* [PATCH] travis: add initial Travis CI script to do builds
@ 2016-02-05 14:48 Doug Goldstein
2016-02-05 15:09 ` Wei Liu
2016-02-05 16:33 ` Ian Campbell
0 siblings, 2 replies; 6+ messages in thread
From: Doug Goldstein @ 2016-02-05 14:48 UTC (permalink / raw)
To: xen-devel
Cc: Keir Fraser, Ian Campbell, Tim Deegan, Doug Goldstein,
Jan Beulich, Ian Jackson
This is just suppose to do a simple compile test on Travis CI. Currently
due to linux86 (bcc/bin86/dev86) not being whitelisted the tools cannot
be built.
Signed-off-by: Doug Goldstein <cardoe@cardoe.com>
---
So this will work great if we get a regular cronjob or a post-receive hook
to push to GitHub (https://github.com/xen-project/xen) since Travis CI provides
free integration with public repositories. So an example run of this is
available at https://travis-ci.org/cardoe/xen/builds/107246160 if this gets
merged I will setup a xen-project space under Travis for us. Currently its
under my namespace because this file only exists in my branches.
The goal here is not to replace osstest by any means but to augment it by
providing some easy to do build tests on every revision and reporting back. It
should be possible in the future to potentially tie this into osstest to
allow this to build and if the tree for some reason failed to build not kick
off a test flight on osstest. Just some thoughts, I know I'm getting a bit
ahead of myself.
Other future tweaks I've got are some simple dom0 boot tests using qemu to
boot Xen and then boot a kernel and see that it came up. I've also got some
scripts to check the code base for code style consistency using clang-format.
If you're curious to see the output of these tests come to #xentest on
Freenode.
---
.travis.yml | 64 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 64 insertions(+)
create mode 100644 .travis.yml
diff --git a/.travis.yml b/.travis.yml
new file mode 100644
index 0000000..9020f16
--- /dev/null
+++ b/.travis.yml
@@ -0,0 +1,64 @@
+language: c
+dist: trusty
+sudo: required
+matrix:
+ allow_failures:
+ - compiler: clang
+ include:
+ - compiler: gcc
+ env: XEN_TARGET_ARCH=x86_64
+ - compiler: gcc
+ env: XEN_TARGET_ARCH=x86_64 COMPILER=gcc-5
+ - compiler: clang
+ env: XEN_TARGET_ARCH=x86_64 clang=y
+ - compiler: clang
+ env: XEN_TARGET_ARCH=x86_64 COMPILER=clang-3.8 clang=y
+ - compiler: gcc
+ env: XEN_TARGET_ARCH=arm32 CROSS_COMPILE=arm-linux-gnueabihf-
+ - compiler: gcc
+ env: XEN_TARGET_ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu-
+addons:
+ apt:
+ sources:
+ - ubuntu-toolchain-r-test
+ - llvm-toolchain-precise
+ packages:
+ - zlib1g-dev
+ - libncurses5-dev
+ - libssl-dev
+ - python2.7-dev
+ - xorg-dev
+ - uuid-dev
+ - libyajl-dev
+ - libaio-dev
+ - libglib2.0-dev
+ - libpixman-1-dev
+ - pkg-config
+ - bridge-utils
+ - iproute2
+ - flex
+ - bison
+ - gettext
+ - acpica-tools
+ - bin86
+ - bcc
+ - libc6-dev-i386
+ - libnl-3-dev
+ - ocaml-nox
+ - libfindlib-ocaml-dev
+ - markdown
+ - transfig
+ - pandoc
+ - gcc-arm-linux-gnueabihf
+ - gcc-aarch64-linux-gnu
+ - gcc-5
+ - clang-3.8
+before_script:
+ - export CC=${COMPILER:-${CC}}
+ - ${CC} --version
+script:
+ - ( ./configure --disable-tools --disable-stubdom --enable-docs &&
+ make CC="${CROSS_COMPILE}${CC}" HOSTCC="${CC}" dist )
+notifications:
+ irc:
+ - "chat.freenode.net#xentest"
--
2.4.10
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH] travis: add initial Travis CI script to do builds
2016-02-05 14:48 [PATCH] travis: add initial Travis CI script to do builds Doug Goldstein
@ 2016-02-05 15:09 ` Wei Liu
2016-02-05 15:19 ` Doug Goldstein
2016-02-05 16:33 ` Ian Campbell
1 sibling, 1 reply; 6+ messages in thread
From: Wei Liu @ 2016-02-05 15:09 UTC (permalink / raw)
To: Doug Goldstein
Cc: Keir Fraser, Ian Campbell, Tim Deegan, Ian Jackson, xen-devel,
Jan Beulich, wei.liu2
On Fri, Feb 05, 2016 at 08:48:49AM -0600, Doug Goldstein wrote:
> This is just suppose to do a simple compile test on Travis CI. Currently
> due to linux86 (bcc/bin86/dev86) not being whitelisted the tools cannot
> be built.
>
> Signed-off-by: Doug Goldstein <cardoe@cardoe.com>
> ---
>
> So this will work great if we get a regular cronjob or a post-receive hook
> to push to GitHub (https://github.com/xen-project/xen) since Travis CI provides
> free integration with public repositories. So an example run of this is
> available at https://travis-ci.org/cardoe/xen/builds/107246160 if this gets
> merged I will setup a xen-project space under Travis for us. Currently its
> under my namespace because this file only exists in my branches.
>
> The goal here is not to replace osstest by any means but to augment it by
> providing some easy to do build tests on every revision and reporting back. It
> should be possible in the future to potentially tie this into osstest to
> allow this to build and if the tree for some reason failed to build not kick
> off a test flight on osstest. Just some thoughts, I know I'm getting a bit
> ahead of myself.
>
I think adding travis CI is a good idea. It would be valuable if
individual developers can run their series through travis before
submitting. Thanks for your work.
> Other future tweaks I've got are some simple dom0 boot tests using qemu to
> boot Xen and then boot a kernel and see that it came up. I've also got some
> scripts to check the code base for code style consistency using clang-format.
>
> If you're curious to see the output of these tests come to #xentest on
> Freenode.
>
The purpose of #xentest is (or was) to coordinate test day activities,
I would rather use a different channel.
However, there has been idea floating that test day should be organised
in developer channels, and there is no test day activity at the moment,
so I think it would fine for now to use it for travis.
A stupid question: does this mean if I fork xen.git on github, travis
will pick up my fork and spam #xentest?
Wei.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] travis: add initial Travis CI script to do builds
2016-02-05 15:09 ` Wei Liu
@ 2016-02-05 15:19 ` Doug Goldstein
2016-02-05 15:24 ` Wei Liu
0 siblings, 1 reply; 6+ messages in thread
From: Doug Goldstein @ 2016-02-05 15:19 UTC (permalink / raw)
To: Wei Liu
Cc: Keir Fraser, Ian Campbell, Tim Deegan, Ian Jackson, xen-devel,
Jan Beulich
[-- Attachment #1.1: Type: text/plain, Size: 2395 bytes --]
On 2/5/16 9:09 AM, Wei Liu wrote:
> On Fri, Feb 05, 2016 at 08:48:49AM -0600, Doug Goldstein wrote:
>> This is just suppose to do a simple compile test on Travis CI. Currently
>> due to linux86 (bcc/bin86/dev86) not being whitelisted the tools cannot
>> be built.
>>
>> Signed-off-by: Doug Goldstein <cardoe@cardoe.com>
>> ---
>>
>> So this will work great if we get a regular cronjob or a post-receive hook
>> to push to GitHub (https://github.com/xen-project/xen) since Travis CI provides
>> free integration with public repositories. So an example run of this is
>> available at https://travis-ci.org/cardoe/xen/builds/107246160 if this gets
>> merged I will setup a xen-project space under Travis for us. Currently its
>> under my namespace because this file only exists in my branches.
>>
>> The goal here is not to replace osstest by any means but to augment it by
>> providing some easy to do build tests on every revision and reporting back. It
>> should be possible in the future to potentially tie this into osstest to
>> allow this to build and if the tree for some reason failed to build not kick
>> off a test flight on osstest. Just some thoughts, I know I'm getting a bit
>> ahead of myself.
>>
>
> I think adding travis CI is a good idea. It would be valuable if
> individual developers can run their series through travis before
> submitting. Thanks for your work.
>
>> Other future tweaks I've got are some simple dom0 boot tests using qemu to
>> boot Xen and then boot a kernel and see that it came up. I've also got some
>> scripts to check the code base for code style consistency using clang-format.
>>
>> If you're curious to see the output of these tests come to #xentest on
>> Freenode.
>>
>
> The purpose of #xentest is (or was) to coordinate test day activities,
> I would rather use a different channel.
>
> However, there has been idea floating that test day should be organised
> in developer channels, and there is no test day activity at the moment,
> so I think it would fine for now to use it for travis.
>
> A stupid question: does this mean if I fork xen.git on github, travis
> will pick up my fork and spam #xentest?
>
> Wei.
>
Hmm. Good point. It will. I'll drop this part of the patch. There might
be a way to properly filter this but I'll look into that as a follow on.
--
Doug Goldstein
[-- Attachment #1.2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 959 bytes --]
[-- Attachment #2: Type: text/plain, Size: 126 bytes --]
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] travis: add initial Travis CI script to do builds
2016-02-05 15:19 ` Doug Goldstein
@ 2016-02-05 15:24 ` Wei Liu
0 siblings, 0 replies; 6+ messages in thread
From: Wei Liu @ 2016-02-05 15:24 UTC (permalink / raw)
To: Doug Goldstein
Cc: Keir Fraser, Ian Campbell, Ian Jackson, Tim Deegan, xen-devel,
Jan Beulich, Wei Liu
On Fri, Feb 05, 2016 at 09:19:30AM -0600, Doug Goldstein wrote:
> On 2/5/16 9:09 AM, Wei Liu wrote:
> > On Fri, Feb 05, 2016 at 08:48:49AM -0600, Doug Goldstein wrote:
> >> This is just suppose to do a simple compile test on Travis CI. Currently
> >> due to linux86 (bcc/bin86/dev86) not being whitelisted the tools cannot
> >> be built.
> >>
> >> Signed-off-by: Doug Goldstein <cardoe@cardoe.com>
> >> ---
> >>
> >> So this will work great if we get a regular cronjob or a post-receive hook
> >> to push to GitHub (https://github.com/xen-project/xen) since Travis CI provides
> >> free integration with public repositories. So an example run of this is
> >> available at https://travis-ci.org/cardoe/xen/builds/107246160 if this gets
> >> merged I will setup a xen-project space under Travis for us. Currently its
> >> under my namespace because this file only exists in my branches.
> >>
> >> The goal here is not to replace osstest by any means but to augment it by
> >> providing some easy to do build tests on every revision and reporting back. It
> >> should be possible in the future to potentially tie this into osstest to
> >> allow this to build and if the tree for some reason failed to build not kick
> >> off a test flight on osstest. Just some thoughts, I know I'm getting a bit
> >> ahead of myself.
> >>
> >
> > I think adding travis CI is a good idea. It would be valuable if
> > individual developers can run their series through travis before
> > submitting. Thanks for your work.
> >
> >> Other future tweaks I've got are some simple dom0 boot tests using qemu to
> >> boot Xen and then boot a kernel and see that it came up. I've also got some
> >> scripts to check the code base for code style consistency using clang-format.
> >>
> >> If you're curious to see the output of these tests come to #xentest on
> >> Freenode.
> >>
> >
> > The purpose of #xentest is (or was) to coordinate test day activities,
> > I would rather use a different channel.
> >
> > However, there has been idea floating that test day should be organised
> > in developer channels, and there is no test day activity at the moment,
> > so I think it would fine for now to use it for travis.
> >
> > A stupid question: does this mean if I fork xen.git on github, travis
> > will pick up my fork and spam #xentest?
> >
> > Wei.
> >
>
> Hmm. Good point. It will. I'll drop this part of the patch. There might
> be a way to properly filter this but I'll look into that as a follow on.
>
I think the best option is for travis to use credential from the current
fork / current user. I *think* that should be supported one way or
another. But there is no need to rush of course. :-)
Wei.
> --
> Doug Goldstein
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] travis: add initial Travis CI script to do builds
2016-02-05 14:48 [PATCH] travis: add initial Travis CI script to do builds Doug Goldstein
2016-02-05 15:09 ` Wei Liu
@ 2016-02-05 16:33 ` Ian Campbell
2016-02-05 16:41 ` Doug Goldstein
1 sibling, 1 reply; 6+ messages in thread
From: Ian Campbell @ 2016-02-05 16:33 UTC (permalink / raw)
To: Doug Goldstein, xen-devel
Cc: Keir Fraser, Tim Deegan, Ian Jackson, Jan Beulich
On Fri, 2016-02-05 at 08:48 -0600, Doug Goldstein wrote:
>
> The goal here is not to replace osstest by any means but to augment it by
> providing some easy to do build tests on every revision and reporting back. It
> should be possible in the future to potentially tie this into osstest to
> allow this to build and if the tree for some reason failed to build not kick
> off a test flight on osstest. Just some thoughts, I know I'm getting a bit
> ahead of myself.
IMHO Free Software projects like Xen Project should be very wary of tying
their infrastructure and workflows to services which are backed by non-free
software (which I know github is and AFAICT Travis CI also is).
IOW while augmenting things to provide extra/non-critical data about the
state of the world might be ok or at least tolerable, projects should be
wary of inserting these tools into the "critical path" of their workflows.
I think inserting Travis CI before osstest would be just such a mistake.
https://mako.cc/writing/hill-free_tools.html is a good essay on the
subject.
Ian.
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] travis: add initial Travis CI script to do builds
2016-02-05 16:33 ` Ian Campbell
@ 2016-02-05 16:41 ` Doug Goldstein
0 siblings, 0 replies; 6+ messages in thread
From: Doug Goldstein @ 2016-02-05 16:41 UTC (permalink / raw)
To: Ian Campbell, xen-devel; +Cc: Keir Fraser, Tim Deegan, Ian Jackson, Jan Beulich
[-- Attachment #1.1: Type: text/plain, Size: 1599 bytes --]
On 2/5/16 10:33 AM, Ian Campbell wrote:
> On Fri, 2016-02-05 at 08:48 -0600, Doug Goldstein wrote:
>>
>> The goal here is not to replace osstest by any means but to augment it by
>> providing some easy to do build tests on every revision and reporting back. It
>> should be possible in the future to potentially tie this into osstest to
>> allow this to build and if the tree for some reason failed to build not kick
>> off a test flight on osstest. Just some thoughts, I know I'm getting a bit
>> ahead of myself.
>
> IMHO Free Software projects like Xen Project should be very wary of tying
> their infrastructure and workflows to services which are backed by non-free
> software (which I know github is and AFAICT Travis CI also is).
>
> IOW while augmenting things to provide extra/non-critical data about the
> state of the world might be ok or at least tolerable, projects should be
> wary of inserting these tools into the "critical path" of their workflows.
> I think inserting Travis CI before osstest would be just such a mistake.
>
> https://mako.cc/writing/hill-free_tools.html is a good essay on the
> subject.
>
> Ian.
>
That's fine. I was throwing out some ideas, we definitely don't have to
do the things I said. My stated just trying to use existing services to
provide some extra testing of Xen and not replace osstest, similarly to
Xen using Coverity which itself is not open source. I've been tinkering
with using llvm's scan-build to provide similar analysis but currently
the tree doesn't build cleanly with llvm.
--
Doug Goldstein
[-- Attachment #1.2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 959 bytes --]
[-- Attachment #2: Type: text/plain, Size: 126 bytes --]
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2016-02-05 16:41 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-02-05 14:48 [PATCH] travis: add initial Travis CI script to do builds Doug Goldstein
2016-02-05 15:09 ` Wei Liu
2016-02-05 15:19 ` Doug Goldstein
2016-02-05 15:24 ` Wei Liu
2016-02-05 16:33 ` Ian Campbell
2016-02-05 16:41 ` Doug Goldstein
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.