| blob | ef7ba39397c69b4f84a49f30120cb568e5ee88a2 |
3 <html>
4 <head>
8 </head>
9 <body>
12 <ol>
15 <ol>
26 </ol></li>
28 <ol>
32 </ol></li>
33 </ol>
36 <!--=========================================================================-->
38 <!--=========================================================================-->
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. This policy covers all llvm.org subprojects,
46 including Clang, LLDB, etc.</p>
49 <ol>
55 </ol>
57 <p>This policy is aimed at frequent contributors to LLVM. People interested in
58 contributing one-off patches can do so in an informal way by sending them to
59 the
61 mailing list</a> and engaging another developer to see it through the
62 process.</p>
63 </div>
65 <!--=========================================================================-->
67 <!--=========================================================================-->
69 <p>This section contains policies that pertain to frequent LLVM developers. We
71 routinely contribute to LLVM, but we expect more from frequent contributors
72 to keep the system as efficient as possible for everyone. Frequent LLVM
73 contributors are expected to meet the following requirements in order for
74 LLVM to maintain a high standard of quality.<p>
75 </div>
77 <!-- _______________________________________________________________________ -->
81 for the projects you are interested in, such as
84 for Clang, or <a
86 for LLDB. If you are doing anything more than just casual work on LLVM, it
87 is suggested that you also subscribe to the "commits" mailing list for the
88 subproject you're interested in, such as
92 Reading the "commits" list and paying attention to changes being made by
93 others is a good way to see what other people are interested in and watching
94 the flow of the project as a whole.</p>
96 <p>We recommend that active developers register an email account with
99 email list to keep track of bugs and enhancements occurring in LLVM. We
100 really appreciate people who are proactive at catching incoming bugs in their
101 components and dealing with them promptly.</p>
102 </div>
104 <!-- _______________________________________________________________________ -->
108 <p>When making a patch for review, the goal is to make it as easy for the
109 reviewer to read it as possible. As such, we recommend that you:</p>
111 <ol>
112 <li>Make your patch against the Subversion trunk, not a branch, and not an old
113 version of LLVM. This makes it easy to apply the patch. For information
114 on how to check out SVN trunk, please see the <a
117 <li>Similarly, patches should be submitted soon after they are generated. Old
118 patches may not apply correctly if the underlying code changes between the
119 time the patch was created and the time it is applied.</li>
123 that it doesn't contain clutter which makes it hard to read.</li>
125 <li>If you are modifying generated files, such as the top-level
127 a separate patch from the rest of your changes.</li>
128 </ol>
130 <p>When sending a patch to a mailing list, it is a good idea to send it as an
132 message. This ensures that your mailer will not mangle the patch when it
133 sends it (e.g. by making whitespace changes or by wrapping lines).</p>
140 attachment</tt>. Apple Mail gamely displays such a file inline, making it
141 difficult to work with for reviewers using that program.</p>
142 </div>
144 <!-- _______________________________________________________________________ -->
147 <p>LLVM has a code review policy. Code review is one way to increase the quality
148 of software. We generally follow these policies:</p>
150 <ol>
151 <li>All developers are required to have significant changes reviewed before
152 they are committed to the repository.</li>
154 <li>Code reviews are conducted by email, usually on the llvm-commits
155 list.</li>
157 <li>Code can be reviewed either before it is committed or after. We expect
158 major changes to be reviewed before being committed, but smaller changes
159 (or changes where the developer owns the component) can be reviewed after
160 commit.</li>
162 <li>The developer responsible for a code change is also responsible for making
163 all necessary review-related changes.</li>
165 <li>Code review can be an iterative process, which continues until the patch
166 is ready to be committed.</li>
167 </ol>
169 <p>Developers should participate in code reviews as both reviewers and
170 reviewees. If someone is kind enough to review your code, you should return
171 the favor for someone else. Note that anyone is welcome to review and give
172 feedback on a patch, but only people with Subversion write access can approve
173 it.</p>
174 </div>
176 <!-- _______________________________________________________________________ -->
180 <p>The LLVM Project relies on two features of its process to maintain rapid
181 development in addition to the high quality of its source base: the
182 combination of code review plus post-commit review for trusted maintainers.
183 Having both is a great way for the project to take advantage of the fact that
184 most people do the right thing most of the time, and only commit patches
185 without pre-commit review when they are confident they are right.</p>
187 <p>The trick to this is that the project has to guarantee that all patches that
188 are committed are reviewed after they go in: you don't want everyone to
189 assume someone else will review it, allowing the patch to go unreviewed. To
190 solve this problem, we have a notion of an 'owner' for a piece of the code.
191 The sole responsibility of a code owner is to ensure that a commit to their
192 area of the code is appropriately reviewed, either by themself or by someone
193 else. The current code owners are:</p>
195 <ol>
205 Windows codegen.</li>
212 </ol>
214 <p>Note that code ownership is completely different than reviewers: anyone can
215 review a piece of code, and we welcome code review from anyone who is
216 interested. Code owners are the "last line of defense" to guarantee that all
217 patches that are committed are actually reviewed.</p>
219 <p>Being a code owner is a somewhat unglamorous position, but it is incredibly
220 important for the ongoing success of the project. Because people get busy,
221 interests change, and unexpected things happen, code ownership is purely
222 opt-in, and anyone can choose to resign their "title" at any time. For now,
223 we do not have an official policy on how one gets elected to be a code
224 owner.</p>
225 </div>
227 <!-- _______________________________________________________________________ -->
230 <p>Developers are required to create test cases for any bugs fixed and any new
231 features added. Some tips for getting your testcase approved:</p>
233 <ol>
234 <li>All feature and regression test cases are added to the
237 details).</li>
240 language</a> unless the feature or regression being tested requires
241 another language (e.g. the bug being fixed or feature being implemented is
242 in the llvm-gcc C++ front-end, in which case it must be written in
243 C++).</li>
245 <li>Test cases, especially for regressions, should be reduced as much as
249 them short.</li>
250 </ol>
252 <p>Note that llvm/test and clang/test are designed for regression and small
253 feature tests only. More extensive test cases (e.g., entire applications,
254 benchmarks, etc)
256 for coverage (correctness, performance, etc) testing, not feature or
257 regression testing.</p>
258 </div>
260 <!-- _______________________________________________________________________ -->
263 <p>The minimum quality standards that any change must satisfy before being
264 committed to the main development branch are:</p>
266 <ol>
270 <li>Code must compile cleanly (no errors, no warnings) on at least one
271 platform.</li>
274 testcase</a> so we know if the fix/feature ever regresses in the
275 future.</li>
279 <li>The code must not cause regressions on a reasonable subset of llvm-test,
280 where "reasonable" depends on the contributor's judgement and the scope of
281 the change (more invasive changes require more testing). A reasonable
282 subset might be something like
284 </ol>
286 <p>Additionally, the committer is responsible for addressing any problems found
287 in the future that the change is responsible for. For example:</p>
289 <ul>
292 <li>The changes should not cause any correctness regressions in the
294 regressions.</li>
296 <li>The change set should not cause performance or correctness regressions for
297 the LLVM tools.</li>
299 <li>The changes should not cause performance or correctness regressions in
300 code compiled by LLVM on all applicable targets.</li>
304 </ul>
306 <p>We prefer for this to be handled before submission but understand that it
307 isn't possible to test all of this for every submission. Our build bots and
308 nightly testing infrastructure normally finds these problems. A good rule of
309 thumb is to check the nightly testers for regressions the day after your
310 change. Build bots will directly email you if a group of commits that
311 included yours caused a failure. You are expected to check the build bot
312 messages to see if they are your fault and, if so, fix the breakage.</p>
314 <p>Commits that violate these quality standards (e.g. are very broken) may be
315 reverted. This is necessary when the change blocks other developers from
316 making progress. The developer is welcome to re-commit the change after the
317 problem has been fixed.</p>
318 </div>
320 <!-- _______________________________________________________________________ -->
325 <p>We grant commit access to contributors with a track record of submitting high
326 quality patches. If you would like commit access, please send an email to
328 information:</p>
330 <ol>
333 <li>The full name and email address you want message to llvm-commits to come
337 Note that you don't ever tell us what your password is, you just give it
338 to us in an encrypted form. To get this, run "htpasswd" (a utility that
339 comes with apache) in crypt mode (often enabled with "-d"), or find a web
340 page that will do it for you.</li>
341 </ol>
343 <p>Once you've been granted commit access, you should be able to check out an
344 LLVM tree with an SVN URL of "https://username@llvm.org/..." instead of the
345 normal anonymous URL of "http://llvm.org/...". The first time you commit
346 you'll have to type in your password. Note that you may get a warning from
347 SVN about an untrusted key, you can ignore this. To verify that your commit
348 access works, please do a test commit (e.g. change a comment or add a blank
349 line). Your first commit to a repository may require the autogenerated email
350 to be approved by a mailing list. This is normal, and will be done when
351 the mailing list owner has time.</p>
355 <ol>
359 When approved you may commit it yourself.</li>
361 <li>You are allowed to commit patches without approval which you think are
362 obvious. This is clearly a subjective decision — we simply expect
363 you to use good judgement. Examples include: fixing build breakage,
364 reverting obviously broken patches, documentation/comment changes, any
365 other minor changes.</li>
367 <li>You are allowed to commit patches without approval to those portions of
368 LLVM that you have contributed or maintain (i.e., have been assigned
369 responsibility for), with the proviso that such commits must not break the
370 build. This is a "trust but verify" policy and commits of this nature are
371 reviewed after they are committed.</li>
373 <li>Multiple violations of these policies or a single egregious violation may
374 cause commit access to be revoked.</li>
375 </ol>
378 review</a> (either before or after they are committed, depending on the
379 nature of the change). You are encouraged to review other peoples' patches
380 as well, but you aren't required to.</p>
381 </div>
383 <!-- _______________________________________________________________________ -->
386 <p>When a developer begins a major new project with the aim of contributing it
387 back to LLVM, s/he should inform the community with an email to
389 email list, to the extent possible. The reason for this is to:
391 <ol>
394 <li>avoid duplication of effort by preventing multiple parties working on the
395 same thing and not knowing about it, and</li>
397 <li>ensure that any technical issues around the proposed work are discussed
398 and resolved before any significant work is done.</li>
399 </ol>
401 <p>The design of LLVM is carefully controlled to ensure that all the pieces fit
402 together well and are as consistent as possible. If you plan to make a major
403 change to the way LLVM works or want to add a major new extension, it is a
404 good idea to get consensus with the development community before you start
405 working on it.</p>
407 <p>Once the design of the new feature is finalized, the work itself should be
409 long-term development branch.</p>
410 </div>
412 <!-- _______________________________________________________________________ -->
414 </div>
416 <p>In the LLVM project, we do all significant changes as a series of incremental
417 patches. We have a strong dislike for huge changes or long-term development
418 branches. Long-term development branches have a number of drawbacks:</p>
420 <ol>
421 <li>Branches must have mainline merged into them periodically. If the branch
422 development and mainline development occur in the same pieces of code,
423 resolving merge conflicts can take a lot of time.</li>
427 <li>Huge changes (produced when a branch is merged back onto mainline) are
430 <li>Branches are not routinely tested by our nightly tester
431 infrastructure.</li>
433 <li>Changes developed as monolithic large changes often don't work until the
434 entire set of changes is done. Breaking it down into a set of smaller
435 changes increases the odds that any of the work will be committed to the
436 main repository.</li>
437 </ol>
439 <p>To address these problems, LLVM uses an incremental development style and we
440 require contributors to follow this practice when making a large/invasive
441 change. Some tips:</p>
443 <ul>
444 <li>Large/invasive changes usually have a number of secondary changes that are
445 required before the big change can be made (e.g. API cleanup, etc). These
446 sorts of changes can often be done before the major change is done,
447 independently of that work.</li>
449 <li>The remaining inter-related work should be decomposed into unrelated sets
450 of changes if possible. Once this is done, define the first increment and
451 get consensus on what the end goal of the change is.</li>
453 <li>Each change in the set can be stand alone (e.g. to fix a bug), or part of
454 a planned series of changes that works towards the development goal.</li>
456 <li>Each change should be kept as small as possible. This simplifies your work
457 (into a logical progression), simplifies code review and reduces the
458 chance that you will get negative feedback on the change. Small increments
459 also facilitate the maintenance of a high quality code base.</li>
461 <li>Often, an independent precursor to a big change is to add a new API and
462 slowly migrate clients to use the new API. Each change to use the new API
463 is often "obvious" and can be committed without review. Once the new API
464 is in place and used, it is much easier to replace the underlying
465 implementation of the API. This implementation change is logically
466 separate from the API change.</li>
467 </ul>
469 <p>If you are interested in making a large change, and this scares you, please
471 then ask about the best way to go about making the change.</p>
472 </div>
474 <!-- _______________________________________________________________________ -->
476 Changes</a></div>
478 <p>We believe in correct attribution of contributions to their contributors.
479 However, we do not want the source code to be littered with random
480 attributions "this code written by J. Random Hacker" (this is noisy and
481 distracting). In practice, the revision control system keeps a perfect
482 history of who changed what, and the CREDITS.txt file describes higher-level
483 contributions. If you commit a patch for someone else, please say "patch
487 </div>
489 <!--=========================================================================-->
492 </div>
493 <!--=========================================================================-->
496 <p>This section addresses the issues of copyright, license and patents for the
497 LLVM project. Currently, the University of Illinois is the LLVM copyright
498 holder and the terms of its license to LLVM users and developers is the
504 legal matters but does not provide legal advice. We are not lawyers, please
505 seek legal counsel from an attorney.</p>
506 </div>
507 </div>
509 <!-- _______________________________________________________________________ -->
513 <p>The LLVM project does not require copyright assignments, which means that the
514 copyright for the code in the project is held by its respective contributors
515 who have each agreed to release their contributed code under the terms of the
518 <p>An implication of this is that the LLVM license is unlikely to ever change:
519 changing it would require tracking down all the contributors to LLVM and
520 getting them to agree that a license change is acceptable for their
521 contribution. Since there are no plans to change the license, this is not a
522 cause for concern.</p>
524 <p>As a contributor to the project, this means that you (or your company) retain
525 ownership of the code you contribute, that it cannot be used in a way that
526 contradicts the license (which is a liberal BSD-style license), and that the
527 license for your contributions won't change without your approval in the
528 future.</p>
530 </div>
532 <!-- _______________________________________________________________________ -->
535 <p>We intend to keep LLVM perpetually open source and to use a liberal open
536 source license. All of the code in LLVM is available under the
540 <ul>
543 <li>Binaries derived from LLVM must reproduce the copyright notice (e.g. in an
544 included readme file).</li>
547 </ul>
550 commercial products to be derived from LLVM</b> with few restrictions and
551 without a requirement for making any derived works also open source (i.e.
552 LLVM's license is not a "copyleft" license like the GPL). We suggest that you
554 if further clarification is needed.</p>
556 <p>In addition to the UIUC license, the runtime library components of LLVM
559 which does not contain the binary redistribution clause. As a user of these
560 runtime libraries, it means that you can choose to use the code under either
561 license (and thus don't need the binary redistribution clause), and as a
562 contributor to the code that you agree that any contributions to these
563 libraries be licensed under both licenses. We feel that this is important
564 for runtime libraries, because they are implicitly linked into applications
565 and therefore should not subject those applications to the binary
566 redistribution clause. This also means that it is ok to move code from (e.g.)
567 libc++ to the LLVM core without concern, but that code cannot be moved from
568 the LLVM core to libc++ without the copyright owner's permission.
569 </p>
572 This means that anything "linked" into llvm-gcc must itself be compatible
573 with the GPL, and must be releasable under the terms of the GPL. This
574 implies that <b>any code linked into llvm-gcc and distributed to others may
575 be subject to the viral aspects of the GPL</b> (for example, a proprietary
576 code generator linked into llvm-gcc must be made available under the GPL).
577 This is not a problem for code already distributed under a more liberal
578 license (like the UIUC license), and does not affect code generated by
579 llvm-gcc. It may be a problem if you intend to base commercial development
580 on llvm-gcc without redistributing your source code.</p>
582 <p>We have no plans to change the license of LLVM. If you have questions or
583 comments about the license, please contact the
585 </div>
587 <!-- _______________________________________________________________________ -->
590 <p>To the best of our knowledge, LLVM does not infringe on any patents (we have
591 actually removed code from LLVM in the past that was found to infringe).
592 Having code in LLVM that infringes on patents would violate an important goal
593 of the project by making it hard or impossible to reuse the code for
594 arbitrary purposes (including commercial use).</p>
596 <p>When contributing code, we expect contributors to notify us of any potential
597 for patent-related trouble with their changes. If you or your employer own
598 the rights to a patent and would like to contribute code to LLVM that relies
599 on it, we require that the copyright owner sign an agreement that allows any
600 other user of LLVM to freely use your patent. Please contact
602 details.</p>
603 </div>
605 <!-- *********************************************************************** -->
606 <hr>
607 <address>
612 Written by the
615 Last modified: $Date$
616 </address>
617 </body>
618 </html>