do its thing. (Network sockets, file locking, |
|
64a33a6…
|
wyoung
|
346 |
etc.) There are certainly well-known platforms Fossil hasn't been ported |
|
64a33a6…
|
wyoung
|
347 |
to yet, but that's most likely due to lack of interest rather than |
|
64a33a6…
|
wyoung
|
348 |
inherent difficulties in doing the port. We believe the most stringent |
|
64a33a6…
|
wyoung
|
349 |
limit on its portability is that it assumes at least a 32-bit CPU and |
|
b95abfd…
|
wyoung
|
350 |
several megs of flat-addressed memory.³ Fossil isn't quite as |
|
64a33a6…
|
wyoung
|
351 |
[https://www.sqlite.org/custombuild.html|portable as SQLite], but it's |
|
64a33a6…
|
wyoung
|
352 |
close. |
|
64a33a6…
|
wyoung
|
353 |
|
|
4b574be…
|
wyoung
|
354 |
Over half of the C code in Fossil is actually an embedded copy of the |
|
64a33a6…
|
wyoung
|
355 |
current version of SQLite. Much of what is Fossil-specific after you set |
|
64a33a6…
|
wyoung
|
356 |
SQLite itself aside is SQL code calling into SQLite. The number of lines |
|
64a33a6…
|
wyoung
|
357 |
of SQL code in Fossil isn't large by percentage, but since SQL is such |
|
64a33a6…
|
wyoung
|
358 |
an expressive, declarative language, it has an outsized contribution to |
|
64a33a6…
|
wyoung
|
359 |
Fossil's user-visible functionality. |
|
64a33a6…
|
wyoung
|
360 |
|
|
21c7f1f…
|
wyoung
|
361 |
Fossil isn't entirely C and SQL code. Its web UI [./javascript.md | |
|
21c7f1f…
|
wyoung
|
362 |
uses JavaScript where |
|
21c7f1f…
|
wyoung
|
363 |
necessary]. The server-side |
|
0996347…
|
wyoung
|
364 |
UI scripting uses a custom minimal |
|
4b574be…
|
wyoung
|
365 |
[https://en.wikipedia.org/wiki/Tcl|Tcl] dialect called |
|
56e59a3…
|
stephan
|
366 |
[./th1.md|TH1], which is |
|
0996347…
|
wyoung
|
367 |
embedded into Fossil itself. Fossil's build system and test suite are |
|
b95abfd…
|
wyoung
|
368 |
largely based on Tcl.⁴ All of this is quite portable. |
|
4b574be…
|
wyoung
|
369 |
|
|
4b574be…
|
wyoung
|
370 |
About half of Git's code is POSIX C, and about a third is POSIX shell |
|
4b574be…
|
wyoung
|
371 |
code. This is largely why the so-called "Git for Windows" distributions |
|
4b574be…
|
wyoung
|
372 |
(both [https://git-scm.com/download/win|first-party] and |
|
64a33a6…
|
wyoung
|
373 |
[https://gitforwindows.org/|third-party]) are actually an |
|
aacd371…
|
danield
|
374 |
[https://www.msys2.org/wiki/Home/|MSYS POSIX portability environment] bundled |
|
64a33a6…
|
wyoung
|
375 |
with all of the Git stuff, because it would be too painful to port Git |
|
64a33a6…
|
wyoung
|
376 |
natively to Windows. Git is a foreign citizen on Windows, speaking to it |
|
b95abfd…
|
wyoung
|
377 |
only through a translator.⁵ |
|
64a33a6…
|
wyoung
|
378 |
|
|
64a33a6…
|
wyoung
|
379 |
While Fossil does lean toward POSIX norms when given a choice — LF-only |
|
64a33a6…
|
wyoung
|
380 |
line endings are treated as first-class citizens over CR+LF, for example |
|
64a33a6…
|
wyoung
|
381 |
— the Windows build of Fossil is truly native. |
|
64a33a6…
|
wyoung
|
382 |
|
|
64a33a6…
|
wyoung
|
383 |
The third-party extensions to Git tend to follow this same pattern. |
|
f8baa04…
|
danield
|
384 |
[https://docs.gitlab.com/ee/install/install_methods.html#microsoft-windows | |
|
f8baa04…
|
danield
|
385 |
GitLab isn't portable to Windows at all], |
|
64a33a6…
|
wyoung
|
386 |
for example. For that matter, GitLab isn't even officially supported on |
|
4b574be…
|
wyoung
|
387 |
macOS, the BSDs, or uncommon Linuxes! We have many users who regularly |
|
4b574be…
|
wyoung
|
388 |
build and run Fossil on all of these systems. |
|
64a33a6…
|
wyoung
|
389 |
|
|
64a33a6…
|
wyoung
|
390 |
|
|
64a33a6…
|
wyoung
|
391 |
<h3 id="vs-linux">2.5 Linux vs. SQLite</h3> |
|
69ec89b…
|
wyoung
|
392 |
|
|
69ec89b…
|
wyoung
|
393 |
Fossil and Git promote different development styles because each one was |
|
0297ae7…
|
wyoung
|
394 |
specifically designed to support the creator's main software |
|
69ec89b…
|
wyoung
|
395 |
development project: [https://en.wikipedia.org/wiki/Linus_Torvalds|Linus |
|
69ec89b…
|
wyoung
|
396 |
Torvalds] designed Git to support development of |
|
69ec89b…
|
wyoung
|
397 |
[https://www.kernel.org/|the Linux kernel], and |
|
69ec89b…
|
wyoung
|
398 |
[https://en.wikipedia.org/wiki/D._Richard_Hipp|D. Richard Hipp] designed |
|
69ec89b…
|
wyoung
|
399 |
Fossil to support the development of [https://sqlite.org/|SQLite]. |
|
69ec89b…
|
wyoung
|
400 |
Both projects must rank high on any objective list of "most |
|
69ec89b…
|
wyoung
|
401 |
important FOSS projects," yet these two projects are almost entirely unlike |
|
466d74c…
|
wyoung
|
402 |
one another, so it is natural that the DVCSes created to support these |
|
466d74c…
|
wyoung
|
403 |
projects also differ in many ways. |
|
69ec89b…
|
wyoung
|
404 |
|
|
e9628d7…
|
wyoung
|
405 |
In the following sections, we will explain how four key differences |
|
466d74c…
|
wyoung
|
406 |
between the Linux and SQLite software development projects dictated the |
|
466d74c…
|
wyoung
|
407 |
design of each DVCS's low-friction usage path. |
|
69ec89b…
|
wyoung
|
408 |
|
|
69ec89b…
|
wyoung
|
409 |
When deciding between these two DVCSes, you should ask yourself, "Is my |
|
69ec89b…
|
wyoung
|
410 |
project more like Linux or more like SQLite?" |
|
69ec89b…
|
wyoung
|
411 |
|
|
69ec89b…
|
wyoung
|
412 |
|
|
64a33a6…
|
wyoung
|
413 |
<h4 id="devorg">2.5.1 Development Organization</h4> |
|
69ec89b…
|
wyoung
|
414 |
|
|
69ec89b…
|
wyoung
|
415 |
Eric S. Raymond's seminal essay-turned-book |
|
69ec89b…
|
wyoung
|
416 |
"[https://en.wikipedia.org/wiki/The_Cathedral_and_the_Bazaar|The |
|
69ec89b…
|
wyoung
|
417 |
Cathedral and the Bazaar]" details the two major development |
|
69ec89b…
|
wyoung
|
418 |
organization styles found in |
|
69ec89b…
|
wyoung
|
419 |
[https://en.wikipedia.org/wiki/Free_and_open-source_software|FOSS] |
|
69ec89b…
|
wyoung
|
420 |
projects. As it happens, Linux and SQLite fall on opposite sides of this |
|
69ec89b…
|
wyoung
|
421 |
dichotomy. Differing development organization styles dictate a different |
|
69ec89b…
|
wyoung
|
422 |
design and low-friction usage path in the tools created to support each |
|
69ec89b…
|
wyoung
|
423 |
project. |
|
69ec89b…
|
wyoung
|
424 |
|
|
69ec89b…
|
wyoung
|
425 |
Git promotes the Linux kernel's bazaar development style, in which a |
|
69ec89b…
|
wyoung
|
426 |
loosely-associated mass of developers contribute their work through |
|
69ec89b…
|
wyoung
|
427 |
[https://git-scm.com/book/en/v2/Distributed-Git-Distributed-Workflows#_dictator_and_lieutenants_workflow|a |
|
69ec89b…
|
wyoung
|
428 |
hierarchy of lieutenants] who manage and clean up these contributions |
|
0996347…
|
wyoung
|
429 |
for consideration by Linus Torvalds, who has the power to cherry-pick |
|
69ec89b…
|
wyoung
|
430 |
individual contributions into his version of the Linux kernel. Git |
|
69ec89b…
|
wyoung
|
431 |
allows an anonymous developer to rebase and push specific locally-named |
|
69ec89b…
|
wyoung
|
432 |
private branches, so that a Git repo clone often isn't really a clone at |
|
69ec89b…
|
wyoung
|
433 |
all: it may have an arbitrary number of differences relative to the |
|
69ec89b…
|
wyoung
|
434 |
repository it originally cloned from. Git encourages siloed development. |
|
69ec89b…
|
wyoung
|
435 |
Select work in a developer's local repository may remain private |
|
69ec89b…
|
wyoung
|
436 |
indefinitely. |
|
69ec89b…
|
wyoung
|
437 |
|
|
69ec89b…
|
wyoung
|
438 |
All of this is exactly what one wants when doing bazaar-style |
|
69ec89b…
|
wyoung
|
439 |
development. |
|
69ec89b…
|
wyoung
|
440 |
|
|
69ec89b…
|
wyoung
|
441 |
Fossil's normal mode of operation differs on every one of these points, |
|
69ec89b…
|
wyoung
|
442 |
with the specific designed-in goal of promoting SQLite's cathedral |
|
69ec89b…
|
wyoung
|
443 |
development model: |
|
69ec89b…
|
wyoung
|
444 |
|
|
1fd407f…
|
wyoung
|
445 |
* <b>Personal engagement:</b> SQLite's developers know each |
|
1fd407f…
|
wyoung
|
446 |
other by name and work together daily on the project. |
|
1fd407f…
|
wyoung
|
447 |
|
|
1fd407f…
|
wyoung
|
448 |
* <b>Trust over hierarchy:</b> SQLite's developers check |
|
1fd407f…
|
wyoung
|
449 |
changes into their local repository, and these are immediately and |
|
1fd407f…
|
wyoung
|
450 |
automatically synchronized up to the central repository; there is no |
|
1fd407f…
|
wyoung
|
451 |
"[https://git-scm.com/book/en/v2/Distributed-Git-Distributed-Workflows#_dictator_and_lieutenants_workflow|dictator |
|
1fd407f…
|
wyoung
|
452 |
and lieutenants]" hierarchy as with Linux kernel contributions. D. |
|
1fd407f…
|
wyoung
|
453 |
Richard Hipp rarely overrides decisions made by those he has trusted |
|
1fd407f…
|
wyoung
|
454 |
with commit access on his repositories. Fossil allows you to give |
|
1fd407f…
|
wyoung
|
455 |
[./caps/admin-v-setup.md|some users] more power over what |
|
1fd407f…
|
wyoung
|
456 |
they can do with the repository, but Fossil [./caps/index.md#ucap | |
|
1fd407f…
|
wyoung
|
457 |
only loosely supports] the enforcement of a development organization's |
|
1fd407f…
|
wyoung
|
458 |
social and power hierarchies. Fossil is a great fit for |
|
1fd407f…
|
wyoung
|
459 |
[https://en.wikipedia.org/wiki/Flat_organization|flat |
|
1fd407f…
|
wyoung
|
460 |
organizations]. |
|
1fd407f…
|
wyoung
|
461 |
|
|
1fd407f…
|
wyoung
|
462 |
* <b>No easy drive-by contributions:</b> Git |
|
1fd407f…
|
wyoung
|
463 |
[https://www.git-scm.com/docs/git-request-pull|pull requests] offer |
|
1fd407f…
|
wyoung
|
464 |
a low-friction path to accepting |
|
1fd407f…
|
wyoung
|
465 |
[https://www.jonobacon.com/2012/07/25/building-strong-community-structural-integrity/|drive-by |
|
1fd407f…
|
wyoung
|
466 |
contributions]. Fossil's closest equivalents are its unique |
|
c64f28d…
|
drh
|
467 |
[/help/bundle|bundle] and [/help/patch|patch] features, which require higher engagement |
|
b95abfd…
|
wyoung
|
468 |
than firing off a PR.⁶ This difference comes directly from the |
|
1fd407f…
|
wyoung
|
469 |
initial designed purpose for each tool: the SQLite project doesn't |
|
1fd407f…
|
wyoung
|
470 |
accept outside contributions from previously-unknown developers, but |
|
1fd407f…
|
wyoung
|
471 |
the Linux kernel does. |
|
1fd407f…
|
wyoung
|
472 |
|
|
1fd407f…
|
wyoung
|
473 |
* <b>No rebasing:</b> When your local repo clone syncs changes |
|
1fd407f…
|
wyoung
|
474 |
up to its parent, those changes are sent exactly as they were |
|
1fd407f…
|
wyoung
|
475 |
committed locally. [#history|There is no rebasing mechanism in |
|
1fd407f…
|
wyoung
|
476 |
Fossil, on purpose.] |
|
1fd407f…
|
wyoung
|
477 |
|
|
1fd407f…
|
wyoung
|
478 |
* <b>Sync over push:</b> Explicit pushes are uncommon in |
|
1fd407f…
|
wyoung
|
479 |
Fossil-based projects: the default is to rely on |
|
c64f28d…
|
drh
|
480 |
[/help/autosync|autosync mode] instead, in which each commit |
|
1fd407f…
|
wyoung
|
481 |
syncs immediately to its parent repository. This is a mode so you |
|
1fd407f…
|
wyoung
|
482 |
can turn it off temporarily when needed, such as when working |
|
1fd407f…
|
wyoung
|
483 |
offline. Fossil is still a truly distributed version control system; |
|
1fd407f…
|
wyoung
|
484 |
it's just that its starting default is to assume you're rarely out |
|
1fd407f…
|
wyoung
|
485 |
of communication with the parent repo. |
|
1fd407f…
|
wyoung
|
486 |
<br><br> |
|
1fd407f…
|
wyoung
|
487 |
This is not merely a reflection of modern always-connected computing |
|
1fd407f…
|
wyoung
|
488 |
environments. It is a conscious decision in direct support of |
|
1fd407f…
|
wyoung
|
489 |
SQLite's cathedral development model: we don't want developers going |
|
1fd407f…
|
wyoung
|
490 |
dark, then showing up weeks later with a massive bolus of changes |
|
1fd407f…
|
wyoung
|
491 |
for us to integrate all at once. |
|
1fd407f…
|
wyoung
|
492 |
[https://en.wikipedia.org/wiki/Jim_McCarthy_(author)|Jim McCarthy] |
|
1fd407f…
|
wyoung
|
493 |
put it well in his book on software project management, |
|
1fd407f…
|
wyoung
|
494 |
<i>[https://www.amazon.com/dp/0735623198/|Dynamics of Software |
|
1fd407f…
|
wyoung
|
495 |
Development]</i>: "[https://www.youtube.com/watch?v=oY6BCHqEbyc|Beware |
|
1fd407f…
|
wyoung
|
496 |
of a guy in a room]." |
|
1fd407f…
|
wyoung
|
497 |
|
|
1fd407f…
|
wyoung
|
498 |
* <b>Branch names sync:</b> Unlike in Git, branch names in |
|
1fd407f…
|
wyoung
|
499 |
Fossil are not purely local labels. They sync along with everything |
|
1fd407f…
|
wyoung
|
500 |
else, so everyone sees the same set of branch names. Fossil's design |
|
1fd407f…
|
wyoung
|
501 |
choice here is a direct reflection of the Linux vs. SQLite project |
|
1fd407f…
|
wyoung
|
502 |
outlook: SQLite's developers collaborate closely on a single |
|
1fd407f…
|
wyoung
|
503 |
coherent project, whereas Linux's developers go off on tangents and |
|
1fd407f…
|
wyoung
|
504 |
occasionally send selected change sets to each other. |
|
1fd407f…
|
wyoung
|
505 |
|
|
1fd407f…
|
wyoung
|
506 |
* <b>Private branches are rare:</b> |
|
1fd407f…
|
wyoung
|
507 |
[/doc/trunk/www/private.wiki|Private branches exist in Fossil], but |
|
1fd407f…
|
wyoung
|
508 |
they're normally used to handle rare exception cases, whereas in |
|
1fd407f…
|
wyoung
|
509 |
many Git projects, they're part of the straight-line development |
|
1fd407f…
|
wyoung
|
510 |
process. |
|
1fd407f…
|
wyoung
|
511 |
|
|
1fd407f…
|
wyoung
|
512 |
* <b>Identical clones:</b> Fossil's autosync system tries to |
|
1fd407f…
|
wyoung
|
513 |
keep each local clone identical to the repository it cloned |
|
1fd407f…
|
wyoung
|
514 |
from. |
|
69ec89b…
|
wyoung
|
515 |
|
|
69ec89b…
|
wyoung
|
516 |
Where Git encourages siloed development, Fossil fights against it. |
|
69e6418…
|
wyoung
|
517 |
Fossil places a lot of emphasis on synchronizing everyone's work and on |
|
69e6418…
|
wyoung
|
518 |
reporting on the state of the project and the work of its developers, so |
|
69e6418…
|
wyoung
|
519 |
that everyone — especially the project leader — can maintain a better |
|
69e6418…
|
wyoung
|
520 |
mental picture of what is happening, leading to better situational |
|
69e6418…
|
wyoung
|
521 |
awareness. |
|
c0269e3…
|
wyoung
|
522 |
|
|
a59af25…
|
danield
|
523 |
By contrast, "…[https://docs.github.com/en/get-started/quickstart/contributing-to-projects|forking is |
|
765b215…
|
wyoung
|
524 |
at the core of social coding at GitHub]". As of June 2026, |
|
765b215…
|
wyoung
|
525 |
[https://github.com/search?q=is:public|Github hosts 324 million distinct |
|
c0269e3…
|
wyoung
|
526 |
software projects], most of which were created by forking a |
|
c0269e3…
|
wyoung
|
527 |
previously-existing project. Since this is |
|
a90f60c…
|
wyoung
|
528 |
[https://www.griddynamics.com/blog/number-software-developers-world | |
|
a90f60c…
|
wyoung
|
529 |
~11× the number of developers in the world], it beggars belief that |
|
c0269e3…
|
wyoung
|
530 |
most of these forks are still under active development. The vast bulk |
|
c0269e3…
|
wyoung
|
531 |
of these must be abandoned one-off efforts. This is part of the nature |
|
c0269e3…
|
wyoung
|
532 |
of bazaar style development. |
|
b0f2a48…
|
wyoung
|
533 |
|
|
b0f2a48…
|
wyoung
|
534 |
You can think about this difference in terms of |
|
b0f2a48…
|
wyoung
|
535 |
[https://en.wikipedia.org/wiki/Feedback | feedback loop size], which we |
|
b0f2a48…
|
wyoung
|
536 |
know from the mathematics of |
|
b0f2a48…
|
wyoung
|
537 |
[https://en.wikipedia.org/wiki/Control_theory | control theory] to |
|
b0f2a48…
|
wyoung
|
538 |
directly affect the speed at which any system can safely make changes. |
|
b0f2a48…
|
wyoung
|
539 |
The larger the feedback loop, the slower the whole system must run in |
|
b0f2a48…
|
wyoung
|
540 |
order to avoid loss of control. The same concept shows up in other |
|
b0f2a48…
|
wyoung
|
541 |
contexts, such as in the [https://en.wikipedia.org/wiki/OODA_loop | OODA |
|
c00b6a4…
|
wyoung
|
542 |
loop] concept. |
|
b0f2a48…
|
wyoung
|
543 |
Committing your changes to private branches in order to delay a public |
|
b0f2a48…
|
wyoung
|
544 |
push to the parent repo increases the size of your collaborators' |
|
b0f2a48…
|
wyoung
|
545 |
control loops, either causing them to slow their work in order to safely |
|
e28c25e…
|
wyoung
|
546 |
react to your work, or to over-correct in response to each change. |
|
64a33a6…
|
wyoung
|
547 |
|
|
69ec89b…
|
wyoung
|
548 |
Each DVCS can be used in the opposite style, but doing so works against |
|
69ec89b…
|
wyoung
|
549 |
their low-friction paths. |
|
69ec89b…
|
wyoung
|
550 |
|
|
8452627…
|
wyoung
|
551 |
|
|
64a33a6…
|
wyoung
|
552 |
<h4 id="scale">2.5.2 Scale</h4> |
|
69ec89b…
|
wyoung
|
553 |
|
|
69ec89b…
|
wyoung
|
554 |
The Linux kernel has a far bigger developer community than that of |
|
69ec89b…
|
wyoung
|
555 |
SQLite: there are thousands and thousands of contributors to Linux, most |
|
e28c25e…
|
wyoung
|
556 |
of whom do not know each other's names. These thousands are responsible |
|
21c98b4…
|
wyoung
|
557 |
for producing roughly 73× more code than is in SQLite. (32.0 |
|
21c98b4…
|
wyoung
|
558 |
[https://en.wikipedia.org/wiki/Source_lines_of_code|MLOC] vs. 0.44 MLOC |
|
e4beb0b…
|
wyoung
|
559 |
according to [https://github.com/boyter/scc | scc].) The Linux |
|
69ec89b…
|
wyoung
|
560 |
kernel and its development process were already uncommonly large back in |
|
69ec89b…
|
wyoung
|
561 |
2005 when Git was designed, specifically to support the consequences of |
|
69ec89b…
|
wyoung
|
562 |
having such a large set of developers working on such a large code base. |
|
69ec89b…
|
wyoung
|
563 |
|
|
21c98b4…
|
wyoung
|
564 |
95% of the code in SQLite comes from just six programmers, and 62% of |
|
69ec89b…
|
wyoung
|
565 |
it is from the lead developer alone. The SQLite developers know each |
|
69ec89b…
|
wyoung
|
566 |
other well and interact daily. Fossil was designed for this development |
|
0e0d76e…
|
wyoung
|
567 |
model. |
|
0e0d76e…
|
wyoung
|
568 |
|
|
5171e59…
|
wyoung
|
569 |
When choosing your DVCS, we think you should ask yourself whether the |
|
5171e59…
|
wyoung
|
570 |
scale of your software configuration management problems is closer to |
|
5171e59…
|
wyoung
|
571 |
those Linus Torvalds designed Git to cope with or whether your work's |
|
5171e59…
|
wyoung
|
572 |
scale is closer to that of SQLite, for which D. Richard Hipp designed |
|
5171e59…
|
wyoung
|
573 |
Fossil. An [https://en.wikipedia.org/wiki/Impact_wrench|automotive air |
|
5171e59…
|
wyoung
|
574 |
impact wrench] running at 8000 RPM driving an M8 socket-cap bolt at 16 |
|
5171e59…
|
wyoung
|
575 |
cm/s is not the best way to hang a picture on the living room wall. |
|
5171e59…
|
wyoung
|
576 |
|
|
5171e59…
|
wyoung
|
577 |
Fossil works well for projects several times the size of SQLite, |
|
ab48b9d…
|
wyoung
|
578 |
[https://core.tcl-lang.org/tcl/ | such as Tcl], with a repository over |
|
5171e59…
|
wyoung
|
579 |
twice the size and with many more core committers. |
|
b3080a2…
|
drh
|
580 |
|
|
b3080a2…
|
drh
|
581 |
|
|
b3080a2…
|
drh
|
582 |
<h4 id="branches">2.5.3 Individual Branches vs. The Entire Change History</h4> |
|
bcea529…
|
drh
|
583 |
|
|
bcea529…
|
drh
|
584 |
Both Fossil and Git store history as a directed acyclic graph (DAG) |
|
fd99012…
|
wyoung
|
585 |
of changes, but Git tends to focus more on individual branches of |
|
bcea529…
|
drh
|
586 |
the DAG, whereas Fossil puts more emphasis on the entire DAG. |
|
bcea529…
|
drh
|
587 |
|
|
03f2393…
|
wyoung
|
588 |
While a common usage pattern in Git is to only synchronize |
|
03f2393…
|
wyoung
|
589 |
a single branch — <tt>git pull upstream feature/branch</tt> — instead |
|
4146f2d…
|
wyoung
|
590 |
of all refs, Fossil does not give you a choice; it |
|
03f2393…
|
wyoung
|
591 |
syncs the entire DAG or nothing. Git commands, |
|
fd99012…
|
wyoung
|
592 |
GitHub, and GitLab tend to show only a single branch at |
|
fd99012…
|
wyoung
|
593 |
a time, whereas Fossil usually shows all parallel branches at |
|
bcea529…
|
drh
|
594 |
once. Git has commands like "rebase" that help keep all relevant |
|
bcea529…
|
drh
|
595 |
changes on a single branch, whereas Fossil encourages a style of |
|
79c2cb0…
|
wyoung
|
596 |
many concurrent branches constantly springing into existence, |
|
bcea529…
|
drh
|
597 |
undergoing active development in parallel for a few days or weeks, then |
|
bcea529…
|
drh
|
598 |
merging back into the main line and disappearing. |
|
bcea529…
|
drh
|
599 |
|
|
bcea529…
|
drh
|
600 |
This difference in emphasis arises from the different purposes of |
|
4146f2d…
|
wyoung
|
601 |
the two systems. Git's focus on individual branches |
|
bcea529…
|
drh
|
602 |
is exactly what you want for a highly-distributed bazaar-style project |
|
bcea529…
|
drh
|
603 |
such as Linux. Linus Torvalds does not want to see every check-in |
|
e28c25e…
|
wyoung
|
604 |
by every contributor to Linux: such extreme visibility does not scale |
|
e28c25e…
|
wyoung
|
605 |
well. Contrast Fossil, which was written for the cathedral-style SQLite project |
|
e28c25e…
|
wyoung
|
606 |
and its handful of active committers. Seeing all |
|
bcea529…
|
drh
|
607 |
changes on all branches all at once helps keep the whole team |
|
fd1282e…
|
drh
|
608 |
up-to-date with what everybody else is doing, resulting in a more |
|
bcea529…
|
drh
|
609 |
tightly focused and cohesive implementation. |
|
bcea529…
|
drh
|
610 |
|
|
69ec89b…
|
wyoung
|
611 |
|
|
64a33a6…
|
wyoung
|
612 |
<h3 id="checkouts">2.6 One vs. Many Check-outs per Repository</h3> |
|
64a33a6…
|
wyoung
|
613 |
|
|
f6e04cc…
|
wyoung
|
614 |
Because Git commingles the repository data with the initial checkout of |
|
f6e04cc…
|
wyoung
|
615 |
that repository, the default mode of operation in Git is to stick to that |
|
f6e04cc…
|
wyoung
|
616 |
single work/repo tree, even when that's a shortsighted way of working. |
|
10a57ce…
|
wyoung
|
617 |
|
|
c0654b1…
|
brickviking
|
618 |
Fossil doesn't work that way. A Fossil repository is an SQLite database |
|
f6e04cc…
|
wyoung
|
619 |
file which is normally stored outside the working checkout directory. You can |
|
c64f28d…
|
drh
|
620 |
[/help/open | open] a Fossil repository any number of times into |
|
10a57ce…
|
wyoung
|
621 |
any number of working directories. A common usage pattern is to have one |
|
10a57ce…
|
wyoung
|
622 |
working directory per active working branch, so that switching branches |
|
10a57ce…
|
wyoung
|
623 |
is done with a <tt>cd</tt> command rather than by checking out the |
|
10a57ce…
|
wyoung
|
624 |
branches successively in a single working directory. |
|
10a57ce…
|
wyoung
|
625 |
|
|
f6e04cc…
|
wyoung
|
626 |
Fossil does allow you to switch branches within a working checkout |
|
f6e04cc…
|
wyoung
|
627 |
directory, and this is also often done. It is simply that there is no |
|
f6e04cc…
|
wyoung
|
628 |
inherent penalty to either choice in Fossil as there is in Git. The |
|
f6e04cc…
|
wyoung
|
629 |
standard advice is to use a switch-in-place workflow in Fossil when |
|
f6e04cc…
|
wyoung
|
630 |
the disturbance from switching branches is small, and to use multiple |
|
f6e04cc…
|
wyoung
|
631 |
checkouts when you have long-lived working branches that are different |
|
f6e04cc…
|
wyoung
|
632 |
enough that switching in place is disruptive. |
|
f6e04cc…
|
wyoung
|
633 |
|
|
2bfb72a…
|
wyoung
|
634 |
While you can [./gitusers.md#worktree | use Git in the Fossil style] via |
|
2bfb72a…
|
wyoung
|
635 |
its worktree feature, tutorials continue to teach the style of having |
|
2bfb72a…
|
wyoung
|
636 |
one working directory only. Git can even fight you on this, as when |
|
2bfb72a…
|
wyoung
|
637 |
working with a forked repository; it is best to have independent clones |
|
2bfb72a…
|
wyoung
|
638 |
of the upstream and your fork, to allow separate "remote" lists and |
|
2bfb72a…
|
wyoung
|
639 |
such. This can result in two working directories but each having a |
|
2bfb72a…
|
wyoung
|
640 |
captive repo clone each, which isn't in the spirit of <tt>git |
|
2bfb72a…
|
wyoung
|
641 |
worktree</tt> at all. Yet, it beats the alternative, which then |
|
2bfb72a…
|
wyoung
|
642 |
highlights a gap in the model of diverging and re-converging forks. |
|
2bfb72a…
|
wyoung
|
643 |
Ideally, Git would let you create one of these forks as a worktree while |
|
2bfb72a…
|
wyoung
|
644 |
maintaining a strong separation between your fork and the upstream repo, |
|
2bfb72a…
|
wyoung
|
645 |
but it ends up being too much hassle to bother with. |
|
2bfb72a…
|
wyoung
|
646 |
|
|
2bfb72a…
|
wyoung
|
647 |
There are |
|
10a57ce…
|
wyoung
|
648 |
[https://duckduckgo.com/?q=git+worktree+problem | several known |
|
2bfb72a…
|
wyoung
|
649 |
problems] with the single worktree style, ones which don't happen in |
|
2bfb72a…
|
wyoung
|
650 |
Fossil because of the clear [./ckout-workflows.md | separation] between |
|
2bfb72a…
|
wyoung
|
651 |
a Fossil repository and each working directory. |
|
10a57ce…
|
wyoung
|
652 |
|
|
f6e04cc…
|
wyoung
|
653 |
This distinction matters because switching branches inside a single working directory loses local context |
|
10a57ce…
|
wyoung
|
654 |
on each switch. |
|
10a57ce…
|
wyoung
|
655 |
|
|
10a57ce…
|
wyoung
|
656 |
For instance, in any software project where the runnable program must be |
|
10a57ce…
|
wyoung
|
657 |
built from source files, you invalidate build objects on each switch, |
|
f6e04cc…
|
wyoung
|
658 |
artificially increasing the time required to switch versions. Most obviously, this |
|
10a57ce…
|
wyoung
|
659 |
affects software written in statically-compiled programming languages |
|
10a57ce…
|
wyoung
|
660 |
such as C, Java, and Haskell, but it can even affect programs written in |
|
f6e04cc…
|
wyoung
|
661 |
dynamic languages like JavaScript. A typical |
|
10a57ce…
|
wyoung
|
662 |
[https://en.wikipedia.org/wiki/Single-page_application | SPA] build |
|
f6e04cc…
|
wyoung
|
663 |
process involves several passes: [http://browserify.org/ | Browserify] to convert |
|
f6e04cc…
|
wyoung
|
664 |
[https://nodejs.org/ | Node] packages so they'll run in a web browser, |
|
f6e04cc…
|
wyoung
|
665 |
[https://sass-lang.com | SASS] to CSS translation, |
|
f6e04cc…
|
wyoung
|
666 |
transpilation of [https://www.typescriptlang.org | Typescript] to JavaScript, |
|
f6e04cc…
|
wyoung
|
667 |
[https://github.com/mishoo/UglifyJS | uglification], etc. |
|
f6e04cc…
|
wyoung
|
668 |
Once all that processing work is done for a given input |
|
10a57ce…
|
wyoung
|
669 |
file in a given working directory, why re-do that work just to switch |
|
f6e04cc…
|
wyoung
|
670 |
versions? If most of the files that differ between versions don't change |
|
f6e04cc…
|
wyoung
|
671 |
very often, you can save substantial time by switching branches with |
|
f6e04cc…
|
wyoung
|
672 |
<tt>cd</tt> rather than swapping versions in-place within a working |
|
f6e04cc…
|
wyoung
|
673 |
checkout directory. |
|
10a57ce…
|
wyoung
|
674 |
|
|
10a57ce…
|
wyoung
|
675 |
For another example, you might have an active long-running test grinding |
|
10a57ce…
|
wyoung
|
676 |
away in a working directory, then get a call from a customer requiring |
|
10a57ce…
|
wyoung
|
677 |
that you switch to a stable branch to answer questions in terms of the |
|
10a57ce…
|
wyoung
|
678 |
version that customer is running. You don't want to stop the test in |
|
10a57ce…
|
wyoung
|
679 |
order to switch your lone working directory to the stable branch. |
|
10a57ce…
|
wyoung
|
680 |
|
|
e28c25e…
|
wyoung
|
681 |
Disk space is cheap. Having several working directories — each with its |
|
e28c25e…
|
wyoung
|
682 |
own local state — makes switching versions cheap and fast. |
|
e28c25e…
|
wyoung
|
683 |
|
|
e28c25e…
|
wyoung
|
684 |
Plus, |
|
f6e04cc…
|
wyoung
|
685 |
<tt>cd</tt> is faster to type than <tt>git checkout</tt> or <tt>fossil |
|
f6e04cc…
|
wyoung
|
686 |
update</tt>. |
|
64a33a6…
|
wyoung
|
687 |
|
|
64a33a6…
|
wyoung
|
688 |
<h3 id="history">2.7 What you should have done vs. What you actually did</h3> |
|
ca891ab…
|
drh
|
689 |
|
|
ca891ab…
|
drh
|
690 |
Git puts a lot of emphasis on maintaining |
|
ca891ab…
|
drh
|
691 |
a "clean" check-in history. Extraneous and experimental branches by |
|
18fc697…
|
wyoung
|
692 |
individual developers often never make it into the main repository. |
|
18fc697…
|
wyoung
|
693 |
Branches may be rebased before being pushed to make |
|
18fc697…
|
wyoung
|
694 |
it appear as if development had been linear, or "squashed" to make it |
|
18fc697…
|
wyoung
|
695 |
appear that multiple commits were made as a single commit. |
|
18fc697…
|
wyoung
|
696 |
There are [https://git-scm.com/book/en/v2/Git-Tools-Rewriting-History | |
|
18fc697…
|
wyoung
|
697 |
other history rewriting mechanisms in Git] as well. Git strives to record what |
|
ca891ab…
|
drh
|
698 |
the development of a project should have looked like had there been no |
|
ca891ab…
|
drh
|
699 |
mistakes. |
|
ca891ab…
|
drh
|
700 |
|
|
ca891ab…
|
drh
|
701 |
Fossil, in contrast, puts more emphasis on recording exactly what happened, |
|
ca891ab…
|
drh
|
702 |
including all of the messy errors, dead-ends, experimental branches, and |
|
ca891ab…
|
drh
|
703 |
so forth. One might argue that this |
|
29997f8…
|
wyoung
|
704 |
makes the history of a Fossil project "messy," but another point of view |
|
fd99012…
|
wyoung
|
705 |
is that this makes the history "accurate." In actual practice, the |
|
e28c25e…
|
wyoung
|
706 |
superior reporting tools available in Fossil mean that this incidental mess |
|
ca891ab…
|
drh
|
707 |
is not a factor. |
|
ca891ab…
|
drh
|
708 |
|
|
c64f28d…
|
drh
|
709 |
Like Git, Fossil has an [/help/amend|amend command] for modifying |
|
e878d54…
|
wyoung
|
710 |
prior commits, but unlike in Git, this works not by replacing data in |
|
e878d54…
|
wyoung
|
711 |
the repository, but by adding a correction record to the repository that |
|
e878d54…
|
wyoung
|
712 |
affects how later Fossil operations present the corrected data. The old |
|
e878d54…
|
wyoung
|
713 |
information is still there in the repository, it is just overridden from |
|
18fc697…
|
wyoung
|
714 |
the amendment point forward. |
|
18fc697…
|
wyoung
|
715 |
|
|
18fc697…
|
wyoung
|
716 |
Fossil lacks almost every other history rewriting mechanism listed on |
|
18fc697…
|
wyoung
|
717 |
the Git documentation page linked above. [./rebaseharm.md | There is no |
|
18fc697…
|
wyoung
|
718 |
rebase] in Fossil, on purpose, thus no way to reorder or copy commits |
|
18fc697…
|
wyoung
|
719 |
around in the commit hash tree. There is no commit squashing, dropping, |
|
18fc697…
|
wyoung
|
720 |
or interactive patch-based cherry-picking of commit elements in Fossil. |
|
18fc697…
|
wyoung
|
721 |
There is nothing like Git's <tt>filter-branch</tt> in Fossil. |
|
18fc697…
|
wyoung
|
722 |
|
|
18fc697…
|
wyoung
|
723 |
The lone exception is deleting commits. Fossil has two methods for doing |
|
18fc697…
|
wyoung
|
724 |
that, both of which have stringent limitations, on purpose. |
|
18fc697…
|
wyoung
|
725 |
|
|
18fc697…
|
wyoung
|
726 |
The first is [/doc/trunk/www/shunning.wiki | shunning]. See that |
|
18fc697…
|
wyoung
|
727 |
document for details, but briefly, you only get mandatory compliance |
|
18fc697…
|
wyoung
|
728 |
for shun requests within a single repository. Shun requests do not |
|
18fc697…
|
wyoung
|
729 |
propagate automatically between repository clones. A Fossil repository |
|
18fc697…
|
wyoung
|
730 |
administrator can <i>cooperatively</i> pull another repo's shun requests |
|
18fc697…
|
wyoung
|
731 |
across a sync boundary, so that two admins can get together and agree to |
|
18fc697…
|
wyoung
|
732 |
shun certain committed artifacts, but a person cannot force their local |
|
18fc697…
|
wyoung
|
733 |
shun requests into another repo without having admin-level control over |
|
18fc697…
|
wyoung
|
734 |
the receiving repo as well. Fossil's shun feature isn't for fixing up |
|
18fc697…
|
wyoung
|
735 |
everyday bad commits, it's for dealing with extreme situations: public |
|
18fc697…
|
wyoung
|
736 |
commits of secret material, ticket/wiki/forum spam, law enforcement |
|
18fc697…
|
wyoung
|
737 |
takedown demands, etc. |
|
18fc697…
|
wyoung
|
738 |
|
|
c64f28d…
|
drh
|
739 |
There is also the experimental [/help/purge | <tt>purge</tt> |
|
18fc697…
|
wyoung
|
740 |
command], which differs from shunning in ways that aren't especially |
|
18fc697…
|
wyoung
|
741 |
important in the context of this document. At a 30000 foot level, you |
|
18fc697…
|
wyoung
|
742 |
can think of purging as useful only when you've turned off Fossil's |
|
18fc697…
|
wyoung
|
743 |
autosync feature and want to pluck artifacts out of its hash tree before |
|
18fc697…
|
wyoung
|
744 |
they get pushed. In that sense, it's approximately the same as |
|
18fc697…
|
wyoung
|
745 |
<tt>git rebase -i, drop</tt>. However, given that Fossil defaults to |
|
18fc697…
|
wyoung
|
746 |
having autosync enabled [#devorg | for good reason], the purge command |
|
18fc697…
|
wyoung
|
747 |
isn't very useful in practice: once a commit has been pushed into |
|
18fc697…
|
wyoung
|
748 |
another repo, shunning is more useful if you need to delete it from |
|
18fc697…
|
wyoung
|
749 |
history. |
|
18fc697…
|
wyoung
|
750 |
|
|
18fc697…
|
wyoung
|
751 |
If these accommodations strike you as incoherent with respect to |
|
18fc697…
|
wyoung
|
752 |
Fossil's philosophy of durable, unchanging commits, realize that if |
|
18fc697…
|
wyoung
|
753 |
shunning and purging were removed from Fossil, you could still remove |
|
18fc697…
|
wyoung
|
754 |
artifacts from the repository with SQL <tt>DELETE</tt> statements; the |
|
18fc697…
|
wyoung
|
755 |
repository database file is, after all, directly modifiable, being |
|
18fc697…
|
wyoung
|
756 |
writable by your user. Where the Fossil philosophy really takes hold is |
|
18fc697…
|
wyoung
|
757 |
in making it difficult to violate the integrity of the hash tree. |
|
18fc697…
|
wyoung
|
758 |
It's somewhat tangential, but the document [./blockchain.md | "Is Fossil |
|
18fc697…
|
wyoung
|
759 |
a Blockchain?"] touches on this and related topics. |
|
e878d54…
|
wyoung
|
760 |
|
|
e878d54…
|
wyoung
|
761 |
One commentator characterized Git as recording history according to |
|
ca891ab…
|
drh
|
762 |
the victors, whereas Fossil records history as it actually happened. |
|
ca891ab…
|
drh
|
763 |
|
|
e5ba457…
|
wyoung
|
764 |
|
|
e5ba457…
|
wyoung
|
765 |
<h3 id="testing">2.8 Test Before Commit</h3> |
|
e5ba457…
|
wyoung
|
766 |
|
|
e5ba457…
|
wyoung
|
767 |
One of the things that falls out of Git's default separation of commit |
|
e5ba457…
|
wyoung
|
768 |
from push is that there are several Git sub-commands that jump straight |
|
e5ba457…
|
wyoung
|
769 |
to the commit step before a change could possibly be tested. Fossil, by |
|
e5ba457…
|
wyoung
|
770 |
contrast, makes the equivalent change to the local working check-out |
|
e5ba457…
|
wyoung
|
771 |
only, requiring a separate check-in step to commit the change. This |
|
e5ba457…
|
wyoung
|
772 |
design difference falls naturally out of Fossil's default-enabled |
|
18fc697…
|
wyoung
|
773 |
autosync feature and its philosophy of [#history | not offering history |
|
18fc697…
|
wyoung
|
774 |
rewriting features]. |
|
e5ba457…
|
wyoung
|
775 |
|
|
e5ba457…
|
wyoung
|
776 |
The prime example in Git is rebasing: the change happens to the local |
|
e5ba457…
|
wyoung
|
777 |
repository immediately if successful, even though you haven't tested the |
|
e5ba457…
|
wyoung
|
778 |
change yet. It's possible to argue for such a design in a tool like Git |
|
18fc697…
|
wyoung
|
779 |
since it lacks an autosync feature, because you can still test the |
|
18fc697…
|
wyoung
|
780 |
change before pushing local changes to the parent repo, but in the |
|
18fc697…
|
wyoung
|
781 |
meantime you've made a durable change to your local Git repository. You |
|
18fc697…
|
wyoung
|
782 |
must do something drastic like <tt>git reset --hard</tt> to revert that |
|
18fc697…
|
wyoung
|
783 |
rebase or rewrite history before pushing it if the rebase causes a |
|
18fc697…
|
wyoung
|
784 |
problem. If you push your rebased local repo up to the parent without |
|
18fc697…
|
wyoung
|
785 |
testing first, you cannot fix it without violating |
|
e5ba457…
|
wyoung
|
786 |
[https://www.atlassian.com/git/tutorials/merging-vs-rebasing#the-golden-rule-of-rebasing |
|
e5ba457…
|
wyoung
|
787 |
| the golden rule of rebasing]. |
|
e5ba457…
|
wyoung
|
788 |
|
|
e5ba457…
|
wyoung
|
789 |
Lesser examples are the Git <tt>merge</tt>, <tt>cherry-pick</tt>, and |
|
e5ba457…
|
wyoung
|
790 |
<tt>revert</tt> commands, all of which apply work from one branch onto |
|
18fc697…
|
wyoung
|
791 |
another, and all of which commit their change to the local repository |
|
18fc697…
|
wyoung
|
792 |
immediately without giving you |
|
18fc697…
|
wyoung
|
793 |
an opportunity to test the change first unless you give the |
|
18fc697…
|
wyoung
|
794 |
<tt>--no-commit</tt> option. Otherwise, you're back in the same boat: |
|
18fc697…
|
wyoung
|
795 |
reset the local repository or rewrite history to fix things, then maybe |
|
18fc697…
|
wyoung
|
796 |
retry. |
|
e5ba457…
|
wyoung
|
797 |
|
|
e5ba457…
|
wyoung
|
798 |
Fossil cannot sensibly work that way because of its default-enabled |
|
18fc697…
|
wyoung
|
799 |
autosync feature and its purposeful paucity of commands for modifying |
|
18fc697…
|
wyoung
|
800 |
commits, as discussed in [#history | the prior section]. |
|
18fc697…
|
wyoung
|
801 |
|
|
18fc697…
|
wyoung
|
802 |
Instead of jumping straight to the commit step, Fossil |
|
e5ba457…
|
wyoung
|
803 |
applies the proposed merge to the local working directory only, |
|
e5ba457…
|
wyoung
|
804 |
requiring a separate check-in step before the change is committed to the |
|
1ddb400…
|
wyoung
|
805 |
repository. This gives you a chance to test the change first, |
|
2cfd125…
|
wyoung
|
806 |
either manually or by running your software's automatic tests. (Ideally, |
|
18fc697…
|
wyoung
|
807 |
both!) Thus, Fossil doesn't need rebase, squashing, |
|
18fc697…
|
wyoung
|
808 |
<tt>reset --hard</tt>, or other Git commit mutating mechanisms. |
|
e5ba457…
|
wyoung
|
809 |
|
|
18fc697…
|
wyoung
|
810 |
Because Fossil requires an explicit commit for a merge, it has the nice |
|
18fc697…
|
wyoung
|
811 |
side benefit that it makes you give an explicit commit <i>message</i> |
|
18fc697…
|
wyoung
|
812 |
for each merge, whereas Git writes that commit message itself by default |
|
e5ba457…
|
wyoung
|
813 |
unless you give the optional <tt>--edit</tt> flag to override it. |
|
e5ba457…
|
wyoung
|
814 |
|
|
e5ba457…
|
wyoung
|
815 |
We don't look at this difference as a workaround in Fossil for autosync, |
|
18fc697…
|
wyoung
|
816 |
but instead as a test-first philosophical difference: |
|
18fc697…
|
wyoung
|
817 |
<tt>fossil commit</tt> is a <i>commitment</i>. When every commit is |
|
18fc697…
|
wyoung
|
818 |
pushed to the parent repo by default, it encourages a working style in |
|
18fc697…
|
wyoung
|
819 |
which every commit is tested first. It encourages thinking before |
|
18fc697…
|
wyoung
|
820 |
acting. We believe this is an inherently good thing. |
|
e5ba457…
|
wyoung
|
821 |
|
|
e5ba457…
|
wyoung
|
822 |
Incidentally, this is a good example of Git's messy command design. |
|
e5ba457…
|
wyoung
|
823 |
These three commands: |
|
e5ba457…
|
wyoung
|
824 |
|
|
e5ba457…
|
wyoung
|
825 |
<pre> |
|
8a1ba49…
|
wyoung
|
826 |
$ git merge HASH |
|
8a1ba49…
|
wyoung
|
827 |
$ git cherry-pick HASH |
|
8a1ba49…
|
wyoung
|
828 |
$ git revert HASH |
|
e5ba457…
|
wyoung
|
829 |
</pre> |
|
e5ba457…
|
wyoung
|
830 |
|
|
e5ba457…
|
wyoung
|
831 |
...are all the same command in Fossil: |
|
e5ba457…
|
wyoung
|
832 |
|
|
e5ba457…
|
wyoung
|
833 |
<pre> |
|
8a1ba49…
|
wyoung
|
834 |
$ fossil merge HASH |
|
8a1ba49…
|
wyoung
|
835 |
$ fossil merge --cherrypick HASH |
|
8a1ba49…
|
wyoung
|
836 |
$ fossil merge --backout HASH |
|
e5ba457…
|
wyoung
|
837 |
</pre> |
|
e5ba457…
|
wyoung
|
838 |
|
|
e5ba457…
|
wyoung
|
839 |
If you think about it, they're all the same function: apply work done on |
|
e5ba457…
|
wyoung
|
840 |
one branch to another. All that changes between these commands is how |
|
e5ba457…
|
wyoung
|
841 |
much work gets applied — just one check-in or a whole branch — and the |
|
e5ba457…
|
wyoung
|
842 |
merge direction. This is the sort of thing we mean when we point out |
|
e5ba457…
|
wyoung
|
843 |
that Fossil's command interface is simpler than Git's: there are fewer |
|
e5ba457…
|
wyoung
|
844 |
concepts to keep track of in your mental model of Fossil's internal |
|
e5ba457…
|
wyoung
|
845 |
operation. |
|
e5ba457…
|
wyoung
|
846 |
|
|
e5ba457…
|
wyoung
|
847 |
Fossil's implementation of the feature is also simpler to describe. The |
|
c64f28d…
|
drh
|
848 |
brief online help for <tt>[/help/merge | fossil merge]</tt> is |
|
26a96e8…
|
wyoung
|
849 |
currently 50 lines long, to which you want to add the ~800 lines of |
|
8f7576b…
|
wyoung
|
850 |
[./branching.wiki | the branching document]. The equivalent |
|
8f7576b…
|
wyoung
|
851 |
documentation in Git is the aggregation of the man pages for the above |
|
26a96e8…
|
wyoung
|
852 |
three commands, which is approaching 1400 lines as of this writing, much of it mutually redundant. |
|
2cfd125…
|
wyoung
|
853 |
(e.g. Git's <tt>--edit</tt> and <tt>--no-commit</tt> options get |
|
d887a6d…
|
wyoung
|
854 |
described three times, each time differently.) Fossil's |
|
8f7576b…
|
wyoung
|
855 |
documentation is not only more concise, it gives a nice split of brief |
|
8f7576b…
|
wyoung
|
856 |
online help and full online documentation. |
|
29997f8…
|
wyoung
|
857 |
|
|
29997f8…
|
wyoung
|
858 |
|
|
e5ba457…
|
wyoung
|
859 |
<h3 id="hash">2.9 Hash Algorithm: SHA-3 vs SHA-2 vs SHA-1</h3> |
|
42b8890…
|
wyoung
|
860 |
|
|
42b8890…
|
wyoung
|
861 |
Fossil started out using 160-bit SHA-1 hashes to identify check-ins, |
|
42b8890…
|
wyoung
|
862 |
just as in Git. That changed in early 2017 when news of the |
|
42b8890…
|
wyoung
|
863 |
[https://shattered.io/|SHAttered attack] broke, demonstrating that SHA-1 |
|
42b8890…
|
wyoung
|
864 |
collisions were now practical to create. Two weeks later, the creator of |
|
42b8890…
|
wyoung
|
865 |
Fossil delivered a new release allowing a clean migration to |
|
42b8890…
|
wyoung
|
866 |
[https://en.wikipedia.org/wiki/SHA-3|256-bit SHA-3] with |
|
42b8890…
|
wyoung
|
867 |
[./hashpolicy.wiki|full backwards compatibility] to old SHA-1 based |
|
42b8890…
|
wyoung
|
868 |
repositories. |
|
42b8890…
|
wyoung
|
869 |
|
|
4fdc8db…
|
wyoung
|
870 |
In October 2019, after the last of the major binary |
|
4fdc8db…
|
wyoung
|
871 |
package repos offering Fossil upgraded to Fossil 2.<i>x</i>, |
|
ad47a44…
|
wyoung
|
872 |
we switched the default hash mode so that |
|
ad47a44…
|
wyoung
|
873 |
the conversion to SHA-3 is fully automatic. |
|
4fdc8db…
|
wyoung
|
874 |
This not |
|
d887a6d…
|
wyoung
|
875 |
only solves the SHAttered problem, it should prevent a reoccurrence of |
|
d887a6d…
|
wyoung
|
876 |
similar problems for the foreseeable future. |
|
d887a6d…
|
wyoung
|
877 |
|
|
4fdc8db…
|
wyoung
|
878 |
Meanwhile, the Git community took until August 2018 to publish |
|
35fe8bf…
|
wyoung
|
879 |
[https://git-scm.com/docs/hash-function-transition/ | their plan] for |
|
35fe8bf…
|
wyoung
|
880 |
solving the same problem by moving to SHA-256, a variant of the |
|
35fe8bf…
|
wyoung
|
881 |
[https://en.wikipedia.org/wiki/SHA-2 | older SHA-2 algorithm]. That is |
|
35fe8bf…
|
wyoung
|
882 |
now technically implemented in the sense that <tt>git init |
|
35fe8bf…
|
wyoung
|
883 |
--object-format=sha256</tt> exists, but note well: this is not only an |
|
35fe8bf…
|
wyoung
|
884 |
optional setting, Git forge support is mixed, most notably |
|
35fe8bf…
|
wyoung
|
885 |
[https://github.com/GitoxideLabs/gitoxide/issues/281 | lacking in |
|
35fe8bf…
|
wyoung
|
886 |
GitHub], plus also BitBucket and others. This is doubtless because of |
|
35fe8bf…
|
wyoung
|
887 |
this warning in the latest ([https://git-scm.com/docs/git-init/2.54.0 | |
|
35fe8bf…
|
wyoung
|
888 |
as of this writing]) <tt>git init</tt> docs: |
|
35fe8bf…
|
wyoung
|
889 |
|
|
35fe8bf…
|
wyoung
|
890 |
<blockquote>Note: At present, there is no interoperability between |
|
35fe8bf…
|
wyoung
|
891 |
SHA-256 repositories and SHA-1 repositories.</blockquote> |
|
35fe8bf…
|
wyoung
|
892 |
|
|
35fe8bf…
|
wyoung
|
893 |
Although we are now in the <i>tenth year</i> of this situation, there |
|
35fe8bf…
|
wyoung
|
894 |
remains hope that Git will manage to make the transition without taking |
|
35fe8bf…
|
wyoung
|
895 |
the full decade: the |
|
35fe8bf…
|
wyoung
|
896 |
[https://www.deployhq.com/blog/git-3-0-on-the-horizon-what-git-users-need-to-know-about-the-next-major-release |
|
35fe8bf…
|
wyoung
|
897 |
| latest plan] is that Git 3.0 will finally <i>(finally!)</i> switch to |
|
35fe8bf…
|
wyoung
|
898 |
SHA256 by default, forcing the issue. Given the track record, we are |
|
35fe8bf…
|
wyoung
|
899 |
taking a "show me" stance on this claim. |
|
35fe8bf…
|
wyoung
|
900 |
|
|
35fe8bf…
|
wyoung
|
901 |
Always remember, attacks only get better, never worse. |
|
4fdc8db…
|
wyoung
|
902 |
Almost three years after Fossil solved this problem, the |
|
4fdc8db…
|
wyoung
|
903 |
[https://sha-mbles.github.io/ | SHAmbles attack] was published, further |
|
4fdc8db…
|
wyoung
|
904 |
weakening the case for continuing to use SHA-1. |
|
4fdc8db…
|
wyoung
|
905 |
|
|
4fdc8db…
|
wyoung
|
906 |
The practical impact of attacks like SHAttered and SHAmbles on the |
|
e673065…
|
wyoung
|
907 |
Git and Fossil Merkle trees isn't clear, but you want to have your repositories |
|
35fe8bf…
|
wyoung
|
908 |
moved over to a stronger hash algorithm <i>before</i> someone figures out how |
|
35fe8bf…
|
wyoung
|
909 |
to make use of the weaknesses in the old one. Fossil's solution is long |
|
35fe8bf…
|
wyoung
|
910 |
since [https://repology.org/project/fossil/versions | universally deployed]. |
|
8452627…
|
wyoung
|
911 |
|
|
8452627…
|
wyoung
|
912 |
<hr/> |
|
8452627…
|
wyoung
|
913 |
|
|
8452627…
|
wyoung
|
914 |
<h3>Asides and Digressions</h3> |
|
8452627…
|
wyoung
|
915 |
|
|
8452627…
|
wyoung
|
916 |
<i><small><ol> |
|
0ef227b…
|
wyoung
|
917 |
<li><p>[./mirrorlimitations.md|Many |
|
0ef227b…
|
wyoung
|
918 |
things are lost] in making a Git mirror of a Fossil repo due to |
|
0ef227b…
|
wyoung
|
919 |
limitations of Git relative to Fossil. GitHub adds some of these |
|
0ef227b…
|
wyoung
|
920 |
missing features to stock |
|
8452627…
|
wyoung
|
921 |
Git, but because they're not part of Git proper, |
|
8452627…
|
wyoung
|
922 |
[./mirrortogithub.md|exporting a Fossil repository to GitHub] will |
|
8452627…
|
wyoung
|
923 |
still not include them; Fossil tickets do not become GitHub issues, |
|
0ef227b…
|
wyoung
|
924 |
for example. |
|
0ef227b…
|
wyoung
|
925 |
|
|
d23a845…
|
wyoung
|
926 |
<li><p>The <tt>fossil-scm.org</tt> web site is actually hosted in |
|
d23a845…
|
wyoung
|
927 |
several parts, so that it is not strictly true that "everything" on |
|
d23a845…
|
wyoung
|
928 |
it is in the self-hosting Fossil project repo. The web forum is |
|
5866461…
|
wyoung
|
929 |
hosted as [https://fossil-scm.org/forum/|a separate Fossil repo] |
|
09908ab…
|
drh
|
930 |
from the [https://fossil-scm.org/home/|main Fossil self-hosting |
|
5866461…
|
wyoung
|
931 |
repo] for administration reasons, and the Download page content |
|
0996347…
|
wyoung
|
932 |
isn't normally synchronized with a "<tt>fossil clone</tt>" command unless |
|
5866461…
|
wyoung
|
933 |
you add the "-u" option. (See "[./aboutdownload.wiki|How the |
|
f5ab488…
|
drh
|
934 |
Download Page Works]" for details.) |
|
f5ab488…
|
drh
|
935 |
Chat history is deliberately not synced as |
|
f5ab488…
|
drh
|
936 |
chat messages are intended to be ephemeral. |
|
f5ab488…
|
drh
|
937 |
There may also be some purely |
|
5866461…
|
wyoung
|
938 |
static elements of the web site served via D. Richard Hipp's own |
|
5866461…
|
wyoung
|
939 |
lightweight web server, |
|
6ef97bd…
|
drh
|
940 |
<tt>[https://sqlite.org/althttpd/|althttpd]</tt>, |
|
5866461…
|
wyoung
|
941 |
which is configured as a front end to Fossil running in CGI mode on |
|
5866461…
|
wyoung
|
942 |
these sites. |
|
5866461…
|
wyoung
|
943 |
|
|
4b574be…
|
wyoung
|
944 |
<li><p>This means you can give up waiting for Fossil to be ported to |
|
4b574be…
|
wyoung
|
945 |
the PDP-11, but we remain hopeful that someone may eventually port |
|
4b574be…
|
wyoung
|
946 |
it to [https://en.wikipedia.org/wiki/Z/OS|z/OS]. |
|
4b574be…
|
wyoung
|
947 |
|
|
4b574be…
|
wyoung
|
948 |
<li><p>"Why is there all this Tcl in and around Fossil?" you may |
|
4b574be…
|
wyoung
|
949 |
ask. It is because D. Richard Hipp is a long-time Tcl user and |
|
4b574be…
|
wyoung
|
950 |
contributor. SQLite started out as an embedded database for Tcl |
|
4b574be…
|
wyoung
|
951 |
specifically. ([https://sqlite.org/tclsqlite.html | [Reference]]) |
|
4b574be…
|
wyoung
|
952 |
When he then created Fossil to manage the development of SQLite, it |
|
4b574be…
|
wyoung
|
953 |
was natural for him to use Tcl-based tools for its scripting, build |
|
4b574be…
|
wyoung
|
954 |
system, test system, etc. It came full circle in 2011 when |
|
4b574be…
|
wyoung
|
955 |
[https://www.reddit.com/r/programming/comments/fwrx5/tcl_and_tk_move_away_from_cvs_to_fossil/ |
|
4b574be…
|
wyoung
|
956 |
| the Tcl and Tk projects moved from CVS to Fossil]. |
|
4b574be…
|
wyoung
|
957 |
|
|
1296ee4…
|
wyoung
|
958 |
<li><p>A minority of the pieces of the Git core software suite are |
|
1296ee4…
|
wyoung
|
959 |
written in other languages, primarily Perl, Python, and Tcl. (e.g. |
|
1296ee4…
|
wyoung
|
960 |
<tt>git-send-mail</tt>, <tt>git-p4</tt>, and <tt>gitk</tt>, |
|
1296ee4…
|
wyoung
|
961 |
respectively.) Although these interpreters are quite portable, they |
|
1296ee4…
|
wyoung
|
962 |
aren't installed by default everywhere, and on some platforms you |
|
1296ee4…
|
wyoung
|
963 |
can't count on them at all. (Not just Windows, but also the BSDs and |
|
1296ee4…
|
wyoung
|
964 |
many other non-Linux platforms.) This expands the dependency |
|
1296ee4…
|
wyoung
|
965 |
footprint of Git considerably. It is why the current Git for Windows |
|
1296ee4…
|
wyoung
|
966 |
distribution is 44.7 MiB but the current <tt>fossil.exe</tt> |
|
1296ee4…
|
wyoung
|
967 |
zip file for Windows is 2.24 MiB. Fossil is much smaller |
|
1296ee4…
|
wyoung
|
968 |
despite using a roughly similar amount of high-level scripting code |
|
1296ee4…
|
wyoung
|
969 |
because its interpreters are compact and built into Fossil itself. |
|
1296ee4…
|
wyoung
|
970 |
|
|
8452627…
|
wyoung
|
971 |
<li><p>Both Fossil and Git support |
|
8452627…
|
wyoung
|
972 |
[https://en.wikipedia.org/wiki/Patch_(Unix)|<tt>patch(1)</tt> |
|
e3e06d3…
|
wyoung
|
973 |
files] — unified diff formatted output — for accepting drive-by contributions, but it's a |
|
8452627…
|
wyoung
|
974 |
lossy contribution path for both systems. Unlike Git PRs and Fossil |
|
79c2cb0…
|
wyoung
|
975 |
bundles, patch files collapse multiple checkins together, they don't |
|
8452627…
|
wyoung
|
976 |
include check-in comments, and they cannot encode changes made above |
|
79c2cb0…
|
wyoung
|
977 |
the individual file content layer: you lose branching decisions, |
|
ad47a44…
|
wyoung
|
978 |
tag changes, file renames, and more when using patch files. The |
|
ad47a44…
|
wyoung
|
979 |
[./patchcmd.md | <tt>fossil patch</tt> command] |
|
e3e06d3…
|
wyoung
|
980 |
also solves these problems, but it is because it works like a Fossil |
|
e3e06d3…
|
wyoung
|
981 |
bundle, only for uncommitted changes; it doesn't use Larry Wall's |
|
e3e06d3…
|
wyoung
|
982 |
<tt>patch</tt> tool to apply unified diff output to the receiving |
|
e3e06d3…
|
wyoung
|
983 |
Fossil checkout.</p></li> |
|
8452627…
|
wyoung
|
984 |
</ol></i></small> |