b4 version 0.5.0 is available

b4 version 0.5.0 is available

[Konstantin Ryabitsev]

[-- Attachment #1: Type: text/plain, Size: 2204 bytes --]

Hello:

I am pleased to release b4 v0.5.0 with the following new features:

b4 diff:

  allows to display range-diffs between patch/series revisions.  It 
  accomplishes this by creating two "fake" detached ranges consisting 
  solely of the files in the patches and then running a "git range-diff" 
  between those ranges. In order to work, the following conditions must 
  all be met:

  1. patches must contain "index" blob information
  2. we must be able to find all of those blobs in the repo

  Behind the scenes, b4 will create a temporary worktree containing just 
  the files mentioned in the patch and run "git am" to create the commit 
  range. After deleting the worktree, this range will continue to exist 
  in the repo as loose objects (until gc cleans them up), allowing to 
  run git range-diff across both patch series.

  In my experience, this works about 60% of the time, but I expect this 
  percentage will be higher for you, since you are more likely to have 
  the needed blobs in your repo due to using the same tree as the 
  submitter.

  Here's a small demo: https://asciinema.org/a/333888

b4 am enhancements:

When we find blob information in the series, we will check if they match 
the current tree. If they do, we will display an "applies clean to 
current tree" message. You can additionally try the following flags:

  -g,--guess-base: will look at the last few tags to check if the 
  patch/series applies cleanly to any of them. I'm not sure how useful 
  this feature is, but it was an easy thing to implement.
  -3,--prep-3way: if you know you have to do a 3-way merge, this will 
  prepare the tree for you by making sure all files exist at indexes 
  specified in the patches. Same caveats apply as for "b4 diff".

The "b4 diff" and "b4 am -3" functionality borrows heavily from Jason 
Gunthorpe's gj_tools [1], for which I am very thankful.

You can upgrade to the latest version by running:

pip install --user --upgrade b4

Please send any inevitable bug reports and feature change suggestions to 
tools@linux.kernel.org.

Best regards,
-K

.. [1]: https://github.com/jgunthorpe/Kernel-Maintainer-Tools

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 235 bytes --]

Re: [kernel.org users] b4 version 0.5.0 is available

[Linus Torvalds]

Hmm.

This fails with the most recent patch-bomb from Andrew when I randomly
picked the email (just because it was most recent in my mailbox
because of a reply, not because of arrival), and I don't see why.

Doing

    b4 am 20200608044014.1sydER6sj%akpm@linux-foundation.org

gets me 51 out of 54 patches, and no cover letter. Strange.

But when I do the same on the _first_ message ID in the series:

    b4 am 20200608044004.04xI8ALrc%akpm@linux-foundation.org

it gets the whole series. It again doesn't find the base commit, but
it's a lot better.

                    Linus

Re: [kernel.org users] b4 version 0.5.0 is available

[James Bottomley]

On Mon, 2020-06-08 at 09:59 -0700, Linus Torvalds wrote:
> Hmm.
> 
> This fails with the most recent patch-bomb from Andrew when I
> randomly
> picked the email (just because it was most recent in my mailbox
> because of a reply, not because of arrival), and I don't see why.
> 
> Doing
> 
>     b4 am 20200608044014.1sydER6sj%akpm@linux-foundation.org
> 
> gets me 51 out of 54 patches, and no cover letter. Strange.
> 
> But when I do the same on the _first_ message ID in the series:
> 
>     b4 am 20200608044004.04xI8ALrc%akpm@linux-foundation.org
> 
> it gets the whole series. It again doesn't find the base commit, but
> it's a lot better.

That's an akpm problem: you can see it here on lore:

https://lore.kernel.org/linux-mm/20200608044004.04xI8ALrc%25akpm@linux-foundation.org/

Lore says it can't find the cover letter because the In-reply-to has
the wrong message id.  This broke threading on my system as well.

James

Re: [kernel.org users] b4 version 0.5.0 is available

[Linus Torvalds]

On Mon, Jun 8, 2020 at 10:30 AM James Bottomley
<James.Bottomley@hansenpartnership.com> wrote:
>
>
> That's an akpm problem: you can see it here on lore:
>
> https://lore.kernel.org/linux-mm/20200608044004.04xI8ALrc%25akpm@linux-foundation.org/
>
> Lore says it can't find the cover letter because the In-reply-to has
> the wrong message id.  This broke threading on my system as well.

Ok, that explains the cover letter missing.

But what about the "patches 1-3 aren't picked up if I picked the
message ID of patch 4" issue?

lore seems to thread the series fine, and sees it all.

               Linus

Re: [kernel.org users] b4 version 0.5.0 is available

[Konstantin Ryabitsev]

On Mon, Jun 08, 2020 at 10:59:19AM -0700, Linus Torvalds wrote:
> > That's an akpm problem: you can see it here on lore:
> >
> > https://lore.kernel.org/linux-mm/20200608044004.04xI8ALrc%25akpm@linux-foundation.org/
> >
> > Lore says it can't find the cover letter because the In-reply-to has
> > the wrong message id.  This broke threading on my system as well.
> 
> Ok, that explains the cover letter missing.
> 
> But what about the "patches 1-3 aren't picked up if I picked the
> message ID of patch 4" issue?

That's a b4 problem -- I have a fix for it in master/stable-0.5.y and I 
will release 0.5.1 in the near future with that fix.

Gotta say, Andrew's patchbombs have been a great fuzzing tool thus far.

-K