Obsessively format this document.

llvm-svn: 68439
This commit is contained in:
Bill Wendling 2009-04-05 12:37:44 +00:00
parent b5fe65091d
commit b95325c543
1 changed files with 394 additions and 370 deletions

View File

@ -37,50 +37,53 @@
<div class="doc_section"><a name="introduction">Introduction</a></div> <div class="doc_section"><a name="introduction">Introduction</a></div>
<!--=========================================================================--> <!--=========================================================================-->
<div class="doc_text"> <div class="doc_text">
<p>This document contains the LLVM Developer Policy which defines the <p>This document contains the LLVM Developer Policy which defines the project's
project's policy towards developers and their contributions. The intent of policy towards developers and their contributions. The intent of this policy
this policy is to eliminate mis-communication, rework, and confusion that is to eliminate miscommunication, rework, and confusion that might arise from
might arise from the distributed nature of LLVM's development. By stating the distributed nature of LLVM's development. By stating the policy in clear
the policy in clear terms, we hope each developer can know ahead of time terms, we hope each developer can know ahead of time what to expect when
what to expect when making LLVM contributions.</p> making LLVM contributions.</p>
<p>This policy is also designed to accomplish the following objectives:</p> <p>This policy is also designed to accomplish the following objectives:</p>
<ol>
<ol>
<li>Attract both users and developers to the LLVM project.</li> <li>Attract both users and developers to the LLVM project.</li>
<li>Make life as simple and easy for contributors as possible.</li> <li>Make life as simple and easy for contributors as possible.</li>
<li>Keep the top of Subversion trees as stable as possible.</li> <li>Keep the top of Subversion trees as stable as possible.</li>
</ol> </ol>
<p>This policy is aimed at frequent contributors to LLVM. People interested in <p>This policy is aimed at frequent contributors to LLVM. People interested in
contributing one-off patches can do so in an informal way by sending them to contributing one-off patches can do so in an informal way by sending them to
the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits"> the
llvm-commits mailing list</a> and engaging another developer to see it through <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits">llvm-commits
the process.</p> mailing list</a> and engaging another developer to see it through the
process.</p>
</div> </div>
<!--=========================================================================--> <!--=========================================================================-->
<div class="doc_section"><a name="policies">Developer Policies</a></div> <div class="doc_section"><a name="policies">Developer Policies</a></div>
<!--=========================================================================--> <!--=========================================================================-->
<div class="doc_text"> <div class="doc_text">
<p>This section contains policies that pertain to frequent LLVM <p>This section contains policies that pertain to frequent LLVM developers. We
developers. We always welcome <a href="#patches">one-off patches</a> from always welcome <a href="#patches">one-off patches</a> from people who do not
people who do not routinely contribute to LLVM, but we expect more from routinely contribute to LLVM, but we expect more from frequent contributors
frequent contributors to keep the system as efficient as possible for to keep the system as efficient as possible for everyone. Frequent LLVM
everyone. contributors are expected to meet the following requirements in order for
Frequent LLVM contributors are expected to meet the following requirements in LLVM to maintain a high standard of quality.<p>
order for LLVM to maintain a high standard of quality.<p>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"> <a name="informed">Stay Informed</a> </div> <div class="doc_subsection"> <a name="informed">Stay Informed</a> </div>
<div class="doc_text"> <div class="doc_text">
<p>Developers should stay informed by reading at least the <p>Developers should stay informed by reading at least the
<a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">llvmdev</a> <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">llvmdev</a> email
email list. If you are doing anything more than just casual work on LLVM, list. If you are doing anything more than just casual work on LLVM, it is
it is suggested that you also subscribe to the suggested that you also subscribe to the
<a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits">llvm-commits</a> <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits">llvm-commits</a>
list and pay attention to changes being made by others.</p> list and pay attention to changes being made by others.</p>
<p>We recommend that active developers register an email account with
<p>We recommend that active developers register an email account with
<a href="http://llvm.org/bugs/">LLVM Bugzilla</a> and preferably subscribe to <a href="http://llvm.org/bugs/">LLVM Bugzilla</a> and preferably subscribe to
the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs">llvm-bugs</a> the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs">llvm-bugs</a>
email list to keep track of bugs and enhancements occurring in LLVM.</p> email list to keep track of bugs and enhancements occurring in LLVM.</p>
@ -90,33 +93,37 @@
<div class="doc_subsection"> <a name="patches">Making a Patch</a></div> <div class="doc_subsection"> <a name="patches">Making a Patch</a></div>
<div class="doc_text"> <div class="doc_text">
<p>When making a patch for review, the goal is to make it as easy for the
<p>When making a patch for review, the goal is to make it as easy for the
reviewer to read it as possible. As such, we recommend that you:</p> reviewer to read it as possible. As such, we recommend that you:</p>
<ol>
<li>Make your patch against the Subversion trunk, not a branch, and not an
old version of LLVM. This makes it easy to apply the patch.</li>
<li>Similarly, patches should be submitted soon after they are generated. <ol>
Old patches may not apply correctly if the underlying code changes between <li>Make your patch against the Subversion trunk, not a branch, and not an old
the time the patch was created and the time it is applied.</li> version of LLVM. This makes it easy to apply the patch.</li>
<li>Similarly, patches should be submitted soon after they are generated. Old
patches may not apply correctly if the underlying code changes between the
time the patch was created and the time it is applied.</li>
<li>Patches should be made with this command: <li>Patches should be made with this command:
<div class="doc_code"><pre>svn diff -x -u</pre></div> <div class="doc_code">
or with the utility <tt>utils/mkpatch</tt>, which makes it easy to read the <pre>
diff.</li> svn diff -x -u
</pre>
</div>
or with the utility <tt>utils/mkpatch</tt>, which makes it easy to read
the diff.</li>
<li>Patches should not include differences in generated code such as the <li>Patches should not include differences in generated code such as the code
code generated by <tt>autoconf</tt> or <tt>tblgen</tt>. The generated by <tt>autoconf</tt> or <tt>tblgen</tt>. The
<tt>utils/mkpatch</tt> utility takes care of this for you.</li> <tt>utils/mkpatch</tt> utility takes care of this for you.</li>
</ol> </ol>
<p>When sending a patch to a mailing list, it is a good idea to send it as an <p>When sending a patch to a mailing list, it is a good idea to send it as an
<em>attachment</em> to the message, not embedded into the text of the <em>attachment</em> to the message, not embedded into the text of the
message. This ensures that your mailer will not mangle the patch when it message. This ensures that your mailer will not mangle the patch when it
sends it (e.g. by making whitespace changes or by wrapping lines).</p> sends it (e.g. by making whitespace changes or by wrapping lines).</p>
<p><em>For Thunderbird users:</em> Before submitting a patch, please open <p><em>For Thunderbird users:</em> Before submitting a patch, please open
<em>Preferences &#8594; Advanced &#8594; General &#8594; Config Editor</em>, <em>Preferences &#8594; Advanced &#8594; General &#8594; Config Editor</em>,
find the key <tt>mail.content_disposition_type</tt>, and set its value to find the key <tt>mail.content_disposition_type</tt>, and set its value to
<tt>1</tt>. Without this setting, Thunderbird sends your attachment using <tt>1</tt>. Without this setting, Thunderbird sends your attachment using
@ -128,149 +135,165 @@
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"> <a name="reviews">Code Reviews</a></div> <div class="doc_subsection"> <a name="reviews">Code Reviews</a></div>
<div class="doc_text"> <div class="doc_text">
<p>LLVM has a code review policy. Code review is one way to increase the <p>LLVM has a code review policy. Code review is one way to increase the quality
quality of software. We generally follow these policies:</p> of software. We generally follow these policies:</p>
<ol>
<li>All developers are required to have significant changes reviewed <ol>
before they are committed to the repository.</li> <li>All developers are required to have significant changes reviewed before
they are committed to the repository.</li>
<li>Code reviews are conducted by email, usually on the llvm-commits <li>Code reviews are conducted by email, usually on the llvm-commits
list.</li> list.</li>
<li>Code can be reviewed either before it is committed or after. We expect <li>Code can be reviewed either before it is committed or after. We expect
major changes to be reviewed before being committed, but smaller major changes to be reviewed before being committed, but smaller changes
changes (or changes where the developer owns the component) can be (or changes where the developer owns the component) can be reviewed after
reviewed after commit.</li> commit.</li>
<li>The developer responsible for a code change is also responsible for
making all necessary review-related changes.</li> <li>The developer responsible for a code change is also responsible for making
all necessary review-related changes.</li>
<li>Code review can be an iterative process, which continues until the patch <li>Code review can be an iterative process, which continues until the patch
is ready to be committed.</li> is ready to be committed.</li>
</ol> </ol>
<p>Developers should participate in code reviews as both reviewers and
reviewees. If someone is kind enough to review your code, you should
return the favor for someone else. Note that anyone is welcome to review
and give feedback on a patch, but only people with Subversion write access
can approve it.</p>
<p>Developers should participate in code reviews as both reviewers and
reviewees. If someone is kind enough to review your code, you should return
the favor for someone else. Note that anyone is welcome to review and give
feedback on a patch, but only people with Subversion write access can approve
it.</p>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"> <a name="owners">Code Owners</a></div> <div class="doc_subsection"> <a name="owners">Code Owners</a></div>
<div class="doc_text"> <div class="doc_text">
<p>The LLVM Project relies on two features of its process to maintain rapid <p>The LLVM Project relies on two features of its process to maintain rapid
development in addition to the high quality of its source base: the development in addition to the high quality of its source base: the
combination of code review plus post-commit review for trusted maintainers. combination of code review plus post-commit review for trusted maintainers.
Having both is a great way for the project to take advantage of the fact Having both is a great way for the project to take advantage of the fact that
that most people do the right thing most of the time, and only commit most people do the right thing most of the time, and only commit patches
patches without pre-commit review when they are confident they are without pre-commit review when they are confident they are right.</p>
right.</p>
<p>The trick to this is that the project has to guarantee that all patches <p>The trick to this is that the project has to guarantee that all patches that
that are committed are reviewed after they go in: you don't want everyone are committed are reviewed after they go in: you don't want everyone to
to assume someone else will review it, allowing the patch to go unreviewed. assume someone else will review it, allowing the patch to go unreviewed. To
To solve this problem, we have a notion of an 'owner' for a piece of the solve this problem, we have a notion of an 'owner' for a piece of the code.
code. The sole responsibility of a code owner is to ensure that a commit The sole responsibility of a code owner is to ensure that a commit to their
to their area of the code is appropriately reviewed, either by themself or area of the code is appropriately reviewed, either by themself or by someone
by someone else. The current code owners are:</p> else. The current code owners are:</p>
<ol> <ol>
<li><b>Anton Korobeynikov</b>: Exception handling, debug information, and <li><b>Anton Korobeynikov</b>: Exception handling, debug information, and
Windows codegen.</li> Windows codegen.</li>
<li><b>Duncan Sands</b>: llvm-gcc 4.2.</li> <li><b>Duncan Sands</b>: llvm-gcc 4.2.</li>
<li><b>Evan Cheng</b>: Code generator and all targets.</li> <li><b>Evan Cheng</b>: Code generator and all targets.</li>
<li><b>Chris Lattner</b>: Everything else.</li> <li><b>Chris Lattner</b>: Everything else.</li>
</ol> </ol>
<p>Note that code ownership is completely different than reviewers: anyone can <p>Note that code ownership is completely different than reviewers: anyone can
review a piece of code, and we welcome code review from anyone who is review a piece of code, and we welcome code review from anyone who is
interested. Code owners are the "last line of defense" to guarantee that interested. Code owners are the "last line of defense" to guarantee that all
all patches that are committed are actually reviewed.</p> patches that are committed are actually reviewed.</p>
<p>Being a code owner is a somewhat unglamorous position, but it is incredibly <p>Being a code owner is a somewhat unglamorous position, but it is incredibly
important for the ongoing success of the project. Because people get busy, important for the ongoing success of the project. Because people get busy,
interests change, and unexpected things happen, code ownership is purely interests change, and unexpected things happen, code ownership is purely
opt-in, and anyone can choose to resign their "title" at any time. For now, opt-in, and anyone can choose to resign their "title" at any time. For now,
we do not have an official policy on how one gets elected to be a code we do not have an official policy on how one gets elected to be a code
owner. owner.</p>
</p>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"> <a name="testcases">Test Cases</a></div> <div class="doc_subsection"> <a name="testcases">Test Cases</a></div>
<div class="doc_text"> <div class="doc_text">
<p>Developers are required to create test cases for any bugs fixed and any new <p>Developers are required to create test cases for any bugs fixed and any new
features added. Some tips for getting your testcase approved:</p> features added. Some tips for getting your testcase approved:</p>
<ol>
<ol>
<li>All feature and regression test cases are added to the <li>All feature and regression test cases are added to the
<tt>llvm/test</tt> directory. The appropriate sub-directory should be <tt>llvm/test</tt> directory. The appropriate sub-directory should be
selected (see the <a href="TestingGuide.html">Testing Guide</a> for selected (see the <a href="TestingGuide.html">Testing Guide</a> for
details).</li> details).</li>
<li>Test cases should be written in
<a href="LangRef.html">LLVM assembly language</a> unless the
feature or regression being tested requires another language (e.g. the
bug being fixed or feature being implemented is in the llvm-gcc C++
front-end, in which case it must be written in C++).</li>
<li>Test cases, especially for regressions, should be reduced as much as
possible, by <a href="Bugpoint.html">bugpoint</a> or
manually. It is unacceptable
to place an entire failing program into <tt>llvm/test</tt> as this creates
a <i>time-to-test</i> burden on all developers. Please keep them short.</li>
</ol>
<p>Note that llvm/test is designed for regression and small feature tests <li>Test cases should be written in <a href="LangRef.html">LLVM assembly
only. More extensive test cases (e.g., entire applications, benchmarks, language</a> unless the feature or regression being tested requires
etc) should be added to the <tt>llvm-test</tt> test suite. The llvm-test another language (e.g. the bug being fixed or feature being implemented is
suite is for coverage (correctness, performance, etc) testing, not feature in the llvm-gcc C++ front-end, in which case it must be written in
or regression testing.</p> C++).</li>
<li>Test cases, especially for regressions, should be reduced as much as
possible, by <a href="Bugpoint.html">bugpoint</a> or manually. It is
unacceptable to place an entire failing program into <tt>llvm/test</tt> as
this creates a <i>time-to-test</i> burden on all developers. Please keep
them short.</li>
</ol>
<p>Note that llvm/test is designed for regression and small feature tests
only. More extensive test cases (e.g., entire applications, benchmarks, etc)
should be added to the <tt>llvm-test</tt> test suite. The llvm-test suite is
for coverage (correctness, performance, etc) testing, not feature or
regression testing.</p>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"> <a name="quality">Quality</a></div> <div class="doc_subsection"> <a name="quality">Quality</a></div>
<div class="doc_text"> <div class="doc_text">
<p>The minimum quality standards that any change must satisfy before being <p>The minimum quality standards that any change must satisfy before being
committed to the main development branch are:</p> committed to the main development branch are:</p>
<ol>
<li>Code must adhere to the <ol>
<a href="CodingStandards.html">LLVM Coding Standards</a>.</li> <li>Code must adhere to the <a href="CodingStandards.html">LLVM Coding
Standards</a>.</li>
<li>Code must compile cleanly (no errors, no warnings) on at least one <li>Code must compile cleanly (no errors, no warnings) on at least one
platform.</li> platform.</li>
<li>Bug fixes and new features should <a href="#testcases">include a <li>Bug fixes and new features should <a href="#testcases">include a
testcase</a> so we know if the fix/feature ever regresses in the testcase</a> so we know if the fix/feature ever regresses in the
future.</li> future.</li>
<li>Code must pass the dejagnu (<tt>llvm/test</tt>) test suite.</li> <li>Code must pass the dejagnu (<tt>llvm/test</tt>) test suite.</li>
<li>The code must not cause regressions on a reasonable subset of llvm-test, <li>The code must not cause regressions on a reasonable subset of llvm-test,
where "reasonable" depends on the contributor's judgement and the scope where "reasonable" depends on the contributor's judgement and the scope of
of the change (more invasive changes require more testing). A reasonable the change (more invasive changes require more testing). A reasonable
subset might be something like subset might be something like
"<tt>llvm-test/MultiSource/Benchmarks</tt>".</li> "<tt>llvm-test/MultiSource/Benchmarks</tt>".</li>
</ol> </ol>
<p>Additionally, the committer is responsible for addressing any problems
found in the future that the change is responsible for. For example:</p> <p>Additionally, the committer is responsible for addressing any problems found
<ul> in the future that the change is responsible for. For example:</p>
<ul>
<li>The code should compile cleanly on all supported platforms.</li> <li>The code should compile cleanly on all supported platforms.</li>
<li>The changes should not cause any correctness regressions in the <li>The changes should not cause any correctness regressions in the
<tt>llvm-test</tt> suite and must not cause any major performance <tt>llvm-test</tt> suite and must not cause any major performance
regressions.</li> regressions.</li>
<li>The change set should not cause performance or correctness regressions
for the LLVM tools.</li> <li>The change set should not cause performance or correctness regressions for
the LLVM tools.</li>
<li>The changes should not cause performance or correctness regressions in <li>The changes should not cause performance or correctness regressions in
code compiled by LLVM on all applicable targets.</li> code compiled by LLVM on all applicable targets.</li>
<li>You are expected to address any <a href="http://llvm.org/bugs/">bugzilla <li>You are expected to address any <a href="http://llvm.org/bugs/">bugzilla
bugs</a> that result from your change.</li> bugs</a> that result from your change.</li>
</ul> </ul>
<p>We prefer for this to be handled before submission but understand that it <p>We prefer for this to be handled before submission but understand that it
isn't possible to test all of this for every submission. Our nightly isn't possible to test all of this for every submission. Our nightly testing
testing
infrastructure normally finds these problems. A good rule of thumb is to infrastructure normally finds these problems. A good rule of thumb is to
check the nightly testers for regressions the day after your change.</p> check the nightly testers for regressions the day after your change.</p>
<p>Commits that violate these quality standards (e.g. are very broken) may <p>Commits that violate these quality standards (e.g. are very broken) may be
be reverted. This is necessary when the change blocks other developers from reverted. This is necessary when the change blocks other developers from
making progress. The developer is welcome to re-commit the change after making progress. The developer is welcome to re-commit the change after the
the problem has been fixed.</p> problem has been fixed.</p>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
@ -278,15 +301,17 @@
<a name="commitaccess">Obtaining Commit Access</a></div> <a name="commitaccess">Obtaining Commit Access</a></div>
<div class="doc_text"> <div class="doc_text">
<p> <p>We grant commit access to contributors with a track record of submitting high
We grant commit access to contributors with a track record of submitting high quality patches. If you would like commit access, please send an email to
quality patches. If you would like commit access, please send an email to <a href="mailto:sabre@nondot.org">Chris</a> with the following
<a href="mailto:sabre@nondot.org">Chris</a> with the following information:</p> information:</p>
<ol> <ol>
<li>The user name you want to commit with, e.g. "sabre".</li> <li>The user name you want to commit with, e.g. "hacker".</li>
<li>The full name and email address you want message to llvm-commits to come <li>The full name and email address you want message to llvm-commits to come
from, e.g. "Chris Lattner &lt;sabre@nondot.org&gt;".</li> from, e.g. "J. Random Hacker &lt;hacker@yoyodyne.com&gt;".</li>
<li>A "password hash" of the password you want to use, e.g. "2ACR96qjUqsyM". <li>A "password hash" of the password you want to use, e.g. "2ACR96qjUqsyM".
Note that you don't ever tell us what your password is, you just give it Note that you don't ever tell us what your password is, you just give it
to us in an encrypted form. To get this, run "htpasswd" (a utility that to us in an encrypted form. To get this, run "htpasswd" (a utility that
@ -307,135 +332,139 @@ quality patches. If you would like commit access, please send an email to
<p>If you have recently been granted commit access, these policies apply:</p> <p>If you have recently been granted commit access, these policies apply:</p>
<ol> <ol>
<li>You are granted <i>commit-after-approval</i> to all parts of LLVM. <li>You are granted <i>commit-after-approval</i> to all parts of LLVM. To get
To get approval, submit a <a href="#patches">patch</a> to approval, submit a <a href="#patches">patch</a> to
<a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits"> <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits">llvm-commits</a>.
llvm-commits</a>. When approved you may commit it yourself.</li> When approved you may commit it yourself.</li>
<li>You are allowed to commit patches without approval which you think are <li>You are allowed to commit patches without approval which you think are
obvious. This is clearly a subjective decision &mdash; we simply expect you obvious. This is clearly a subjective decision &mdash; we simply expect
to use good judgement. Examples include: fixing build breakage, reverting you to use good judgement. Examples include: fixing build breakage,
obviously broken patches, documentation/comment changes, any other minor reverting obviously broken patches, documentation/comment changes, any
changes.</li> other minor changes.</li>
<li>You are allowed to commit patches without approval to those portions
of LLVM that you have contributed or maintain (i.e., have been assigned <li>You are allowed to commit patches without approval to those portions of
LLVM that you have contributed or maintain (i.e., have been assigned
responsibility for), with the proviso that such commits must not break the responsibility for), with the proviso that such commits must not break the
build. This is a "trust but verify" policy and commits of this nature are build. This is a "trust but verify" policy and commits of this nature are
reviewed after they are committed.</li> reviewed after they are committed.</li>
<li>Multiple violations of these policies or a single egregious violation
may cause commit access to be revoked.</li> <li>Multiple violations of these policies or a single egregious violation may
cause commit access to be revoked.</li>
</ol> </ol>
<p>In any case, your changes are still subject to <a href="#reviews">code <p>In any case, your changes are still subject to <a href="#reviews">code
review</a> (either before or after they are committed, depending on the nature review</a> (either before or after they are committed, depending on the
of the change). You are encouraged to review other peoples' patches as well, nature of the change). You are encouraged to review other peoples' patches
but you aren't required to.</p> as well, but you aren't required to.</p>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"> <a name="newwork">Making a Major Change</a></div> <div class="doc_subsection"> <a name="newwork">Making a Major Change</a></div>
<div class="doc_text"> <div class="doc_text">
<p>When a developer begins a major new project with the aim of contributing <p>When a developer begins a major new project with the aim of contributing it
it back to LLVM, s/he should inform the community with an email to back to LLVM, s/he should inform the community with an email to
the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">llvmdev</a> the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">llvmdev</a>
email list, to the extent possible. The reason for this is to: email list, to the extent possible. The reason for this is to:
<ol>
<ol>
<li>keep the community informed about future changes to LLVM, </li> <li>keep the community informed about future changes to LLVM, </li>
<li>avoid duplication of effort by preventing multiple parties working on
the same thing and not knowing about it, and</li>
<li>ensure that any technical issues around the proposed work are
discussed and resolved before any significant work is done.</li>
</ol>
<p>The design of LLVM is carefully controlled to ensure that all the pieces <li>avoid duplication of effort by preventing multiple parties working on the
fit together well and are as consistent as possible. If you plan to make a same thing and not knowing about it, and</li>
major change to the way LLVM works or want to add a major new extension, it
is a good idea to get consensus with the development
community before you start working on it.</p>
<p>Once the design of the new feature is finalized, the work itself should be <li>ensure that any technical issues around the proposed work are discussed
done as a series of <a href="#incremental">incremental changes</a>, not as and resolved before any significant work is done.</li>
a long-term development branch.</p> </ol>
<p>The design of LLVM is carefully controlled to ensure that all the pieces fit
together well and are as consistent as possible. If you plan to make a major
change to the way LLVM works or want to add a major new extension, it is a
good idea to get consensus with the development community before you start
working on it.</p>
<p>Once the design of the new feature is finalized, the work itself should be
done as a series of <a href="#incremental">incremental changes</a>, not as a
long-term development branch.</p>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"> <a name="incremental">Incremental Development</a> <div class="doc_subsection"> <a name="incremental">Incremental Development</a>
</div> </div>
<div class="doc_text"> <div class="doc_text">
<p>In the LLVM project, we do all significant changes as a series of <p>In the LLVM project, we do all significant changes as a series of incremental
incremental patches. We have a strong dislike for huge changes or patches. We have a strong dislike for huge changes or long-term development
long-term development branches. Long-term development branches have a branches. Long-term development branches have a number of drawbacks:</p>
number of drawbacks:</p>
<ol> <ol>
<li>Branches must have mainline merged into them periodically. If the branch <li>Branches must have mainline merged into them periodically. If the branch
development and mainline development occur in the same pieces of code, development and mainline development occur in the same pieces of code,
resolving merge conflicts can take a lot of time.</li> resolving merge conflicts can take a lot of time.</li>
<li>Other people in the community tend to ignore work on branches.</li> <li>Other people in the community tend to ignore work on branches.</li>
<li>Huge changes (produced when a branch is merged back onto mainline) are <li>Huge changes (produced when a branch is merged back onto mainline) are
extremely difficult to <a href="#reviews">code review</a>.</li> extremely difficult to <a href="#reviews">code review</a>.</li>
<li>Branches are not routinely tested by our nightly tester <li>Branches are not routinely tested by our nightly tester
infrastructure.</li> infrastructure.</li>
<li>Changes developed as monolithic large changes often don't work until the <li>Changes developed as monolithic large changes often don't work until the
entire set of changes is done. Breaking it down into a set of smaller entire set of changes is done. Breaking it down into a set of smaller
changes increases the odds that any of the work will be committed to the changes increases the odds that any of the work will be committed to the
main repository.</li> main repository.</li>
</ol> </ol>
<p> <p>To address these problems, LLVM uses an incremental development style and we
To address these problems, LLVM uses an incremental development style and we
require contributors to follow this practice when making a large/invasive require contributors to follow this practice when making a large/invasive
change. Some tips:</p> change. Some tips:</p>
<ul> <ul>
<li>Large/invasive changes usually have a number of secondary changes that <li>Large/invasive changes usually have a number of secondary changes that are
are required before the big change can be made (e.g. API cleanup, etc). required before the big change can be made (e.g. API cleanup, etc). These
These sorts of changes can often be done before the major change is done, sorts of changes can often be done before the major change is done,
independently of that work.</li> independently of that work.</li>
<li>The remaining inter-related work should be decomposed into unrelated
sets of changes if possible. Once this is done, define the first increment
and get consensus on what the end goal of the change is.</li>
<li>Each change in the set can be stand alone (e.g. to fix a bug), or part <li>The remaining inter-related work should be decomposed into unrelated sets
of a planned series of changes that works towards the development goal.</li> of changes if possible. Once this is done, define the first increment and
get consensus on what the end goal of the change is.</li>
<li>Each change should be kept as small as possible. This simplifies your <li>Each change in the set can be stand alone (e.g. to fix a bug), or part of
work (into a logical progression), simplifies code review and reduces the a planned series of changes that works towards the development goal.</li>
<li>Each change should be kept as small as possible. This simplifies your work
(into a logical progression), simplifies code review and reduces the
chance that you will get negative feedback on the change. Small increments chance that you will get negative feedback on the change. Small increments
also facilitate the maintenance of a high quality code base.</li> also facilitate the maintenance of a high quality code base.</li>
<li>Often, an independent precursor to a big change is to add a new API and <li>Often, an independent precursor to a big change is to add a new API and
slowly migrate clients to use the new API. Each change to use the new slowly migrate clients to use the new API. Each change to use the new API
API is often "obvious" and can be committed without review. Once the is often "obvious" and can be committed without review. Once the new API
new API is in place and used, it is much easier to replace the is in place and used, it is much easier to replace the underlying
underlying implementation of the API. This implementation change is implementation of the API. This implementation change is logically
logically separate from the API change.</li> separate from the API change.</li>
</ul> </ul>
<p>If you are interested in making a large change, and this scares you, please <p>If you are interested in making a large change, and this scares you, please
make sure to first <a href="#newwork">discuss the change/gather make sure to first <a href="#newwork">discuss the change/gather consensus</a>
consensus</a> then ask about the best way to go about making then ask about the best way to go about making the change.</p>
the change.</p>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"><a name="attribution">Attribution of <div class="doc_subsection"><a name="attribution">Attribution of
Changes</a></div> Changes</a></div>
<div class="doc_text"> <div class="doc_text">
<p>We believe in correct attribution of contributions to <p>We believe in correct attribution of contributions to their contributors.
their contributors. However, we do not want the source code to be littered However, we do not want the source code to be littered with random
with random attributions "this code written by J Random Guy" (this is noisy attributions "this code written by J. Random Hacker" (this is noisy and
and distracting). In practice, the revision control system keeps a perfect distracting). In practice, the revision control system keeps a perfect
history of who changed what, and the CREDITS.txt file describes higher-level history of who changed what, and the CREDITS.txt file describes higher-level
contributions. If you commit a patch for someone else, please say "patch contributions. If you commit a patch for someone else, please say "patch
contributed by J Random Guy!" in the commit message.</p> contributed by J. Random Hacker!" in the commit message.</p>
<p>Overall, please do not add contributor names to the source code.</p> <p>Overall, please do not add contributor names to the source code.</p>
</div> </div>
<!--=========================================================================--> <!--=========================================================================-->
<div class="doc_section"> <div class="doc_section">
<a name="clp">Copyright, License, and Patents</a> <a name="clp">Copyright, License, and Patents</a>
@ -443,115 +472,110 @@ Changes</a></div>
<!--=========================================================================--> <!--=========================================================================-->
<div class="doc_text"> <div class="doc_text">
<p>This section addresses the issues of copyright, license and patents for <p>This section addresses the issues of copyright, license and patents for the
the LLVM project. LLVM project. Currently, the University of Illinois is the LLVM copyright
Currently, the University of Illinois is the LLVM copyright holder and the holder and the terms of its license to LLVM users and developers is the
terms of its license to LLVM users and developers is the
<a href="http://www.opensource.org/licenses/UoI-NCSA.php">University of <a href="http://www.opensource.org/licenses/UoI-NCSA.php">University of
Illinois/NCSA Open Source License</a>.</p> Illinois/NCSA Open Source License</a>.</p>
<div class="doc_notes"> <div class="doc_notes">
<p style="text-align:center;font-weight:bold">NOTE: This section <p style="text-align:center;font-weight:bold">NOTE: This section deals with
deals with legal matters but does not provide legal advice. We are not legal matters but does not provide legal advice. We are not lawyers, please
lawyers, please seek legal counsel from an attorney.</p> seek legal counsel from an attorney.</p>
</div> </div>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"><a name="copyright">Copyright</a></div> <div class="doc_subsection"><a name="copyright">Copyright</a></div>
<div class="doc_text"> <div class="doc_text">
<p> <p>For consistency and ease of management, the project requires the copyright
<p>For consistency and ease of management, the project requires the for all LLVM software to be held by a single copyright holder: the University
copyright for all LLVM software to be held by a single copyright holder: of Illinois (UIUC).</p>
the University of Illinois (UIUC).</p>
<p> <p>Although UIUC may eventually reassign the copyright of the software to
Although UIUC may eventually reassign the copyright of the software to another another entity (e.g. a dedicated non-profit "LLVM Organization") the intent
entity (e.g. a dedicated non-profit "LLVM Organization") for the project is to always have a single entity hold the copyrights to LLVM
the intent for the project is to always have a single entity hold the at any given time.</p>
copyrights to LLVM at any given time.</p>
<p>We believe that having a single copyright
holder is in the best interests of all developers and users as it greatly
reduces the managerial burden for any kind of administrative or technical
decisions about LLVM. The goal of the LLVM project is to always keep the code
open and <a href="#license">licensed under a very liberal license</a>.</p>
<p>We believe that having a single copyright holder is in the best interests of
all developers and users as it greatly reduces the managerial burden for any
kind of administrative or technical decisions about LLVM. The goal of the
LLVM project is to always keep the code open and <a href="#license">licensed
under a very liberal license</a>.</p>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"><a name="license">License</a></div> <div class="doc_subsection"><a name="license">License</a></div>
<div class="doc_text"> <div class="doc_text">
<p>We intend to keep LLVM perpetually open source <p>We intend to keep LLVM perpetually open source and to use a liberal open
and to use a liberal open source license. The current license is the source license. The current license is the
<a href="http://www.opensource.org/licenses/UoI-NCSA.php"> <a href="http://www.opensource.org/licenses/UoI-NCSA.php">University of
University of Illinois/NCSA Open Source License</a>, which boils llinois/NCSA Open Source License</a>, which boils down to this:</p>
down to this:</p>
<ul>
<li>You can freely distribute LLVM.</li>
<li>You must retain the copyright notice if you redistribute LLVM.</li>
<li>Binaries derived from LLVM must reproduce the copyright notice (e.g.
in an included readme file).</li>
<li>You can't use our names to promote your LLVM derived products.</li>
<li>There's no warranty on LLVM at all.</li>
</ul>
<p>We believe this fosters the widest adoption of LLVM because it <b>allows <ul>
<li>You can freely distribute LLVM.</li>
<li>You must retain the copyright notice if you redistribute LLVM.</li>
<li>Binaries derived from LLVM must reproduce the copyright notice (e.g. in
an included readme file).</li>
<li>You can't use our names to promote your LLVM derived products.</li>
<li>There's no warranty on LLVM at all.</li>
</ul>
<p>We believe this fosters the widest adoption of LLVM because it <b>allows
commercial products to be derived from LLVM</b> with few restrictions and commercial products to be derived from LLVM</b> with few restrictions and
without a requirement for making any derived works also open source (i.e. without a requirement for making any derived works also open source (i.e.
LLVM's license is not a "copyleft" license like the GPL). We suggest that you LLVM's license is not a "copyleft" license like the GPL). We suggest that you
read the <a href="http://www.opensource.org/licenses/UoI-NCSA.php">License</a> read the <a href="http://www.opensource.org/licenses/UoI-NCSA.php">License</a>
if further clarification is needed.</p> if further clarification is needed.</p>
<p>Note that the LLVM Project does distribute llvm-gcc, <b>which is GPL.</b> <p>Note that the LLVM Project does distribute llvm-gcc, <b>which is GPL.</b>
This means that anything "linked" into llvm-gcc must itself be compatible This means that anything "linked" into llvm-gcc must itself be compatible
with the GPL, and must be releasable under the terms of the GPL. This implies with the GPL, and must be releasable under the terms of the GPL. This
that <b>any code linked into llvm-gcc and distributed to others may be subject implies that <b>any code linked into llvm-gcc and distributed to others may
to the viral aspects of the GPL</b> (for example, a proprietary code generator be subject to the viral aspects of the GPL</b> (for example, a proprietary
linked into llvm-gcc must be made available under the GPL). This is not a code generator linked into llvm-gcc must be made available under the GPL).
problem for code already distributed under a more liberal license (like the This is not a problem for code already distributed under a more liberal
UIUC license), and does not affect code generated by llvm-gcc. It may be a license (like the UIUC license), and does not affect code generated by
problem if you intend to base commercial development on llvm-gcc without llvm-gcc. It may be a problem if you intend to base commercial development
redistributing your source code.</p> on llvm-gcc without redistributing your source code.</p>
<p>We have no plans to change the license of LLVM. If you have questions
or comments about the license, please contact the <a
href="mailto:llvm-oversight@cs.uiuc.edu">LLVM Oversight Group</a>.</p>
<p>We have no plans to change the license of LLVM. If you have questions or
comments about the license, please contact the
<a href="mailto:llvm-oversight@cs.uiuc.edu">LLVM Oversight Group</a>.</p>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"><a name="patents">Patents</a></div> <div class="doc_subsection"><a name="patents">Patents</a></div>
<div class="doc_text"> <div class="doc_text">
<p>To the best of our knowledge, LLVM does not infringe on any patents (we have <p>To the best of our knowledge, LLVM does not infringe on any patents (we have
actually removed code from LLVM in the past that was found to infringe). actually removed code from LLVM in the past that was found to infringe).
Having code in LLVM that infringes on patents would violate an important Having code in LLVM that infringes on patents would violate an important goal
goal of the project by making it hard or impossible to reuse the code for of the project by making it hard or impossible to reuse the code for
arbitrary purposes (including commercial use).</p> arbitrary purposes (including commercial use).</p>
<p>When contributing code, we expect contributors to notify us of any potential <p>When contributing code, we expect contributors to notify us of any potential
for patent-related trouble with their changes. If you or your employer for patent-related trouble with their changes. If you or your employer own
own the rights to a the rights to a patent and would like to contribute code to LLVM that relies
patent and would like to contribute code to LLVM that relies on it, we on it, we require that the copyright owner sign an agreement that allows any
require that other user of LLVM to freely use your patent. Please contact
the copyright owner sign an agreement that allows any other user of LLVM to the <a href="mailto:llvm-oversight@cs.uiuc.edu">oversight group</a> for more
freely use your patent. Please contact the <a
href="mailto:llvm-oversight@cs.uiuc.edu">oversight group</a> for more
details.</p> details.</p>
</div> </div>
<!-- _______________________________________________________________________ --> <!-- _______________________________________________________________________ -->
<div class="doc_subsection"><a name="devagree">Developer Agreements</a></div> <div class="doc_subsection"><a name="devagree">Developer Agreements</a></div>
<div class="doc_text"> <div class="doc_text">
<p>With regards to the LLVM copyright and licensing, developers agree to <p>With regards to the LLVM copyright and licensing, developers agree to assign
assign their copyrights to UIUC for any contribution made so that their copyrights to UIUC for any contribution made so that the entire
the entire software base can be managed by a single copyright holder. This software base can be managed by a single copyright holder. This implies that
implies that any contributions can be licensed under the license that the any contributions can be licensed under the license that the project
project uses.</p> uses.</p>
<p>When contributing code, you also affirm that you are legally entitled to <p>When contributing code, you also affirm that you are legally entitled to
grant this copyright, personally or on behalf of your employer. If the code grant this copyright, personally or on behalf of your employer. If the code
belongs to some other entity, please raise this issue with the oversight belongs to some other entity, please raise this issue with the oversight
group before the code is committed.</p> group before the code is committed.</p>