All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jacob Stiffler <j-stiffler@ti.com>
To: Denys Dmytriyenko <denys@ti.com>
Cc: meta-arago@arago-project.org
Subject: Re: [morty/master][PATCH 1/2] ocl: update to version 1.1.12.0
Date: Fri, 17 Mar 2017 10:46:35 -0400	[thread overview]
Message-ID: <77986226-486b-b9eb-5ae2-2455688245df@ti.com> (raw)
In-Reply-To: <281d4261-9aba-1257-d0e9-dec0b8f53d81@ti.com>



On 3/17/2017 9:38 AM, Jacob Stiffler wrote:
>
>
> On 3/16/2017 4:28 PM, Denys Dmytriyenko wrote:
>> On Thu, Mar 16, 2017 at 03:23:59PM -0400, Denys Dmytriyenko wrote:
>>> On Thu, Mar 16, 2017 at 03:20:33PM -0400, Denys Dmytriyenko wrote:
>>>> Jake,
>>>>
>>>> It appears the new version no longer provides libOpenCL.so - is it 
>>>> expected?
>>> Actually, it does provide it, just messing up the debian package 
>>> renaming due
>>> to all the new stuff... I'll see if I can correct it.
>> It's not just few packagegroups, it's everything that depends on opencl
>> (opencv, openmp, examples, etc.) is affected. If you build from 
>> scratch every
>> time, you won't see it.
>>
>> I've submitted a patch to hopefully resolve it:
>> http://arago-project.org/pipermail/meta-arago/2017-March/008854.html
>>
>> This doesn't only affect incremental builds, but also package feeds 
>> and would
>> break package/distro upgrades in the field. Unfortunately, we haven't 
>> been
>> paying much attention to this problem in the past, but as we've already
>> discussed, we should keep an eye on it...
>>
> I'm trying to understand this automatic package naming, and how the 
> opencl package gets renamed to libopencl when the IPK is created. 
> Could you point to where this occurs in OE?
>

Reading the comment on your patch shows why this happens, and I have 
found that the code responsible is in the debian.bbclass.

> I also see that on krogoth, everything is OK. Even with the new 
> binaries, the IPKs are still name libopencl*. However, I do see that 
> on morty, the addition of the binaries does alter the IPK names.
>

I see the it was OK because your patch was already committed. Removing 
that patch shows the problem.

> I'll look at the patch and provide comments once I better understand 
> what is happening.
>

We may need to take a closer look at the packaging and dependencies. I 
would assume that the library RDEPENDS on the daemon, so it might not 
make sense to have opposite dependency in the packages (runtime RDEPENDS 
on libraries).

> - Jake



  reply	other threads:[~2017-03-17 14:46 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-03-15 18:01 [morty/master][PATCH 1/2] ocl: update to version 1.1.12.0 Jacob Stiffler
2017-03-15 18:01 ` [morty/master][PATCH 2/2] opencl-examples: add persistent examples for dra7xx Jacob Stiffler
2017-03-16 19:20 ` [morty/master][PATCH 1/2] ocl: update to version 1.1.12.0 Denys Dmytriyenko
2017-03-16 19:23   ` Denys Dmytriyenko
2017-03-16 20:28     ` Denys Dmytriyenko
2017-03-17 13:38       ` Jacob Stiffler
2017-03-17 14:46         ` Jacob Stiffler [this message]
2017-03-17 18:51           ` Denys Dmytriyenko

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=77986226-486b-b9eb-5ae2-2455688245df@ti.com \
    --to=j-stiffler@ti.com \
    --cc=denys@ti.com \
    --cc=meta-arago@arago-project.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.