Rodrigo Vivi authored 1 month ago and Rodrigo Vivi committed 3 weeks ago

The original fix of fix on commit 28155089 ("dim: Consider
fix-of-fix for -fixes cherry picking.") was:

fix_of_fix=$(git log --grep="cherry picked from commit $fixes" \
		  --after=6months --format=format:%h \
		  $intel_remote/drm-intel-next-fixes -1)
if [[ -z "$fix_of_fix" ]]; then
 fix_of_fix=$(git log --grep="cherry picked from commit $fixes" \
 --after=6months --format=format:%h $intel_remote/drm-intel-fixes -1)

But commit e7f73056 ("dim: Generalize cherry-pick fixes")
generalized it to for b in drm-intel-next drm-intel-gt-next

So, we need to get back to look for the commits in the fix branch
not in the next branch. However, let's only look to the current
fix branch instead of both fixes branches, otherwise we will can
reach a point where we can try to cherry-pick a fix_of_fix to a branch
which doesn't have the first-fix patch.

Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

The section about `dim cherry-pick` precedes the docs split in a committer
and maintainer sections. Having it in the committer/getting-started.rst is
misleading, because only maintainers are supposed to cherry-pick commits.

Suggested-by: Simona Vetter <simona.vetter@ffwll.ch>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

Bypass any search for a config file if DIM_CONFIG is set to an empty
string.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

Even if git handles a non-existent path correctly while cloning, it's
distracting and people cut the log short when reporting. Just omit the
argument if DIM_KERNEL_REFERENCE is empty.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

Tejas reported that dim fails on a simple `dim setup` with the following
message:

	The DIM_REPO maintainer kernel repository 'src' doesn't exist. Clone upstream? (y/N) y
	Cloning into 'src'...
	info: Could not add alternate for '': path '' does not exist

git should continue if the path doesn't exist, but apparently this
doesn't happen with all git versions if the argument is an empty string.
The original commit tried to workaround that by setting it to a
non-existent path, but there was a typo in the variable name. Fix the
typo.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

The subcommand "retip" allows passing extra git rebase options, but that
requires the branch name to be explicitly passed as first argument.
Since we do not expect branch names to start with a dash, let's allow
the user to pass extra rebase options without the need to explicitly
pass the branch name if the current one is to be used.

Signed-off-by: Gustavo Sousa <gustavo.sousa@intel.com>

Just for generating pulls, that avoids tampering with the user's
config.

References: https://lwn.net/Articles/931180/


Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

Disappointingly gitk doesn't obey .mailmap, but everything else does.

Signed-off-by: Simona Vetter <simona.vetter@ffwll.ch>

Jani Nikula authored 7 months ago and Jani Nikula committed 7 months ago

Add the missing subcommand documentation.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Jani Nikula authored 7 months ago and Jani Nikula committed 7 months ago

'make mancheck' has complained about missing documentation since commit
e93a36ec ("dim: add a simple facility to require minimum dim
version") with:

mancheck: version not documented

But it does not actually fail the check, just prints that out, so nobody
noticed.

Use 'false' to bail out on errors and fail CI on mancheck failures.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Link to kernel.org instead of 01.org. Use https instead of http. Drop
the inline link markup. Avoid redirects where feasible. Reference
kernel.org DRM UAPI requirements page instead of Sima's old blog.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

The man page has excessive (and broken) links to various pages. Just
link to the maintainer tools main page, and describe what's there.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Replace all links to the dim man page with :ref:`dim`, fixing a number
of broken links while at it. The :doc:`/dim` method would still work,
but unify on :ref:.

Fix contributing link to point at the right document instead of the man
page.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

The usual way to add links across documents in Sphinx is to use
automatic labels and ref as described in [1].

However, adding e.g. ".. _dim:" automatic label triggers warnings with
rst2man --strict such as:

dim.rst:5: (INFO/1) Duplicate implicit target name: "dim".

and trying with a different label such as ".. _dim-man:":

dim.rst:1: (INFO/1) Hyperlink target "dim-man" is not referenced.

Just relax the strictness a notch by ignoring INFO level messages, and
halting on WARNING level instead, to allow us to improve
cross-referencing in Sphinx documentation.

[1] https://www.sphinx-doc.org/en/master/usage/referencing.html#role-ref



Signed-off-by: Jani Nikula <jani.nikula@intel.com>

When moving pages around I, for whatever reason, didn't consider that
this creates a bunch of stale links. The drm-misc link in MAINTAINERS
being perhaps the worst one to go 404.

We could, and should, fix a lot of the now stale links, but in the mean
time, use the sphinx-reredirects extension [1] to create redirects for
the relocated pages.

[1] https://documatt.com/sphinx-reredirects/



Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Add a place to log the required Python packages for building
documentation. It's only Sphinx for starters, but we'll add more.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Just a light link update.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Note the -f option to dim may be required. Mention intel-xe mailing
list. Rephrase reverts to make them okay to apply instead of removing
commits. Add links, brush up.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Unify access request text for all the repos. This changes the text
slightly for all repos:

- drm-misc gains "Commit rights may be automatically revoked after a
  year of inactivity (no commits or reviews)". For drm-intel/drm-xe that
  changes from "will be" to "may be" revoked.

- drm-intel/drm-xe lose the text on rate limiting the number of new
  committers. This was barely an issue when we started adding committers
  in the first place, but certainly not anymore.

- Remove drm-intel/drm-xe repo details. They belong in repo
  documentation and MAINTAINERS.

Add links to creating access request issues to drm-intel/drm-xe
too. They URLs contain the info about the issue template to use, which
we don't have in place yet, but there's no harm in using that already.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Remove extra fluff, add links, etc.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Remove extra fluff, add links.

The flowchart and timeline also need update, but leave that for later.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Streamline drm-misc description into the top section, and remove the
superfluous (and stale) commit access descriptions.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Feels a bit silly to repeat the same thing for each repository. Keep it,
but make it a note in the main index page.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Brush up, remove duplication, add links.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Brush up, remove duplication, add links.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

The kernel.org MAINTAINERS document has anchors letting us link directly
there. Take advantage of it.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Add link to kernel.org, don't single out i915 as an example.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Fix the 404.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Add hyperlinks to referenced content.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Minor brush-up, add links, etc.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Update Makefile from sphinx-quickstart. Switch to using the new "make
mode" option.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Jani Nikula authored 7 months ago and Jani Nikula committed 7 months ago

The getting started document is really about getting started as a
committer. Move it to the suitable place under committer guidelines.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Jani Nikula authored 7 months ago and Jani Nikula committed 7 months ago

This info is for maintainers, not really about repositories.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Jani Nikula authored 7 months ago and Jani Nikula committed 7 months ago

This info is for committers, not really about repositories.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Jani Nikula authored 7 months ago and Jani Nikula committed 7 months ago

This info is for committers, not really about repositories.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Jani Nikula authored 7 months ago and Jani Nikula committed 7 months ago

Seems like a more fitting place than the top level.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Jani Nikula authored 7 months ago and Jani Nikula committed 7 months ago

Group maintainer guidelines together in a subdirectory

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Jani Nikula authored 7 months ago and Jani Nikula committed 7 months ago

Group committer guidelines together in a subdirectory

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Jani Nikula authored 7 months ago and Jani Nikula committed 7 months ago

It was always a bit odd that we had spread the information to several
places. Describe each repository in its own page.

This is mostly just movement, with just minor updates, and will lead to
slight duplication that will need to be cleaned up in follow-up.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>