docs: Update release notes with current process

There's a lot of stuff here that's out of date, update it to something
more modern.

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4066>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4066>
This commit is contained in:
Dylan Baker 2020-03-05 10:31:22 -08:00 committed by Marge Bot
parent 7451eb9a27
commit d4cb9ef826
1 changed files with 20 additions and 214 deletions

View File

@ -23,7 +23,6 @@
<li><a href="#pickntest">Cherry-pick and test</a>
<li><a href="#stagingbranch">Staging branch</a>
<li><a href="#branch">Making a branchpoint</a>
<li><a href="#prerelease">Pre-release announcement</a>
<li><a href="#release">Making a new release</a>
<li><a href="#announce">Announce the release</a>
<li><a href="#gitlab">Update Gitlab Issues</a>
@ -74,16 +73,11 @@ on the mesa-announce@ mailing list.
<li>Typically, the final release will happen after 4
candidates. Additional ones may be needed in order to resolve blocking
regressions, though.
<li>A <a href="#prerelease">pre-release</a> announcement should be available
approximately 24 hours before the final (non-rc) release.
</ul>
<h2>Stable releases</h2>
<ul>
<li>Normally available once every two weeks.
<li>Only the latest branch has releases. See note below.
<li>A <a href="#prerelease">pre-release</a> announcement should be available
approximately 48 hours before the actual release.
</ul>
<p>
@ -113,20 +107,24 @@ described in the same section.
</p>
<p>
Nomination happens in the mesa-stable@ mailing list. However,
maintainer is responsible of checking for forgotten candidates in the
master branch. This is achieved by a combination of ad-hoc scripts and
a casual search for terms such as regression, fix, broken and similar.
Nominations happen via special tags in the commit messages, and via gitlab
merge requests against the staging branches. There are special scripts used
to read the tags.
</p>
<p>
Maintainer is also responsible for testing in various possible permutations of
the meson and scons build.
The maintainer should watch or be in contact with the Intel CI team, as well
as watch the gitlab CI for regressions.
</p>
<h2>Cherry-picking and build/check testing</h2>
<p>
Cherry picking should be done with the '-x' switch (to automatically add
"cherry picked from ..." to the commit message):
</p>
<p>Done continuously up-to the <a href="#prerelease">pre-release</a> announcement.</p>
<code>
git cherry-pick -x abcdef12345667890
</code>
<p>
Developers can request, <em>as an exception</em>, patches to be applied up-to
@ -141,20 +139,7 @@ well contained. Thus it cannot affect more than one driver/subsystem.
<li><em>AMD team</em>
</ul>
<p>The following must pass:</p>
<ul>
<li>meson test, scons and scons check
<li>Testing with different version of system components - LLVM and others is also
performed where possible.
<li>As a general rule, testing with various combinations of configure
switches, depending on the specific patchset.
</ul>
<p>
These are achieved by combination of <a href="basictesting">local testing</a>,
which includes mingw-w64 cross compilation and AppVeyor plus Travis-CI, the
latter two as part of their Github integration.
</p>
<p>The gitlab CI must pass.</p>
<p>
For Windows related changes, the main contact point is Brian
@ -178,10 +163,6 @@ fix(es), then they should be squashed together. The commit messages and the
&quot;<code>cherry picked from</code>&quot;-tags must be preserved.
</p>
<p>
This should be noted in the <a href="#prerelease">pre-announce</a> email.
</p>
<pre>
git show b10859ec41d09c57663a258f43fe57c12332698e
@ -208,27 +189,12 @@ Date: Wed Mar 1 18:11:10 2017 +0100
<h2>Regression/functionality testing</h2>
<p>
Less often (once or twice), shortly before the pre-release announcement.
Ensure that testing is redone if Intel devs have requested an exception, as per above.
</p>
<ul>
<li><em>no regressions should be observed for Piglit/dEQP/CTS/Vulkan on Intel platforms</em>
<li><em>no regressions should be observed for Piglit using the swrast, softpipe
and llvmpipe drivers</em>
</ul>
<p>
Currently testing is performed courtesy of the Intel OTC team and their Jenkins CI setup. Check with the Intel team over IRC how to get things setup.
</p>
<p>
Installing the built driver from the pre-announced RC branch in the
system and making some every day's use until the release may be a good
idea too.
</p>
<h2 id="stagingbranch">Staging branch</h2>
<p>
@ -277,8 +243,8 @@ git checkout -b X.Y
git checkout master
$EDITOR VERSION # bump the version number
git commit -as
cp docs/relnotes/{X.Y,X.Y+1}.html # copy/create relnotes template
git commit -as
truncate docs/relnotes/new_features.txt
git commit -a
git push origin X.Y-branchpoint X.Y
</pre>
@ -293,162 +259,6 @@ For example: files being overwritten on install, etc. Happens extremely rarely -
we had only one case so far (see commit 2ced8eb136528914e1bf4e000dea06a9d53c7e04).
</p>
<p>
Proceed to <a href="#release">release</a> -rc1.
</p>
<h2 id="prerelease">Pre-release announcement</h2>
<p>
It comes shortly after outstanding patches in the respective branch are pushed.
Developers can check, in brief, what's the status of their patches. They,
alongside very early testers, are strongly encouraged to test the branch and
report any regressions.
</p>
<p>
It is followed by a brief period (normally 24 or 48 hours) before the actual
release is made.
</p>
<p>
Be aware to add a note to warn about a final release in a series, if
that is the case.
</p>
<h2>Terminology used</h2>
<ul><li>Nominated</ul>
<p>
Patch that is nominated but yet to to merged in the patch queue/branch.
</p>
<ul><li>Queued</ul>
<p>
Patch is in the queue/branch and will feature in the next release.
Barring reported regressions or objections from developers.
</p>
<ul><li>Rejected</ul>
<p>
Patch does not fit the
<a href="submittingpatches.html#criteria" target="_parent">criteria</a> and
is followed by a brief information. The release maintainer is human so if you
believe you've spotted a mistake do let them know.
</p>
<h2>Format/template</h2>
<pre>
Subject: [ANNOUNCE] Mesa X.Y.Z release candidate
To: mesa-announce@...
Cc: mesa-dev@...
Hello list,
The candidate for the Mesa X.Y.Z is now available. Currently we have:
- NUMBER queued
- NUMBER nominated (outstanding)
- and NUMBER rejected patches
[If applicable:
Note: this is the final anticipated release in the SERIES series. Users are
encouraged to migrate to the NEXT_SERIES series in order to obtain future fixes.]
BRIEF SUMMARY OF CHANGES
Take a look at section "Mesa stable queue" for more information.
Testing reports/general approval
--------------------------------
Any testing reports (or general approval of the state of the branch) will be
greatly appreciated.
The plan is to have X.Y.Z this DAY (DATE), around or shortly after TIME.
If you have any questions or suggestions - be that about the current patch
queue or otherwise, please go ahead.
Trivial merge conflicts
-----------------------
List of commits where manual intervention was required.
Keep the authors in the CC list.
commit SHA
Author: AUTHOR
COMMIT SUMMARY
CHERRY PICKED FROM
For example:
commit 990f395e007c3204639daa34efc3049f350ee819
Author: Emil Velikov &lt;emil.velikov@collabora.com&gt;
anv: automake: cleanup the generated json file during make clean
(cherry picked from commit 8df581520a823564be0ab5af7dbb7d501b1c9670)
Cheers,
Emil
Mesa stable queue
-----------------
Nominated (NUMBER)
==================
AUTHOR (NUMBER):
SHA COMMIT SUMMARY
For example:
Dave Airlie (1):
2de85eb radv: fix texturesamples to handle single sample case
Queued (NUMBER)
===============
AUTHOR (NUMBER):
COMMIT SUMMARY
[If applicable:
Squashed with
COMMIT SUMMARY]
For example:
Jonas Pfeil (1):
ralloc: Make sure ralloc() allocations match malloc()'s alignment.
Squashed with
ralloc: don't leave out the alignment factor
Rejected (NUMBER)
=================
AUTHOR (NUMBER):
SHA COMMIT SUMMARY
Reason: ...
For example:
Emil Velikov (1)
a39ad18 configure.ac: honour LLVM_LIBDIR when linking against LLVM
Reason: The patch was reverted shortly after it was merged.
</pre>
<h2 id="release">Making a new release</h2>
<p>
@ -466,8 +276,7 @@ relevant branch.
<p>
Most of the testing should already be done during the
<a href="#pickntest">cherry-pick</a> and
<a href="#prerelease">pre-announce</a> stages.
<a href="#pickntest">cherry-pick</a>
So we do a quick 'touch test'
</p>
@ -521,13 +330,11 @@ unset VK_ICD_FILENAMES
<p>
The release notes are completely generated by the
<code>bin/gen_release_notes.py</code> script. Simply run this script before
bumping the version
bumping the version, and commit the results.
The only thing left to do is add the sha256 sums.
</p>
<h3>Update version in file VERSION</h3>
<p>
Increment the version contained in the file VERSION at Mesa's top-level, then
commit this change.
@ -549,8 +356,6 @@ Start the release process.
</p>
<pre>
# For the dist/distcheck, you may want to specify which LLVM to use:
# export LLVM_CONFIG=/usr/lib/llvm-3.9/bin/llvm-config
../relative/path/to/release.sh . # append --dist if you've already done distcheck above
</pre>
@ -562,7 +367,7 @@ and SSH passphrase(s) to sign and upload the files, respectively.
<h3>Add the sha256sums to the release notes</h3>
<p>
Edit docs/relnotes/X.Y.Z.html to add the sha256sums as available in the mesa-X.Y.Z.announce template. Commit this change.
Edit docs/relnotes/X.Y.Z.html to add the sha256sum as available in the mesa-X.Y.Z.announce template. Commit this change.
</p>
<h3>Back on mesa master, add the new release notes into the tree</h3>
@ -578,7 +383,8 @@ git cherry-pick -x X.Y
<p>Then run the <code>./bin/post_verison.py X.Y.Z</code>, where X.Y.Z is the
version you just made. This will updated docs/relnotes.html and
docs/index.html. Remove docs/release-calendar.html. Then commit and push:
docs/index.html. Remove the new release from docs/release-calendar.html. Then
commit and push:
</p>
<pre>