mbox series

[PULL,00/21] Docker and shippable updates

Message ID 20170621144814.15324-1-alex.bennee@linaro.org
Headers show
Series Docker and shippable updates | expand

Message

Alex Bennée June 21, 2017, 2:47 p.m. UTC
The following changes since commit 8dfaf23ae1f2273a9730a9b309cc8471269bb524:

  tcg/tci: fix tcg-interpreter build (2017-06-20 18:39:15 +0100)

are available in the git repository at:

  https://github.com/stsquad/qemu.git tags/pull-ci-updates-210617-2

for you to fetch changes up to 32b9ca986855a5d56daf47fdb516743008788b71:

  MAINTAINERS: self-appoint me as reviewer in build/test automation (2017-06-21 15:03:06 +0100)

----------------------------------------------------------------
This is mostly Philippe's updates

We add the following cross-compile targets:
  - mipsel-softmmu,mipsel-linux-user,mips64el-linux-user
  - armeb-linux-user

While I was rolling I discovered we could also back out a bunch of the
emdebian hacks as the newly released stretch handles cross compilers
as first class citizens. Unfortunately this also meant I had to drop
the powerpc support as that is no longer in Debian stable.

----------------------------------------------------------------
Alex Bennée (1):
      docker: update qemu:debian base following stretch release

Philippe Mathieu-Daudé (20):
      docker: let _copy_with_mkdir() sub_path argument be optional
      docker: add --include-files argument to 'build' command
      docker: rebuild image if 'extra files' checksum does not match
      docker: install ca-certificates package in base image
      docker: use better regex to generate deb-src entries
      docker: use eatmydata, install common build packages in base image
      docker: use eatmydata in debian armhf image
      docker: use eatmydata in debian arm64 image
      docker: add extra libs to armhf target to extend codebase coverage
      docker: add extra libs to arm64 target to extend codebase coverage
      docker: add extra libs to s390x target to extend codebase coverage
      docker: add mipsel build target
      shippable: use C locale to simplify console output
      shippable: build using all available cpus
      shippable: do not initialize submodules automatically
      shippable: be verbose while building docker images
      shippable: add armeb-linux-user target
      shippable: add mipsel target
      MAINTAINERS: add Shippable automation platform URL
      MAINTAINERS: self-appoint me as reviewer in build/test automation

 .shippable.yml                                     | 21 ++++++++++++---
 MAINTAINERS                                        |  2 ++
 tests/docker/Makefile.include                      |  4 +++
 tests/docker/docker.py                             | 30 +++++++++++++++++-----
 tests/docker/dockerfiles/debian-arm64-cross.docker | 18 +++++++++++--
 tests/docker/dockerfiles/debian-armhf-cross.docker | 18 +++++++++++--
 .../docker/dockerfiles/debian-mipsel-cross.docker  | 29 +++++++++++++++++++++
 tests/docker/dockerfiles/debian-s390x-cross.docker | 10 ++++++++
 tests/docker/dockerfiles/debian.docker             | 15 ++++-------
 9 files changed, 123 insertions(+), 24 deletions(-)
 create mode 100644 tests/docker/dockerfiles/debian-mipsel-cross.docker

Comments

Peter Maydell June 22, 2017, 2:09 p.m. UTC | #1
On 21 June 2017 at 15:47, Alex Bennée <alex.bennee@linaro.org> wrote:
> The following changes since commit 8dfaf23ae1f2273a9730a9b309cc8471269bb524:

>

>   tcg/tci: fix tcg-interpreter build (2017-06-20 18:39:15 +0100)

>

> are available in the git repository at:

>

>   https://github.com/stsquad/qemu.git tags/pull-ci-updates-210617-2

>

> for you to fetch changes up to 32b9ca986855a5d56daf47fdb516743008788b71:

>

>   MAINTAINERS: self-appoint me as reviewer in build/test automation (2017-06-21 15:03:06 +0100)

>

> ----------------------------------------------------------------

> This is mostly Philippe's updates

>

> We add the following cross-compile targets:

>   - mipsel-softmmu,mipsel-linux-user,mips64el-linux-user

>   - armeb-linux-user

>

> While I was rolling I discovered we could also back out a bunch of the

> emdebian hacks as the newly released stretch handles cross compilers

> as first class citizens. Unfortunately this also meant I had to drop

> the powerpc support as that is no longer in Debian stable.

>


Applied, thanks.

-- PMM
Philippe Mathieu-Daudé June 22, 2017, 2:58 p.m. UTC | #2
On 06/22/2017 11:09 AM, Peter Maydell wrote:
> On 21 June 2017 at 15:47, Alex Bennée <alex.bennee@linaro.org> wrote:

[...]>> We add the following cross-compile targets:
>>    - mipsel-softmmu,mipsel-linux-user,mips64el-linux-user

>>    - armeb-linux-user

>>

>> While I was rolling I discovered we could also back out a bunch of the

>> emdebian hacks as the newly released stretch handles cross compilers

>> as first class citizens. Unfortunately this also meant I had to drop

>> the powerpc support as that is no longer in Debian stable.

>>

> 

> Applied, thanks.


Yay! Shippable happily working again \o/

https://app.shippable.com/github/qemu/qemu/runs/238/summary/console

   Overall Status: Success
Alex Bennée June 22, 2017, 3:13 p.m. UTC | #3
Philippe Mathieu-Daudé <f4bug@amsat.org> writes:

> On 06/22/2017 11:09 AM, Peter Maydell wrote:

>> On 21 June 2017 at 15:47, Alex Bennée <alex.bennee@linaro.org> wrote:

> [...]>> We add the following cross-compile targets:

>>>    - mipsel-softmmu,mipsel-linux-user,mips64el-linux-user

>>>    - armeb-linux-user

>>>

>>> While I was rolling I discovered we could also back out a bunch of the

>>> emdebian hacks as the newly released stretch handles cross compilers

>>> as first class citizens. Unfortunately this also meant I had to drop

>>> the powerpc support as that is no longer in Debian stable.

>>>

>>

>> Applied, thanks.

>

> Yay! Shippable happily working again \o/

>

> https://app.shippable.com/github/qemu/qemu/runs/238/summary/console

>

>   Overall Status: Success


Now we've been running shippable for a while is it worth turning on the
IRC notifications?

--
Alex Bennée
Philippe Mathieu-Daudé June 22, 2017, 3:42 p.m. UTC | #4
On Thu, Jun 22, 2017 at 12:13 PM, Alex Bennée <alex.bennee@linaro.org> wrote:
> Now we've been running shippable for a while is it worth turning on the

> IRC notifications?


What about moving such bot/scripts notifications in another channel
like #QEMU-notifications?
It is often hard to follow 3 concurrent topics while the travis-ci bot
disrupts entering/broadcasting/leaving...
I'd also like to see here checkpatch/patchew stripped output, and
eventually gcov/gprof reports.
I don't think full patchew reports are useful in a mailbox.
Philippe Mathieu-Daudé July 10, 2017, 12:04 p.m. UTC | #5
On Thu, Jun 22, 2017 at 12:42 PM, Philippe Mathieu-Daudé
<f4bug@amsat.org> wrote:
> On Thu, Jun 22, 2017 at 12:13 PM, Alex Bennée <alex.bennee@linaro.org> wrote:

>> Now we've been running shippable for a while is it worth turning on the

>> IRC notifications?

>

> What about moving such bot/scripts notifications in another channel

> like #QEMU-notifications?


ping?

> It is often hard to follow 3 concurrent topics while the travis-ci bot

> disrupts entering/broadcasting/leaving...

> I'd also like to see here checkpatch/patchew stripped output, and

> eventually gcov/gprof reports.

> I don't think full patchew reports are useful in a mailbox.
Alex Bennée July 10, 2017, 2:17 p.m. UTC | #6
Philippe Mathieu-Daudé <f4bug@amsat.org> writes:

> On Thu, Jun 22, 2017 at 12:42 PM, Philippe Mathieu-Daudé

> <f4bug@amsat.org> wrote:

>> On Thu, Jun 22, 2017 at 12:13 PM, Alex Bennée <alex.bennee@linaro.org> wrote:

>>> Now we've been running shippable for a while is it worth turning on the

>>> IRC notifications?

>>

>> What about moving such bot/scripts notifications in another channel

>> like #QEMU-notifications?

>

> ping?


I'm in two minds so I suspect this is something that needs to be taken
to the users of #qemu.

>> It is often hard to follow 3 concurrent topics while the travis-ci bot

>> disrupts entering/broadcasting/leaving...


I find it useful as its somewhere I look at a lot. That said it is not
exactly a trial to sub to yet another IRC channel.

>> I'd also like to see here checkpatch/patchew stripped output, and

>> eventually gcov/gprof reports.

>> I don't think full patchew reports are useful in a mailbox.


I disagree. They could certainly be made terser but I think the general
workflow of posting automated testing reports to patch series is a
useful one.

--
Alex Bennée
Peter Maydell July 10, 2017, 2:22 p.m. UTC | #7
On 10 July 2017 at 15:17, Alex Bennée <alex.bennee@linaro.org> wrote:
> Philippe Mathieu-Daudé <f4bug@amsat.org> writes:

>>> I don't think full patchew reports are useful in a mailbox.

>

> I disagree. They could certainly be made terser but I think the general

> workflow of posting automated testing reports to patch series is a

> useful one.


Agreed. What we want is a mode for the makefile where
it doesn't even print the "   CC foo.o" lines (except
perhaps for the case where it's going to print an error
message, if possible, but that might not be doable), and
then we can use that for patchew emails and also maybe
for the Travis builds, whose logs are currently too long
for the web viewer to cope with (might run into issues with
travis thinking the build has stalled though.)

Is there a standard-ish convention for a really-silent
build we can borrow from the linux kernel or somewhere?

thanks
-- PMM
Philippe Mathieu-Daudé July 12, 2017, 12:17 p.m. UTC | #8
On Mon, Jul 10, 2017 at 11:22 AM, Peter Maydell
<peter.maydell@linaro.org> wrote:
> On 10 July 2017 at 15:17, Alex Bennée <alex.bennee@linaro.org> wrote:

>> Philippe Mathieu-Daudé <f4bug@amsat.org> writes:

>>>> I don't think full patchew reports are useful in a mailbox.

>>

>> I disagree. They could certainly be made terser but I think the general

>> workflow of posting automated testing reports to patch series is a

>> useful one.

>

> Agreed. What we want is a mode for the makefile where

> it doesn't even print the "   CC foo.o" lines (except

> perhaps for the case where it's going to print an error

> message, if possible, but that might not be doable), and

> then we can use that for patchew emails and also maybe

> for the Travis builds, whose logs are currently too long

> for the web viewer to cope with (might run into issues with

> travis thinking the build has stalled though.)


Using webmail (gmail) you get:

---
Hi,

This series failed automatic build test. Please find the testing commands and
their output below. If you have docker installed, you can probably reproduce it
locally.
[...]
  CC      aarch64-softmmu/hw/scsi/virtio-scsi.o
  CC      arm-softmmu/hw/net/vhost_net.o
  CC      aarch64-softmmu/hw/scsi/virtio-scsi-dataplane.o
  CC      arm-softmmu/hw/pcmcia/...

[Message clipped]  View entire message
---

also this one 132KB alone which makes thunderbird feature "filter
messages [Body]" slow to unresponsive:

---
Hi,

This series failed build test on s390x host. Please find the details below.

Type: series
Message-id: 1496820931-27416-10-git-send-email-a.perevalov@samsung.com
Subject: [Qemu-devel] [[PATCH V7] 09/11] migration: calculate vCPU
blocktime on dst side
---

I'm happy with the script output workflow, just say it could be more
useful if we can shorter it to focus on the problem, tailing last 500
stderr lines should be more than enough.