[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH v2] docker: Don't enable networking as a side-ef
From: |
Fam Zheng |
Subject: |
Re: [Qemu-devel] [PATCH v2] docker: Don't enable networking as a side-effect of DEBUG=1 |
Date: |
Fri, 14 Jul 2017 19:50:03 +0800 |
User-agent: |
Mutt/1.8.3 (2017-05-23) |
On Fri, 07/14 12:41, Daniel P. Berrange wrote:
> On Fri, Jul 14, 2017 at 10:29:30AM +0800, Fam Zheng wrote:
> > On Thu, 07/13 15:43, Daniel P. Berrange wrote:
> > > When trying to debug problems with tests it is natural to set
> > > DEBUG=1 when starting the docker environment. Unfortunately
> > > this has a side-effect of enabling an eth0 network interface
> > > in the container, which changes the operating environment of
> > > the test suite. IOW tests with fail may suddenly start
> > > working again if DEBUG=1 is set, due to changed network setup.
> > >
> > > Add a separate NETWORK variable to allow enablement of
> > > networking separately from DEBUG=1. This can be used in two
> > > ways. To enable the default docker network backend
> > >
> > > make address@hidden NETWORK=1
> > >
> > > while to enable a specific network backend, eg join the network
> > > associated with the container 'wibble':
> > >
> > > make address@hidden NETWORK=container:wibble
> > >
> > > Signed-off-by: Daniel P. Berrange <address@hidden>
> > > ---
> > >
> > > Changed in v2:
> > >
> > > - Remove reference to eth0
> > > - Allow arbitrary docker network backend to be given
> > >
> > > tests/docker/Makefile.include | 5 ++++-
> > > 1 file changed, 4 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/tests/docker/Makefile.include b/tests/docker/Makefile.include
> > > index 037cb9e9e7..ec8e8d935d 100644
> > > --- a/tests/docker/Makefile.include
> > > +++ b/tests/docker/Makefile.include
> > > @@ -106,6 +106,8 @@ docker:
> > > @echo ' (default is 1)'
> > > @echo ' DEBUG=1 Stop and drop to shell in the created
> > > container'
> > > @echo ' before running the command.'
> > > + @echo ' NETWORK=1 Enable virtual network interface with
> > > default backend.'
> > > + @echo ' NETWORK=$BACKEND Enable virtual network interface with
> > > $BACKEND.'
> > > @echo ' NOUSER Define to disable adding current user
> > > to containers passwd.'
> > > @echo ' NOCACHE=1 Ignore cache when build images.'
> > > @echo ' EXECUTABLE=<path> Include executable in image.'
> > > @@ -132,7 +134,8 @@ docker-run: docker-qemu-src
> > > $(SRC_PATH)/tests/docker/docker.py run \
> > > $(if $(NOUSER),,-u $(shell id -u)) -t \
> > > $(if $V,,--rm) \
> > > - $(if $(DEBUG),-i,--net=none) \
> > > + $(if $(DEBUG),-i,) \
> > > + $(if $(NETWORK),$(if $(subst $(NETWORK),,1)$(subst
> > > 1,,$(NETWORK)),--net=$(NETWORK),),--net=none) \
> >
> > Isn't the first subst enough? We already know $(NETWORK) is non-empty. If
> > it is
> > "1", $(subst 1,,1) is empty; otherwise, $(subst foo,,1) is non-empty:
> >
> > $(if $(NETWORK),$(if $(subst
> > $(NETWORK),,1),,--net=$(NETWORK)),--net=none) \
>
> This looks inverted to me - NETWORK=1, will generate --net=1 and
> NETWORK=foo generates empty string.
You are right, I was confused (subst succeed is FALSE branch of if). I think
s/,,/,/ and it will work?
Fam