- Aug 08, 2019
-
-
If the commit range that is being checked with 'dim sparse' introduces new Kconfig options 'make' prompts us for our opinion. Since we are capturing the stdout and -err the user sees nothing and we are seemingly stuck. To avoid that and be as non-interactive as possible let's just squeeze 'make olddefconfig' in-between 'git checkout' and the actual 'make'. Reported-by:
Jani Nikula <jani.nikula@linux.intel.com> Signed-off-by:
Arkadiusz Hiler <arkadiusz.hiler@intel.com> Signed-off-by:
Jani Nikula <jani.nikula@intel.com>
-
- Aug 07, 2019
-
-
Jani Nikula authored
Resurrect commit e03f7356 ("dim: set magic fdo cookie push-option on git push") after server side was fixed. Signed-off-by:
-
Jani Nikula authored
This reverts commit e03f7356 ("dim: set magic fdo cookie push-option on git push"), because: fatal: the receiving end does not support push options fatal: The remote end hung up unexpectedly Signed-off-by:
-
- Aug 06, 2019
-
-
Jani Nikula authored
Make it possible for the server side git pre-receive hook to determine the client is dim and refuse the push otherwise. Obviously it's possible to set the push-option manually on the git command line as well, but there's really no plausible denial in that case. Cc: Daniel Stone <daniel@fooishbar.org> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Emil Velikov <emil.l.velikov@gmail.com> Reviewed-by:
Emil Velikov <emil.velikov@collabora.com> Acked-by:
Daniel Stone <daniels@collabora.com> Signed-off-by:
-
- Aug 02, 2019
-
-
I myself missed that we had this page for a long time since it's not positioned as something to be the first thing to do. Signed-off-by:
Lucas De Marchi <lucas.demarchi@intel.com> Reviewed-by:
Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by:
-
If patch happens to move lines around it's easy to end in a situation where we end up with errors just moving around: -error1:123 error2:124 +error1:125 With remapping, sorting and then diffing without context we end up with: -error1:123 +error1:123 Which is quite confusing and should not generate warning in the first place. This can be fixed by sorting the lines *after* the remapping. Cc: Daniel Vetter <daniel@ffwll.ch> Reported-by:
Oleg Vasilev <oleg.vasilev@intel.com> Signed-off-by:
-
Let's change my email address to the kernel.org one. Signed-off-by:
Maxime Ripard <maxime.ripard@bootlin.com> Signed-off-by:
-
- May 02, 2019
-
-
To stop sending those pesky failure emails, that condition people to ignore anything sent by gitlab-ci. Now we will have to click the button manually, whenever we want to bump the docs. Bugzilla: freedesktop/freedesktop#46 Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by:
Daniel Vetter <daniel.vetter@ffwll.ch>
-
- Mar 11, 2019
-
-
Lucas De Marchi authored
Otherwise we need to document it better so user doesn't try to find out why he doesn't have the remotes he was supposed to have. Signed-off-by:
-
Lucas De Marchi authored
Signed-off-by:
-
Lucas De Marchi authored
Signed-off-by:
-
- Feb 19, 2019
-
-
Simona Vetter authored
We want explicit backmerges when a pull request pulls in random other stuff. Motivated by a few recent examples. v2: - Fix typo in warn_or_fail (Maxime) - check baselines before checking commits, usually a wrong baseline will result in tons of complaints about unreviewed patches and other stuff dim likes to be obsessed about. Cc: Maxime Ripard <maxime.ripard@bootlin.com> Cc: Dave Airlie <airlied@linux.ie> Acked-by:
Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Signed-off-by:
Daniel Vetter <daniel.vetter@intel.com>
-
Simona Vetter authored
Sometimes they get split, and then we lose the branch, pulling the default branch. Cc: Maxime Ripard <maxime.ripard@bootlin.com> Tested-by:
-
- Feb 13, 2019
-
-
Simona Vetter authored
Somewhat inspired by Stephen Rotwell's code, but because random unfriendly encounters with bash it ended up looking quite a bit different. Also only really checks sha1 for now, but that should be good enough. v2: Actually forward the error code from the subshell (which was automatically created because of the pipeline, now I made it explicit). Spotted by shellcheck, somehow I didn't notice that my checks didn't actually abort the pull in my testing. v3: - quiet some noise in the output - also check sha1 length and whether the subject matches. That latter check is a bit more strict than what sfr uses, but matches what dim cite produces. Cc: Dave Airlie <airlied@linux.ie> Acked-by:
Dave Airlie <airlied@linux.ie> Signed-off-by:
-
Simona Vetter authored
fd.o had to switch to mangling From: addresses for a lot of domains. Catch them. Cc: Daniel Stone <daniel@fooishbar.org> References: https://lists.freedesktop.org/archives/freedesktop/2019-February/000396.html Acked-by:
-
- Jan 29, 2019
-
-
dim rebuild-tip was failing for me and after investigation with Daniel Vetter, the culprit was having a pre-commit git hook enabled. Let's pass --no-verify to avoid the same mistake for other people. Reviewed-by:
-
- Jan 21, 2019
-
-
Jani Nikula authored
With the C99 types to kernel types conversion well on the way in drm/i915, let's start warning about C99 types usage, and encourage kernel types. Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by:
-
- Jan 18, 2019
-
-
Jani Nikula authored
Fix: local prompt="$@ (y/N) " ^-- SC2124: Assigning an array to a string! Assign as array, or use * instead of @ to concatenate. Reported-by:
-
- Jan 17, 2019
-
-
Jani Nikula authored
Commit e9846737 ("dim: Dumbest base64 parser ever for apply-pull") broke dim apply-pull for complex multipart messages, such as [1] with mime structure: └┬╴multipart/mixed 11398 bytes ├┬╴multipart/signed inline 2051 bytes │├─╴text/plain inline 1470 bytes │└─╴application/pgp-signature [signature.asc] 195 bytes └─╴text/plain inline 219 bytes The problem lies in the original message_print_body() which would not recurse beyond the first multipart nesting level. Use the message walk() function to also find the text/plain part within the multipart/signed part. [1] http://mid.mail-archive.com/20190117074604.GI18272@zhen-hp.sh.intel.com v2: only print text/plain (Daniel) Fixes: f9bd24c2 ("dim: Add extract-tags command") Fixes: e9846737 ("dim: Dumbest base64 parser ever for apply-pull") Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by:
-
Simona Vetter authored
Some early notes from my first few pull requests. Lots of things that need to be discussed/clarified/agreed upon, but should be good for a starting point at least. Not entirely sure how we should split this all up, so all in one file for now. v2: Fix typo Cc: Dave Airlie <airlied@gmail.com> Acked-by:
Dave Airlie <airlied@gmail.com> Signed-off-by:
-
- Jan 16, 2019
-
-
Simona Vetter authored
This partially reverts commit 6ce0305a. We need to filter both using --first-parent (for proper backmerges) and --committer (for fast-forward merges). v2: Rebase over the push limit from Jani. Cc: Sean Paul <sean@poorly.run> Cc: Maxime Ripard <maxime.ripard@bootlin.com> Cc: Jani Nikula <jani.nikula@linux.intel.com> Reviewed-by:
-
- Jan 15, 2019
-
-
Jani Nikula authored
In most cases, even when pushing long series, it's better to push commits in moderately sized batches. One functional reason is resolving potential drm-tip conflicts in smaller and more isolated units, which is easier to resolve, and is less likely to need a new massive conflict resolution later. Ask the user for confirmation when they're trying to push more than (arbitrarily chosen) 10 patches in one go. This also acts as a safeguard for pushing more than the user intended. Obviously maintainers will also face the question when pushing merges or rebases, but there's no harm in that. They also need to be sure this is what they mean. v2: add --first-parent (Daniel) Reviewed-by:
-
Jani Nikula authored
Use dim_cite() to produce pretty commit message references. Unify and improve error messages while at it. Reviewed-by:
-
Jani Nikula authored
Make dim_cite safer to use within dim. Reviewed-by:
-
- Jan 14, 2019
-
-
Jani Nikula authored
The default repository is master nowadays. Reviewed-by:
-
Jani Nikula authored
It's a separate repository now. Reviewed-by:
-
Jani Nikula authored
Help the users if they so desire. Reviewed-by:
-
Jani Nikula authored
If the dim repo name matches the git remote, only print the repo name. Reviewed-by:
-
Jani Nikula authored
If the dim repo name matches the git remote, only print the repo name. Reviewed-by:
-
- Jan 11, 2019
-
-
Simona Vetter authored
Originally we used the committer to filter for this, but --first-parent is better. Reviewed-and-tested-by:
Sean Paul <sean@poorly.run> Cc: Sean Paul <sean@poorly.run> Cc: Maxime Ripard <maxime.ripard@bootlin.com> Signed-off-by:
-
Jani Nikula authored
The pull requests are more than just FYI for Daniel nowadays. Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by:
-
- Jan 10, 2019
-
-
Jani Nikula authored
Now that drm-intel-next is defined in nightly.conf, we can remove some of the special treatment. Reviewed-by:
-
Jani Nikula authored
Ignore the bool member warning for drm-intel: CHECK:BOOL_MEMBER: Avoid using bool structure members because of possible alignment issues - see: https://lkml.org/lkml/2017/11/21/384 For most of our use cases having the compiler DTRT without explicit !! is preferrable, even at the cost of the relatively small space waste. For more discussion, see also the subthread starting at [1]. [1] http://mid.mail-archive.com/5d0ac1d4-d85d-2238-978a-417362a830bd@intel.com Reviewed-by:
-
- Dec 19, 2018
-
-
Simona Vetter authored
Avoids me having to ask patchwork to decode them for me. Worked on the 2 pulls from Dave and Inki I tried it on. v2: Use message_print_body (Jani). Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Dave Airlie <airlied@gmail.com> Reviewed-by:
-
- Dec 18, 2018
-
-
Using rev-parse git option is safer than manually parsing git files. Signed-off-by:
Andrzej Hajda <a.hajda@samsung.com> Signed-off-by:
-
update_rerere_cache does not work correctly if drm-tip and $DIM_REPO are workdirs - it cannot locate properly rr-cache. Let's use git rev-parse to resolve rr-cache location. Signed-off-by:
-
Assumption that git directory is always located at REPO/.git is incorrect, especially in case of git worktrees. There is already function to deal with it correctly - git_dir, let's then use it. Signed-off-by:
-
git_dir function returns git directory for current working directory. Allowing specifying any directory allows to reuse it more widely. Signed-off-by:
-
- Dec 13, 2018
-
-
Simona Vetter authored
In dim line 976: for patch in $dir/*; do ^-- SC2231: Quote expansions in this for loop glob to prevent wordsplitting, e.g. "$dir"/*.txt . Apparently the shellcheck I have is newer than the one we use for build testing on gitlab. Reviewed-by: -
Simona Vetter authored
Somehow I screwed up and broke make check while at it :-/ Reviewed-by:
-
