Openembedded Core Discussions
 help / color / mirror / Atom feed
From: Markus Lehtonen <markus.lehtonen@linux.intel.com>
To: Paul Eggleton <paul.eggleton@linux.intel.com>
Cc: openembedded-core@lists.openembedded.org
Subject: Re: [PATCH 0/6] devtool: improve handling of local source files
Date: Thu, 11 Jun 2015 15:53:07 +0300	[thread overview]
Message-ID: <1434027187.12084.54.camel@linux.intel.com> (raw)
In-Reply-To: <4504055.vUCZMjBN1g@peggleto-mobl.ger.corp.intel.com>

Hi Paul,

On Thu, 2015-06-04 at 14:49 +0100, Paul Eggleton wrote:
> On Thursday 04 June 2015 16:12:07 Markus Lehtonen wrote:
> > On Tue, 2015-05-12 at 19:01 +0100, Paul Eggleton wrote:
> > > On Thursday 30 April 2015 12:16:06 Markus Lehtonen wrote:
> > > > This patchset tries to improve handling of local source files (i.e.
> > > > file://
> > > > in SRC_URI). First, it improves packages for which S=WORKDIR (that
> > > > possibly
> > > > only have local sources. Second, it makes local sources available in the
> > > > srctree for all packages.
> > > > 
> > > > See yocto bug #7602
> > > 
> > > I've finally looked at these, apologies for the delay. Some comments:
> > > 
> > > * I don't think we really want the local files to become part of the git
> > > repository by default - they shouldn't be committed. Once users have
> > > finished with devtool, we want them to be able to push the source tree to
> > > their own repo and point to that within the recipe, whilst keeping the
> > > local files next to the recipe.
> > 
> > So you suggest to add a new command line option to devtool extract and
> > modify (--local-files or smth)? What to do when there are only local
> > files (no source tarball / repo) - automatically enable --local-files in
> > this case?
> 
> Is another option really required? Unless I'm missing something, I would have 
> thought the behaviour for local files ought to be the same regardless of 
> whether they are in addition to the upstream source, or the only files in 
> SRC_URI.

Currently, the local files are already committed into Git in some cases
(i.e. basically when S==WORKDIR, e.g. with makedevs).

What would you suggest to do with the local files? Copy the files to git
repo worktree but add them to .git/info/exclude? Or not copy them at
all? In this case some packages would not be supported anymore (e.g. the
makedevs mentioned before).

I agree that the behavior should be consistent and the same regardless
of the existence of upstream sources.


> > > * This implies that new files added to the local files dir when we do
> > > devtool update-recipe should not be added as a patch, they should be
> > > copied next to the recipe and added to SRC_URI. I'm more than happy for
> > > us to implement this separately as a follow-up (i.e. we could start by
> > > not handling adding files to the local files directory at all.)
> > 
> > Yeah, I actually have this WIP. Currently (i.e. with the current
> > patchset), new files added to 'local-files' are just ignored. They are
> > not copied and no patches is generated out of these.
> 
> OK, then it sounds like the behaviour for added files is reasonable for the 
> moment and we can extend it as a follow-up.

OK, thanks.


> > > * The local-files directory needs to be named specific to OE -
> > > "oe-local-files" would be ideal. If we could have one place in the code
> > > where this was defined that would be ideal as well (maybe at some point
> > > we'd allow it to be configured).
> > 
> > This is not a big deal. Should it perhaps be "bb-local-files" instead?
> 
> Well strictly speaking all of this is being defined in OE, not bitbake, hence 
> my suggestion of "oe-local-files".

OK, "oe-local-files" is probably fine. I was just thinking about Poky or
some other derived project where the files might not be from OE.


Thanks,
  Markus





      reply	other threads:[~2015-06-11 12:53 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-04-30  9:16 [PATCH 0/6] devtool: improve handling of local source files Markus Lehtonen
2015-04-30  9:16 ` [PATCH 1/6] devtool: extract: remove patches when S=WORKDIR Markus Lehtonen
2015-04-30  9:16 ` [PATCH 2/6] recipeutils: implement get_recipe_local_files() Markus Lehtonen
2015-04-30  9:16 ` [PATCH 3/6] oe.patch.GitApplyTree: add paths argument to extractPatches Markus Lehtonen
2015-04-30  9:16 ` [PATCH 4/6] devtool: update-recipe: update local files directly Markus Lehtonen
2015-04-30  9:16 ` [PATCH 5/6] devtool: extract: always import local files to srctree Markus Lehtonen
2015-04-30  9:16 ` [PATCH 6/6] devtool: modify: make bitbake use local files from srctree Markus Lehtonen
2015-05-12 18:01 ` [PATCH 0/6] devtool: improve handling of local source files Paul Eggleton
2015-06-04 13:12   ` Markus Lehtonen
2015-06-04 13:49     ` Paul Eggleton
2015-06-11 12:53       ` Markus Lehtonen [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1434027187.12084.54.camel@linux.intel.com \
    --to=markus.lehtonen@linux.intel.com \
    --cc=openembedded-core@lists.openembedded.org \
    --cc=paul.eggleton@linux.intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox