Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 78c8fcef7bc01a4a1774cc8cf95f9fc9b2e9c2da / . / docs / DeveloperPolicy.html
blob: 987b622f2edff4af13cd45982fb89acd409f857a [file] [log] [blame]
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <title>LLVM Developer Policy</title>
6 <link rel="stylesheet" href="llvm.css" type="text/css">
7</head>
8<body>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]9
10<div class="doc_title">LLVM Developer Policy</div>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]11<ol>
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]12 <li><a href="#introduction">Introduction</a></li>
Chris Lattner1ff20cd2007-02-19 06:05:58 +0000[diff] [blame]13 <li><a href="#policies">Developer Policies</a>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]14 <ol>
Chris Lattner1653fec2007-02-19 05:57:29 +0000[diff] [blame]15 <li><a href="#informed">Stay Informed</a></li>
16 <li><a href="#patches">Making a Patch</a></li>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]17 <li><a href="#reviews">Code Reviews</a></li>
Chris Lattnere268a402007-12-03 19:00:47 +0000[diff] [blame]18 <li><a href="#owners">Code Owners</a></li>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]19 <li><a href="#testcases">Test Cases</a></li>
Chris Lattner1acdc952007-02-19 05:49:11 +0000[diff] [blame]20 <li><a href="#quality">Quality</a></li>
21 <li><a href="#commitaccess">Obtaining Commit Access</a></li>
Chris Lattner1ff20cd2007-02-19 06:05:58 +0000[diff] [blame]22 <li><a href="#newwork">Making a Major Change</a></li>
23 <li><a href="#incremental">Incremental Development</a></li>
Chris Lattnerc7d954e2007-02-19 05:43:04 +0000[diff] [blame]24 <li><a href="#attribution">Attribution of Changes</a></li>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]25 </ol></li>
Chris Lattner793aa382007-02-19 06:19:16 +0000[diff] [blame]26 <li><a href="#clp">Copyright, License, and Patents</a>
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]27 <ol>
28 <li><a href="#copyright">Copyright</a></li>
29 <li><a href="#license">License</a></li>
Chris Lattner1ff20cd2007-02-19 06:05:58 +0000[diff] [blame]30 <li><a href="#patents">Patents</a></li>
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]31 <li><a href="#devagree">Developer Agreements</a></li>
32 </ol></li>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]33</ol>
Chris Lattner2ae49dd2007-02-19 06:24:23 +0000[diff] [blame]34<div class="doc_author">Written by the LLVM Oversight Team</div>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]35
36<!--=========================================================================-->
Reid Spencere7bd7d62007-02-14 17:24:04 +0000[diff] [blame]37<div class="doc_section"><a name="introduction">Introduction</a></div>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]38<!--=========================================================================-->
39<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]40<p>This document contains the LLVM Developer Policy which defines the project's
41 policy towards developers and their contributions. The intent of this policy
42 is to eliminate miscommunication, rework, and confusion that might arise from
43 the distributed nature of LLVM's development. By stating the policy in clear
44 terms, we hope each developer can know ahead of time what to expect when
45 making LLVM contributions.</p>
46<p>This policy is also designed to accomplish the following objectives:</p>
47
48<ol>
49 <li>Attract both users and developers to the LLVM project.</li>
50
51 <li>Make life as simple and easy for contributors as possible.</li>
52
53 <li>Keep the top of Subversion trees as stable as possible.</li>
54</ol>
Chris Lattnerc7d954e2007-02-19 05:43:04 +0000[diff] [blame]55
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]56<p>This policy is aimed at frequent contributors to LLVM. People interested in
57 contributing one-off patches can do so in an informal way by sending them to
58 the
59 <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits">llvm-commits
60 mailing list</a> and engaging another developer to see it through the
61 process.</p>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]62</div>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]63
64<!--=========================================================================-->
Chris Lattner1ff20cd2007-02-19 06:05:58 +0000[diff] [blame]65<div class="doc_section"><a name="policies">Developer Policies</a></div>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]66<!--=========================================================================-->
67<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]68<p>This section contains policies that pertain to frequent LLVM developers. We
69 always welcome <a href="#patches">one-off patches</a> from people who do not
70 routinely contribute to LLVM, but we expect more from frequent contributors
71 to keep the system as efficient as possible for everyone. Frequent LLVM
72 contributors are expected to meet the following requirements in order for
73 LLVM to maintain a high standard of quality.<p>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]74</div>
75
76<!-- _______________________________________________________________________ -->
77<div class="doc_subsection"> <a name="informed">Stay Informed</a> </div>
78<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]79<p>Developers should stay informed by reading at least the
80 <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">llvmdev</a> email
81 list. If you are doing anything more than just casual work on LLVM, it is
82 suggested that you also subscribe to the
83 <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits">llvm-commits</a>
Reid Spencer01f56be2007-02-14 16:21:10 +0000[diff] [blame]84 list and pay attention to changes being made by others.</p>
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]85
86<p>We recommend that active developers register an email account with
87 <a href="http://llvm.org/bugs/">LLVM Bugzilla</a> and preferably subscribe to
88 the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs">llvm-bugs</a>
89 email list to keep track of bugs and enhancements occurring in LLVM.</p>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]90</div>
91
92<!-- _______________________________________________________________________ -->
Chris Lattner1653fec2007-02-19 05:57:29 +0000[diff] [blame]93<div class="doc_subsection"> <a name="patches">Making a Patch</a></div>
94
95<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]96<p>When making a patch for review, the goal is to make it as easy for the
97 reviewer to read it as possible. As such, we recommend that you:</p>
Chris Lattner1653fec2007-02-19 05:57:29 +0000[diff] [blame]98
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]99<ol>
100 <li>Make your patch against the Subversion trunk, not a branch, and not an old
101 version of LLVM. This makes it easy to apply the patch.</li>
Chris Lattner1653fec2007-02-19 05:57:29 +0000[diff] [blame]102
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]103 <li>Similarly, patches should be submitted soon after they are generated. Old
104 patches may not apply correctly if the underlying code changes between the
105 time the patch was created and the time it is applied.</li>
106
107 <li>Patches should be made with this command:
108<div class="doc_code">
109<pre>
110svn diff -x -u
111</pre>
112</div>
113 or with the utility <tt>utils/mkpatch</tt>, which makes it easy to read
114 the diff.</li>
115
116 <li>Patches should not include differences in generated code such as the code
117 generated by <tt>autoconf</tt> or <tt>tblgen</tt>. The
118 <tt>utils/mkpatch</tt> utility takes care of this for you.</li>
119</ol>
Chris Lattner0cca50c2007-02-19 06:57:16 +0000[diff] [blame]120
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]121<p>When sending a patch to a mailing list, it is a good idea to send it as an
122 <em>attachment</em> to the message, not embedded into the text of the
123 message. This ensures that your mailer will not mangle the patch when it
124 sends it (e.g. by making whitespace changes or by wrapping lines).</p>
125
126<p><em>For Thunderbird users:</em> Before submitting a patch, please open
127 <em>Preferences &#8594; Advanced &#8594; General &#8594; Config Editor</em>,
128 find the key <tt>mail.content_disposition_type</tt>, and set its value to
129 <tt>1</tt>. Without this setting, Thunderbird sends your attachment using
130 <tt>Content-Disposition: inline</tt> rather than <tt>Content-Disposition:
131 attachment</tt>. Apple Mail gamely displays such a file inline, making it
132 difficult to work with for reviewers using that program.</p>
Chris Lattner1653fec2007-02-19 05:57:29 +0000[diff] [blame]133</div>
134
135<!-- _______________________________________________________________________ -->
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]136<div class="doc_subsection"> <a name="reviews">Code Reviews</a></div>
137<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]138<p>LLVM has a code review policy. Code review is one way to increase the quality
139 of software. We generally follow these policies:</p>
Chris Lattner0cca50c2007-02-19 06:57:16 +0000[diff] [blame]140
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]141<ol>
142 <li>All developers are required to have significant changes reviewed before
143 they are committed to the repository.</li>
144
145 <li>Code reviews are conducted by email, usually on the llvm-commits
146 list.</li>
147
148 <li>Code can be reviewed either before it is committed or after. We expect
149 major changes to be reviewed before being committed, but smaller changes
150 (or changes where the developer owns the component) can be reviewed after
151 commit.</li>
152
153 <li>The developer responsible for a code change is also responsible for making
154 all necessary review-related changes.</li>
155
156 <li>Code review can be an iterative process, which continues until the patch
157 is ready to be committed.</li>
158</ol>
159
160<p>Developers should participate in code reviews as both reviewers and
161 reviewees. If someone is kind enough to review your code, you should return
162 the favor for someone else. Note that anyone is welcome to review and give
163 feedback on a patch, but only people with Subversion write access can approve
164 it.</p>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]165</div>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]166
167<!-- _______________________________________________________________________ -->
Chris Lattnere268a402007-12-03 19:00:47 +0000[diff] [blame]168<div class="doc_subsection"> <a name="owners">Code Owners</a></div>
169<div class="doc_text">
170
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]171<p>The LLVM Project relies on two features of its process to maintain rapid
172 development in addition to the high quality of its source base: the
173 combination of code review plus post-commit review for trusted maintainers.
174 Having both is a great way for the project to take advantage of the fact that
175 most people do the right thing most of the time, and only commit patches
176 without pre-commit review when they are confident they are right.</p>
Chris Lattnere268a402007-12-03 19:00:47 +0000[diff] [blame]177
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]178<p>The trick to this is that the project has to guarantee that all patches that
179 are committed are reviewed after they go in: you don't want everyone to
180 assume someone else will review it, allowing the patch to go unreviewed. To
181 solve this problem, we have a notion of an 'owner' for a piece of the code.
182 The sole responsibility of a code owner is to ensure that a commit to their
183 area of the code is appropriately reviewed, either by themself or by someone
184 else. The current code owners are:</p>
Chris Lattnere268a402007-12-03 19:00:47 +0000[diff] [blame]185
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]186<ol>
187 <li><b>Anton Korobeynikov</b>: Exception handling, debug information, and
188 Windows codegen.</li>
189
190 <li><b>Duncan Sands</b>: llvm-gcc 4.2.</li>
191
192 <li><b>Evan Cheng</b>: Code generator and all targets.</li>
193
194 <li><b>Chris Lattner</b>: Everything else.</li>
195</ol>
Chris Lattnere268a402007-12-03 19:00:47 +0000[diff] [blame]196
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]197<p>Note that code ownership is completely different than reviewers: anyone can
198 review a piece of code, and we welcome code review from anyone who is
199 interested. Code owners are the "last line of defense" to guarantee that all
200 patches that are committed are actually reviewed.</p>
Chris Lattnere268a402007-12-03 19:00:47 +0000[diff] [blame]201
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]202<p>Being a code owner is a somewhat unglamorous position, but it is incredibly
203 important for the ongoing success of the project. Because people get busy,
204 interests change, and unexpected things happen, code ownership is purely
205 opt-in, and anyone can choose to resign their "title" at any time. For now,
206 we do not have an official policy on how one gets elected to be a code
207 owner.</p>
Chris Lattnere268a402007-12-03 19:00:47 +0000[diff] [blame]208</div>
209
Chris Lattnere268a402007-12-03 19:00:47 +0000[diff] [blame]210<!-- _______________________________________________________________________ -->
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]211<div class="doc_subsection"> <a name="testcases">Test Cases</a></div>
212<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]213<p>Developers are required to create test cases for any bugs fixed and any new
214 features added. Some tips for getting your testcase approved:</p>
215
216<ol>
217 <li>All feature and regression test cases are added to the
218 <tt>llvm/test</tt> directory. The appropriate sub-directory should be
219 selected (see the <a href="TestingGuide.html">Testing Guide</a> for
220 details).</li>
221
222 <li>Test cases should be written in <a href="LangRef.html">LLVM assembly
223 language</a> unless the feature or regression being tested requires
224 another language (e.g. the bug being fixed or feature being implemented is
225 in the llvm-gcc C++ front-end, in which case it must be written in
226 C++).</li>
227
228 <li>Test cases, especially for regressions, should be reduced as much as
229 possible, by <a href="Bugpoint.html">bugpoint</a> or manually. It is
230 unacceptable to place an entire failing program into <tt>llvm/test</tt> as
231 this creates a <i>time-to-test</i> burden on all developers. Please keep
232 them short.</li>
233</ol>
Chris Lattner0cca50c2007-02-19 06:57:16 +0000[diff] [blame]234
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]235<p>Note that llvm/test is designed for regression and small feature tests
236 only. More extensive test cases (e.g., entire applications, benchmarks, etc)
237 should be added to the <tt>llvm-test</tt> test suite. The llvm-test suite is
238 for coverage (correctness, performance, etc) testing, not feature or
239 regression testing.</p>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]240</div>
241
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]242<!-- _______________________________________________________________________ -->
Chris Lattner39582bd2007-02-19 05:47:13 +0000[diff] [blame]243<div class="doc_subsection"> <a name="quality">Quality</a></div>
244<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]245<p>The minimum quality standards that any change must satisfy before being
246 committed to the main development branch are:</p>
247
248<ol>
249 <li>Code must adhere to the <a href="CodingStandards.html">LLVM Coding
250 Standards</a>.</li>
251
252 <li>Code must compile cleanly (no errors, no warnings) on at least one
253 platform.</li>
254
255 <li>Bug fixes and new features should <a href="#testcases">include a
256 testcase</a> so we know if the fix/feature ever regresses in the
257 future.</li>
258
259 <li>Code must pass the dejagnu (<tt>llvm/test</tt>) test suite.</li>
260
261 <li>The code must not cause regressions on a reasonable subset of llvm-test,
262 where "reasonable" depends on the contributor's judgement and the scope of
263 the change (more invasive changes require more testing). A reasonable
264 subset might be something like
265 "<tt>llvm-test/MultiSource/Benchmarks</tt>".</li>
266</ol>
267
268<p>Additionally, the committer is responsible for addressing any problems found
269 in the future that the change is responsible for. For example:</p>
270
271<ul>
272 <li>The code should compile cleanly on all supported platforms.</li>
273
274 <li>The changes should not cause any correctness regressions in the
275 <tt>llvm-test</tt> suite and must not cause any major performance
276 regressions.</li>
277
278 <li>The change set should not cause performance or correctness regressions for
279 the LLVM tools.</li>
280
281 <li>The changes should not cause performance or correctness regressions in
282 code compiled by LLVM on all applicable targets.</li>
283
284 <li>You are expected to address any <a href="http://llvm.org/bugs/">bugzilla
285 bugs</a> that result from your change.</li>
286</ul>
Chris Lattner39582bd2007-02-19 05:47:13 +0000[diff] [blame]287
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]288<p>We prefer for this to be handled before submission but understand that it
289 isn't possible to test all of this for every submission. Our nightly testing
290 infrastructure normally finds these problems. A good rule of thumb is to
291 check the nightly testers for regressions the day after your change.</p>
292
293<p>Commits that violate these quality standards (e.g. are very broken) may be
294 reverted. This is necessary when the change blocks other developers from
295 making progress. The developer is welcome to re-commit the change after the
296 problem has been fixed.</p>
Chris Lattner39582bd2007-02-19 05:47:13 +0000[diff] [blame]297</div>
298
299<!-- _______________________________________________________________________ -->
Chris Lattner1acdc952007-02-19 05:49:11 +0000[diff] [blame]300<div class="doc_subsection">
301 <a name="commitaccess">Obtaining Commit Access</a></div>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]302<div class="doc_text">
303
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]304<p>We grant commit access to contributors with a track record of submitting high
305 quality patches. If you would like commit access, please send an email to
306 <a href="mailto:sabre@nondot.org">Chris</a> with the following
307 information:</p>
Chris Lattnerbebcdabd82007-12-03 00:36:20 +0000[diff] [blame]308
309<ol>
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]310 <li>The user name you want to commit with, e.g. "hacker".</li>
311
Chris Lattnerbebcdabd82007-12-03 00:36:20 +0000[diff] [blame]312 <li>The full name and email address you want message to llvm-commits to come
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]313 from, e.g. "J. Random Hacker &lt;hacker@yoyodyne.com&gt;".</li>
314
Chris Lattnerbebcdabd82007-12-03 00:36:20 +0000[diff] [blame]315 <li>A "password hash" of the password you want to use, e.g. "2ACR96qjUqsyM".
316 Note that you don't ever tell us what your password is, you just give it
317 to us in an encrypted form. To get this, run "htpasswd" (a utility that
318 comes with apache) in crypt mode (often enabled with "-d"), or find a web
319 page that will do it for you.</li>
320</ol>
321
322<p>Once you've been granted commit access, you should be able to check out an
323 LLVM tree with an SVN URL of "https://username@llvm.org/..." instead of the
324 normal anonymous URL of "http://llvm.org/...". The first time you commit
325 you'll have to type in your password. Note that you may get a warning from
326 SVN about an untrusted key, you can ignore this. To verify that your commit
327 access works, please do a test commit (e.g. change a comment or add a blank
328 line). Your first commit to a repository may require the autogenerated email
329 to be approved by a mailing list. This is normal, and will be done when
330 the mailing list owner has time.</p>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]331
332<p>If you have recently been granted commit access, these policies apply:</p>
Chris Lattnerbebcdabd82007-12-03 00:36:20 +0000[diff] [blame]333
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]334<ol>
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]335 <li>You are granted <i>commit-after-approval</i> to all parts of LLVM. To get
336 approval, submit a <a href="#patches">patch</a> to
337 <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits">llvm-commits</a>.
338 When approved you may commit it yourself.</li>
339
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]340 <li>You are allowed to commit patches without approval which you think are
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]341 obvious. This is clearly a subjective decision &mdash; we simply expect
342 you to use good judgement. Examples include: fixing build breakage,
343 reverting obviously broken patches, documentation/comment changes, any
344 other minor changes.</li>
345
346 <li>You are allowed to commit patches without approval to those portions of
347 LLVM that you have contributed or maintain (i.e., have been assigned
348 responsibility for), with the proviso that such commits must not break the
349 build. This is a "trust but verify" policy and commits of this nature are
350 reviewed after they are committed.</li>
351
352 <li>Multiple violations of these policies or a single egregious violation may
353 cause commit access to be revoked.</li>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]354</ol>
Chris Lattner0cca50c2007-02-19 06:57:16 +0000[diff] [blame]355
356<p>In any case, your changes are still subject to <a href="#reviews">code
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]357 review</a> (either before or after they are committed, depending on the
358 nature of the change). You are encouraged to review other peoples' patches
359 as well, but you aren't required to.</p>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]360</div>
361
362<!-- _______________________________________________________________________ -->
Chris Lattnerc7d954e2007-02-19 05:43:04 +0000[diff] [blame]363<div class="doc_subsection"> <a name="newwork">Making a Major Change</a></div>
364<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]365<p>When a developer begins a major new project with the aim of contributing it
366 back to LLVM, s/he should inform the community with an email to
367 the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">llvmdev</a>
368 email list, to the extent possible. The reason for this is to:
369
370<ol>
371 <li>keep the community informed about future changes to LLVM, </li>
372
373 <li>avoid duplication of effort by preventing multiple parties working on the
374 same thing and not knowing about it, and</li>
375
376 <li>ensure that any technical issues around the proposed work are discussed
377 and resolved before any significant work is done.</li>
378</ol>
Chris Lattnerc7d954e2007-02-19 05:43:04 +0000[diff] [blame]379
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]380<p>The design of LLVM is carefully controlled to ensure that all the pieces fit
381 together well and are as consistent as possible. If you plan to make a major
382 change to the way LLVM works or want to add a major new extension, it is a
383 good idea to get consensus with the development community before you start
384 working on it.</p>
Chris Lattnerc7d954e2007-02-19 05:43:04 +0000[diff] [blame]385
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]386<p>Once the design of the new feature is finalized, the work itself should be
387 done as a series of <a href="#incremental">incremental changes</a>, not as a
388 long-term development branch.</p>
Chris Lattnerc7d954e2007-02-19 05:43:04 +0000[diff] [blame]389</div>
390
391<!-- _______________________________________________________________________ -->
Chris Lattner1ff20cd2007-02-19 06:05:58 +0000[diff] [blame]392<div class="doc_subsection"> <a name="incremental">Incremental Development</a>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]393</div>
394<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]395<p>In the LLVM project, we do all significant changes as a series of incremental
396 patches. We have a strong dislike for huge changes or long-term development
397 branches. Long-term development branches have a number of drawbacks:</p>
398
399<ol>
400 <li>Branches must have mainline merged into them periodically. If the branch
401 development and mainline development occur in the same pieces of code,
402 resolving merge conflicts can take a lot of time.</li>
403
404 <li>Other people in the community tend to ignore work on branches.</li>
405
406 <li>Huge changes (produced when a branch is merged back onto mainline) are
407 extremely difficult to <a href="#reviews">code review</a>.</li>
408
409 <li>Branches are not routinely tested by our nightly tester
410 infrastructure.</li>
411
412 <li>Changes developed as monolithic large changes often don't work until the
413 entire set of changes is done. Breaking it down into a set of smaller
414 changes increases the odds that any of the work will be committed to the
415 main repository.</li>
416</ol>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]417
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]418<p>To address these problems, LLVM uses an incremental development style and we
419 require contributors to follow this practice when making a large/invasive
420 change. Some tips:</p>
421
422<ul>
423 <li>Large/invasive changes usually have a number of secondary changes that are
424 required before the big change can be made (e.g. API cleanup, etc). These
425 sorts of changes can often be done before the major change is done,
426 independently of that work.</li>
427
428 <li>The remaining inter-related work should be decomposed into unrelated sets
429 of changes if possible. Once this is done, define the first increment and
430 get consensus on what the end goal of the change is.</li>
431
432 <li>Each change in the set can be stand alone (e.g. to fix a bug), or part of
433 a planned series of changes that works towards the development goal.</li>
434
435 <li>Each change should be kept as small as possible. This simplifies your work
436 (into a logical progression), simplifies code review and reduces the
437 chance that you will get negative feedback on the change. Small increments
438 also facilitate the maintenance of a high quality code base.</li>
439
440 <li>Often, an independent precursor to a big change is to add a new API and
441 slowly migrate clients to use the new API. Each change to use the new API
442 is often "obvious" and can be committed without review. Once the new API
443 is in place and used, it is much easier to replace the underlying
444 implementation of the API. This implementation change is logically
445 separate from the API change.</li>
446</ul>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]447
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]448<p>If you are interested in making a large change, and this scares you, please
449 make sure to first <a href="#newwork">discuss the change/gather consensus</a>
450 then ask about the best way to go about making the change.</p>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]451</div>
452
453<!-- _______________________________________________________________________ -->
Chris Lattnerc7d954e2007-02-19 05:43:04 +0000[diff] [blame]454<div class="doc_subsection"><a name="attribution">Attribution of
455Changes</a></div>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]456<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]457<p>We believe in correct attribution of contributions to their contributors.
458 However, we do not want the source code to be littered with random
459 attributions "this code written by J. Random Hacker" (this is noisy and
460 distracting). In practice, the revision control system keeps a perfect
461 history of who changed what, and the CREDITS.txt file describes higher-level
462 contributions. If you commit a patch for someone else, please say "patch
463 contributed by J. Random Hacker!" in the commit message.</p>
Chris Lattnerab2b10c2007-12-29 19:56:08 +0000[diff] [blame]464
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]465<p>Overall, please do not add contributor names to the source code.</p>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]466</div>
467
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]468<!--=========================================================================-->
Chris Lattner793aa382007-02-19 06:19:16 +0000[diff] [blame]469<div class="doc_section">
470 <a name="clp">Copyright, License, and Patents</a>
471</div>
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]472<!--=========================================================================-->
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]473
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]474<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]475<p>This section addresses the issues of copyright, license and patents for the
476 LLVM project. Currently, the University of Illinois is the LLVM copyright
477 holder and the terms of its license to LLVM users and developers is the
478 <a href="http://www.opensource.org/licenses/UoI-NCSA.php">University of
479 Illinois/NCSA Open Source License</a>.</p>
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]480
481<div class="doc_notes">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]482<p style="text-align:center;font-weight:bold">NOTE: This section deals with
483 legal matters but does not provide legal advice. We are not lawyers, please
484 seek legal counsel from an attorney.</p>
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]485</div>
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]486</div>
487
488<!-- _______________________________________________________________________ -->
489<div class="doc_subsection"><a name="copyright">Copyright</a></div>
490<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]491<p>For consistency and ease of management, the project requires the copyright
492 for all LLVM software to be held by a single copyright holder: the University
493 of Illinois (UIUC).</p>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]494
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]495<p>Although UIUC may eventually reassign the copyright of the software to
496 another entity (e.g. a dedicated non-profit "LLVM Organization") the intent
497 for the project is to always have a single entity hold the copyrights to LLVM
498 at any given time.</p>
Chris Lattner305915b2008-05-20 20:06:53 +0000[diff] [blame]499
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]500<p>We believe that having a single copyright holder is in the best interests of
501 all developers and users as it greatly reduces the managerial burden for any
502 kind of administrative or technical decisions about LLVM. The goal of the
503 LLVM project is to always keep the code open and <a href="#license">licensed
504 under a very liberal license</a>.</p>
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]505</div>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]506
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]507<!-- _______________________________________________________________________ -->
508<div class="doc_subsection"><a name="license">License</a></div>
509<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]510<p>We intend to keep LLVM perpetually open source and to use a liberal open
511 source license. The current license is the
512 <a href="http://www.opensource.org/licenses/UoI-NCSA.php">University of
513 llinois/NCSA Open Source License</a>, which boils down to this:</p>
Chris Lattner450f40a2007-02-19 03:50:31 +0000[diff] [blame]514
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]515<ul>
516 <li>You can freely distribute LLVM.</li>
517
518 <li>You must retain the copyright notice if you redistribute LLVM.</li>
519
520 <li>Binaries derived from LLVM must reproduce the copyright notice (e.g. in
521 an included readme file).</li>
522
523 <li>You can't use our names to promote your LLVM derived products.</li>
524
525 <li>There's no warranty on LLVM at all.</li>
526</ul>
527
528<p>We believe this fosters the widest adoption of LLVM because it <b>allows
529 commercial products to be derived from LLVM</b> with few restrictions and
530 without a requirement for making any derived works also open source (i.e.
531 LLVM's license is not a "copyleft" license like the GPL). We suggest that you
532 read the <a href="http://www.opensource.org/licenses/UoI-NCSA.php">License</a>
533 if further clarification is needed.</p>
534
535<p>Note that the LLVM Project does distribute llvm-gcc, <b>which is GPL.</b>
536 This means that anything "linked" into llvm-gcc must itself be compatible
537 with the GPL, and must be releasable under the terms of the GPL. This
538 implies that <b>any code linked into llvm-gcc and distributed to others may
539 be subject to the viral aspects of the GPL</b> (for example, a proprietary
540 code generator linked into llvm-gcc must be made available under the GPL).
541 This is not a problem for code already distributed under a more liberal
542 license (like the UIUC license), and does not affect code generated by
543 llvm-gcc. It may be a problem if you intend to base commercial development
544 on llvm-gcc without redistributing your source code.</p>
545
546<p>We have no plans to change the license of LLVM. If you have questions or
547 comments about the license, please contact the
548 <a href="mailto:llvm-oversight@cs.uiuc.edu">LLVM Oversight Group</a>.</p>
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]549</div>
Chris Lattner1ff20cd2007-02-19 06:05:58 +0000[diff] [blame]550
551<!-- _______________________________________________________________________ -->
552<div class="doc_subsection"><a name="patents">Patents</a></div>
553<div class="doc_text">
Chris Lattnerb87fb2a2007-02-19 06:13:50 +0000[diff] [blame]554<p>To the best of our knowledge, LLVM does not infringe on any patents (we have
555 actually removed code from LLVM in the past that was found to infringe).
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]556 Having code in LLVM that infringes on patents would violate an important goal
557 of the project by making it hard or impossible to reuse the code for
Chris Lattnerd0742ce2007-02-19 06:15:33 +0000[diff] [blame]558 arbitrary purposes (including commercial use).</p>
Chris Lattnerb87fb2a2007-02-19 06:13:50 +0000[diff] [blame]559
560<p>When contributing code, we expect contributors to notify us of any potential
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]561 for patent-related trouble with their changes. If you or your employer own
562 the rights to a patent and would like to contribute code to LLVM that relies
563 on it, we require that the copyright owner sign an agreement that allows any
564 other user of LLVM to freely use your patent. Please contact
565 the <a href="mailto:llvm-oversight@cs.uiuc.edu">oversight group</a> for more
Chris Lattnerb87fb2a2007-02-19 06:13:50 +0000[diff] [blame]566 details.</p>
Chris Lattner1ff20cd2007-02-19 06:05:58 +0000[diff] [blame]567</div>
568
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]569<!-- _______________________________________________________________________ -->
570<div class="doc_subsection"><a name="devagree">Developer Agreements</a></div>
571<div class="doc_text">
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]572<p>With regards to the LLVM copyright and licensing, developers agree to assign
573 their copyrights to UIUC for any contribution made so that the entire
574 software base can be managed by a single copyright holder. This implies that
575 any contributions can be licensed under the license that the project
576 uses.</p>
Chris Lattner64a4c112008-05-22 03:06:14 +0000[diff] [blame]577
Bill Wendling78c8fce2009-04-05 12:37:44 +0000[diff] [blame^]578<p>When contributing code, you also affirm that you are legally entitled to
579 grant this copyright, personally or on behalf of your employer. If the code
580 belongs to some other entity, please raise this issue with the oversight
581 group before the code is committed.</p>
Reid Spencer78bade22007-02-14 07:22:19 +0000[diff] [blame]582</div>
583
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]584<!-- *********************************************************************** -->
585<hr>
586<address>
587 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]588 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]589 <a href="http://validator.w3.org/check/referer"><img
Misha Brukmanf00ddb02008-12-11 18:23:24 +0000[diff] [blame]590 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Chris Lattnerd0742ce2007-02-19 06:15:33 +0000[diff] [blame]591 Written by the
Reid Spencer3eedbd32007-02-14 07:57:48 +0000[diff] [blame]592 <a href="mailto:llvm-oversight@cs.uiuc.edu">LLVM Oversight Group</a><br>
Reid Spencer8d0ac692007-02-13 09:06:01 +0000[diff] [blame]593 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
594 Last modified: $Date$
595</address>
596</body>
597</html>