no-preinstall: Ready to merge

[Joel Sherrill]

On Mon, Jan 22, 2018 at 6:23 PM, Joel Sherrill <joel at rtems.org> wrote:

>
>
> On Mon, Jan 22, 2018 at 5:59 PM, Chris Johns <chrisj at rtems.org> wrote:
>
>> On 23/01/2018 10:50, Joel Sherrill wrote:
>> > On Mon, Jan 22, 2018 at 4:35 PM, Chris Johns <chrisj at rtems.org
>> > <mailto:chrisj at rtems.org>> wrote:
>> >
>> >     On 23/01/2018 01:49, Sebastian Huber wrote:
>> >     > Hello,
>> >     >
>> >     > I installed all BSPs of the master and then all BSPs of the
>> no-preinstall
>> >     > branch.  The difference in installed BSP files is this:
>> >     >
>> >     > https://ftp.rtems.org/pub/rtems/people/sebh/tmp/headers-diff.txt
>> >     <https://ftp.rtems.org/pub/rtems/people/sebh/tmp/headers-diff.txt>
>> >     >
>> >     > There are no removals.  In the no-preinstall installation, there
>> are more header
>> >     > files present (this is expected).  If someone has a problem with
>> this, then we
>> >     > can fine tune this later.  For the preinstall removal this is not
>> a blocking point.
>> >     >
>> >     > The squashed commit is at the top of the no-preinstall branch:
>> >     >
>> >     > https://git.rtems.org/sebh/rtems.git/log/?h=no-preinstall
>> >     <https://git.rtems.org/sebh/rtems.git/log/?h=no-preinstall>
>> >     >
>> >
>> >     I OK with this change being pushed to master.
>> >
>> >     It would be great if someone who experienced the original problem
>> of the
>> >     'install' command failing on Linux during the preinstall phase
>> could please run
>> >     a test build with a large number of jobs to see if the problem has
>> been
>> >     resolved.
>> >
>> >
>> > I think that was a not so subtle jab at me.
>>
>> Not really, it was to anyone effected and willing to help test. I seem to
>> remember other reports of this happening.
>>
>
> Thanks Chris for the help fetching this. I have started a build of all BSPs
> on the no-preinstall branch on rtbf64c. I could pretty reliably get
> failures
> on this machine with preinstall and make -j.
>
> I will report in the morning.
>

No issues using make -j with the number of jobs being 8 on a machine with
12 cores (6 real). Same failures as master. powerpc-qoriq_core_0 and
powerpc-qoriq_core_1 did not build.

This was with my test script. I will rebuild today with the bsp-builder[1]

[1] The bsp-builder still has a %s incorrectly in the output

[   2/1565] arm/altcycv_devkit (no-posix)                      Building
run: %s: arm/altcycv_devkit.no-posix\
         make -j 3



>
>>
>> > Do you want me to test before or after Sebastian pushes the branch.
>>
>> I do not mind. This is the original reason for the change and so it would
>> be
>> nice to be able to close any related tickets. I cannot do this because I
>> have
>> not seen the problem.
>>
>> >
>> > I am ok pushing it and cleaning up problems as we get more users.
>> >
>>
>> Sure.
>>
>> Chris
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20180123/f918f879/attachment-0002.html>