Add -w or --workdir to compose run to override workdir from commandline

[simonvanderveldt]

This adds the possibility to override the workdir from the command line, using either -w or --workdir just like docker run.
In essence it does the same as #332, though that code couldn't really be rebased so I wrote my own. Do we want to attribute anything of this to @scottynomad?

Also not really sure about the acceptance tests, there seem to be multiple different ways of doing them within tests/acceptance/cli_test.py. I basically mirrored the -u / --user tests.

Relevant for #363
Fixes#2811

[simonvanderveldt]

@GordonTheTurtle doesn't seem to be happy because the DinD tag used in the test is missing

++ docker run -d --name compose-dind-1.10.2-rc1-2391 --privileged --volume=/var/lib/docker dockerswarm/dind:1.10.2-rc1 docker daemon -H tcp://0.0.0.0:2375 --storage-driver=aufs
Unable to find image 'dockerswarm/dind:1.10.2-rc1' locally
Pulling repository docker.io/dockerswarm/dind
Tag 1.10.2-rc1 not found in repository docker.io/dockerswarm/dind

[dnephin]

Edit: Oops, nevermind, this is different

I'm fixing the dind image build now

[simonvanderveldt]

I noticed there were markdown files documenting the commands as well, so I've updated docs/reference/run.md to reflect the change in this PR.

The test on Janky now fails on the flake8 test, which works fine locally. Not sure why this is happening.
Note the test on Janky passed before I added the markdown file, so I guess something is broken/has changed in the test?

[dnephin]

I've got a fix for the max-complexity error in this commit: dnephin@8246c2a

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

[simonvanderveldt]

I've got a fix for the max-complexity error in this commit: dnephin/compose@8246c2a

Thanks!
Do you think the build_container_options() could be mistaken for docker build options?

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

Done, works fine locally, hope Janky is happy this time :)

[dnephin]

I just noticed, this should match what docker run --help says "Working directory inside the container"

[simonvanderveldt]

I just noticed, this should match what docker run --help says "Working directory inside the container"

Done, thx for the hint!

[dnephin]

With the fix to the help text LGTM

[dnephin]

LGTM

[simonvanderveldt]

Janky errors seem to be unrelated?

____________________________ CLITestCase.test_kill _____________________________
tests/acceptance/cli_test.py:1163: in test_kill
self.dispatch(['up', '-d'], None)
tests/acceptance/cli_test.py:128: in dispatch
return wait_on_process(proc, returncode=returncode)
tests/acceptance/cli_test.py:48: in wait_on_process
assert proc.returncode == returncode
E AssertionError: assert 1 == 0
E + where 1 = <subprocess.Popen object at 0x7f30c77aacd0>.returncode
----------------------------- Captured stdout call -----------------------------
Running process: 205
Stderr: Creating simplecomposefile_simple_1
Starting simplecomposefile_another_1
Container is marked for removal and cannot be started.
Stdout: ----------------------------- Captured stderr call -----------------------------
Killing simplecomposefile_simple_1 ... Killing simplecomposefile_simple_1 ... done
Removing simplecomposefile_simple_1 ... Removing simplecomposefile_another_1 ... Removing simplecomposefile_simple_1 ... done
Removing simplecomposefile_another_1 ... error
ERROR: for simplecomposefile_another_1 Driver aufs failed to remove root filesystem f75f290d301456234f1d37c18bd63254a3f6bd48c3efcc2b099768760c071b3c: rename /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8 /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8-removing: device or resource busy 

[dnephin]

yup, unrelated

[dnephin]

Needs a rebase to be merged

[simonvanderveldt]

Needs a rebase to be merged

@dnephin When will you need it to be rebased? I'm on holidays till the 25th of march :)

[dnephin]

Ok, don't worry about it. I will pick this up and carry the commits for the release. Code freeze is this week.

[dnephin]

Carried in #3136

[simonvanderveldt]

Awesome, thanks!

[simonvanderveldt]

This adds the possibility to override the workdir from the command line, using either -w or --workdir just like docker run.
In essence it does the same as #332, though that code couldn't really be rebased so I wrote my own. Do we want to attribute anything of this to @scottynomad?

Also not really sure about the acceptance tests, there seem to be multiple different ways of doing them within tests/acceptance/cli_test.py. I basically mirrored the -u / --user tests.

Relevant for #363
Fixes#2811

[simonvanderveldt]

@GordonTheTurtle doesn't seem to be happy because the DinD tag used in the test is missing

++ docker run -d --name compose-dind-1.10.2-rc1-2391 --privileged --volume=/var/lib/docker dockerswarm/dind:1.10.2-rc1 docker daemon -H tcp://0.0.0.0:2375 --storage-driver=aufs
Unable to find image 'dockerswarm/dind:1.10.2-rc1' locally
Pulling repository docker.io/dockerswarm/dind
Tag 1.10.2-rc1 not found in repository docker.io/dockerswarm/dind

[dnephin]

Edit: Oops, nevermind, this is different

I'm fixing the dind image build now

[simonvanderveldt]

I noticed there were markdown files documenting the commands as well, so I've updated docs/reference/run.md to reflect the change in this PR.

The test on Janky now fails on the flake8 test, which works fine locally. Not sure why this is happening.
Note the test on Janky passed before I added the markdown file, so I guess something is broken/has changed in the test?

[dnephin]

I've got a fix for the max-complexity error in this commit: dnephin@8246c2a

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

[simonvanderveldt]

I've got a fix for the max-complexity error in this commit: dnephin/compose@8246c2a

Thanks!
Do you think the build_container_options() could be mistaken for docker build options?

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

Done, works fine locally, hope Janky is happy this time :)

[dnephin]

I just noticed, this should match what docker run --help says "Working directory inside the container"

[simonvanderveldt]

I just noticed, this should match what docker run --help says "Working directory inside the container"

Done, thx for the hint!

[dnephin]

With the fix to the help text LGTM

[dnephin]

LGTM

[simonvanderveldt]

Janky errors seem to be unrelated?

____________________________ CLITestCase.test_kill _____________________________
tests/acceptance/cli_test.py:1163: in test_kill
self.dispatch(['up', '-d'], None)
tests/acceptance/cli_test.py:128: in dispatch
return wait_on_process(proc, returncode=returncode)
tests/acceptance/cli_test.py:48: in wait_on_process
assert proc.returncode == returncode
E AssertionError: assert 1 == 0
E + where 1 = <subprocess.Popen object at 0x7f30c77aacd0>.returncode
----------------------------- Captured stdout call -----------------------------
Running process: 205
Stderr: Creating simplecomposefile_simple_1
Starting simplecomposefile_another_1
Container is marked for removal and cannot be started.
Stdout: ----------------------------- Captured stderr call -----------------------------
Killing simplecomposefile_simple_1 ... Killing simplecomposefile_simple_1 ... done
Removing simplecomposefile_simple_1 ... Removing simplecomposefile_another_1 ... Removing simplecomposefile_simple_1 ... done
Removing simplecomposefile_another_1 ... error
ERROR: for simplecomposefile_another_1 Driver aufs failed to remove root filesystem f75f290d301456234f1d37c18bd63254a3f6bd48c3efcc2b099768760c071b3c: rename /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8 /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8-removing: device or resource busy 

[dnephin]

yup, unrelated

[dnephin]

Needs a rebase to be merged

[simonvanderveldt]

Needs a rebase to be merged

@dnephin When will you need it to be rebased? I'm on holidays till the 25th of march :)

[dnephin]

Ok, don't worry about it. I will pick this up and carry the commits for the release. Code freeze is this week.

[dnephin]

Carried in #3136

[simonvanderveldt]

Awesome, thanks!

[simonvanderveldt]

This adds the possibility to override the workdir from the command line, using either -w or --workdir just like docker run.
In essence it does the same as #332, though that code couldn't really be rebased so I wrote my own. Do we want to attribute anything of this to @scottynomad?

Also not really sure about the acceptance tests, there seem to be multiple different ways of doing them within tests/acceptance/cli_test.py. I basically mirrored the -u / --user tests.

Relevant for #363
Fixes#2811

[simonvanderveldt]

@GordonTheTurtle doesn't seem to be happy because the DinD tag used in the test is missing

++ docker run -d --name compose-dind-1.10.2-rc1-2391 --privileged --volume=/var/lib/docker dockerswarm/dind:1.10.2-rc1 docker daemon -H tcp://0.0.0.0:2375 --storage-driver=aufs
Unable to find image 'dockerswarm/dind:1.10.2-rc1' locally
Pulling repository docker.io/dockerswarm/dind
Tag 1.10.2-rc1 not found in repository docker.io/dockerswarm/dind

[dnephin]

Edit: Oops, nevermind, this is different

I'm fixing the dind image build now

[simonvanderveldt]

I noticed there were markdown files documenting the commands as well, so I've updated docs/reference/run.md to reflect the change in this PR.

The test on Janky now fails on the flake8 test, which works fine locally. Not sure why this is happening.
Note the test on Janky passed before I added the markdown file, so I guess something is broken/has changed in the test?

[dnephin]

I've got a fix for the max-complexity error in this commit: dnephin@8246c2a

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

[simonvanderveldt]

I've got a fix for the max-complexity error in this commit: dnephin/compose@8246c2a

Thanks!
Do you think the build_container_options() could be mistaken for docker build options?

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

Done, works fine locally, hope Janky is happy this time :)

[dnephin]

I just noticed, this should match what docker run --help says "Working directory inside the container"

[simonvanderveldt]

I just noticed, this should match what docker run --help says "Working directory inside the container"

Done, thx for the hint!

[dnephin]

With the fix to the help text LGTM

[dnephin]

LGTM

[simonvanderveldt]

Janky errors seem to be unrelated?

____________________________ CLITestCase.test_kill _____________________________
tests/acceptance/cli_test.py:1163: in test_kill
self.dispatch(['up', '-d'], None)
tests/acceptance/cli_test.py:128: in dispatch
return wait_on_process(proc, returncode=returncode)
tests/acceptance/cli_test.py:48: in wait_on_process
assert proc.returncode == returncode
E AssertionError: assert 1 == 0
E + where 1 = <subprocess.Popen object at 0x7f30c77aacd0>.returncode
----------------------------- Captured stdout call -----------------------------
Running process: 205
Stderr: Creating simplecomposefile_simple_1
Starting simplecomposefile_another_1
Container is marked for removal and cannot be started.
Stdout: ----------------------------- Captured stderr call -----------------------------
Killing simplecomposefile_simple_1 ... Killing simplecomposefile_simple_1 ... done
Removing simplecomposefile_simple_1 ... Removing simplecomposefile_another_1 ... Removing simplecomposefile_simple_1 ... done
Removing simplecomposefile_another_1 ... error
ERROR: for simplecomposefile_another_1 Driver aufs failed to remove root filesystem f75f290d301456234f1d37c18bd63254a3f6bd48c3efcc2b099768760c071b3c: rename /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8 /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8-removing: device or resource busy 

[dnephin]

yup, unrelated

[dnephin]

Needs a rebase to be merged

[simonvanderveldt]

Needs a rebase to be merged

@dnephin When will you need it to be rebased? I'm on holidays till the 25th of march :)

[dnephin]

Ok, don't worry about it. I will pick this up and carry the commits for the release. Code freeze is this week.

[dnephin]

Carried in #3136

[simonvanderveldt]

Awesome, thanks!

[simonvanderveldt]

This adds the possibility to override the workdir from the command line, using either -w or --workdir just like docker run.
In essence it does the same as #332, though that code couldn't really be rebased so I wrote my own. Do we want to attribute anything of this to @scottynomad?

Also not really sure about the acceptance tests, there seem to be multiple different ways of doing them within tests/acceptance/cli_test.py. I basically mirrored the -u / --user tests.

Relevant for #363
Fixes#2811

[simonvanderveldt]

@GordonTheTurtle doesn't seem to be happy because the DinD tag used in the test is missing

++ docker run -d --name compose-dind-1.10.2-rc1-2391 --privileged --volume=/var/lib/docker dockerswarm/dind:1.10.2-rc1 docker daemon -H tcp://0.0.0.0:2375 --storage-driver=aufs
Unable to find image 'dockerswarm/dind:1.10.2-rc1' locally
Pulling repository docker.io/dockerswarm/dind
Tag 1.10.2-rc1 not found in repository docker.io/dockerswarm/dind

[dnephin]

Edit: Oops, nevermind, this is different

I'm fixing the dind image build now

[simonvanderveldt]

I noticed there were markdown files documenting the commands as well, so I've updated docs/reference/run.md to reflect the change in this PR.

The test on Janky now fails on the flake8 test, which works fine locally. Not sure why this is happening.
Note the test on Janky passed before I added the markdown file, so I guess something is broken/has changed in the test?

[dnephin]

I've got a fix for the max-complexity error in this commit: dnephin@8246c2a

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

[simonvanderveldt]

I've got a fix for the max-complexity error in this commit: dnephin/compose@8246c2a

Thanks!
Do you think the build_container_options() could be mistaken for docker build options?

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

Done, works fine locally, hope Janky is happy this time :)

[dnephin]

I just noticed, this should match what docker run --help says "Working directory inside the container"

[simonvanderveldt]

I just noticed, this should match what docker run --help says "Working directory inside the container"

Done, thx for the hint!

[dnephin]

With the fix to the help text LGTM

[dnephin]

LGTM

[simonvanderveldt]

Janky errors seem to be unrelated?

____________________________ CLITestCase.test_kill _____________________________
tests/acceptance/cli_test.py:1163: in test_kill
self.dispatch(['up', '-d'], None)
tests/acceptance/cli_test.py:128: in dispatch
return wait_on_process(proc, returncode=returncode)
tests/acceptance/cli_test.py:48: in wait_on_process
assert proc.returncode == returncode
E AssertionError: assert 1 == 0
E + where 1 = <subprocess.Popen object at 0x7f30c77aacd0>.returncode
----------------------------- Captured stdout call -----------------------------
Running process: 205
Stderr: Creating simplecomposefile_simple_1
Starting simplecomposefile_another_1
Container is marked for removal and cannot be started.
Stdout: ----------------------------- Captured stderr call -----------------------------
Killing simplecomposefile_simple_1 ... Killing simplecomposefile_simple_1 ... done
Removing simplecomposefile_simple_1 ... Removing simplecomposefile_another_1 ... Removing simplecomposefile_simple_1 ... done
Removing simplecomposefile_another_1 ... error
ERROR: for simplecomposefile_another_1 Driver aufs failed to remove root filesystem f75f290d301456234f1d37c18bd63254a3f6bd48c3efcc2b099768760c071b3c: rename /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8 /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8-removing: device or resource busy 

[dnephin]

yup, unrelated

[dnephin]

Needs a rebase to be merged

[simonvanderveldt]

Needs a rebase to be merged

@dnephin When will you need it to be rebased? I'm on holidays till the 25th of march :)

[dnephin]

Ok, don't worry about it. I will pick this up and carry the commits for the release. Code freeze is this week.

[dnephin]

Carried in #3136

[simonvanderveldt]

Awesome, thanks!

[simonvanderveldt]

This adds the possibility to override the workdir from the command line, using either -w or --workdir just like docker run.
In essence it does the same as #332, though that code couldn't really be rebased so I wrote my own. Do we want to attribute anything of this to @scottynomad?

Also not really sure about the acceptance tests, there seem to be multiple different ways of doing them within tests/acceptance/cli_test.py. I basically mirrored the -u / --user tests.

Relevant for #363
Fixes#2811

[simonvanderveldt]

@GordonTheTurtle doesn't seem to be happy because the DinD tag used in the test is missing

++ docker run -d --name compose-dind-1.10.2-rc1-2391 --privileged --volume=/var/lib/docker dockerswarm/dind:1.10.2-rc1 docker daemon -H tcp://0.0.0.0:2375 --storage-driver=aufs
Unable to find image 'dockerswarm/dind:1.10.2-rc1' locally
Pulling repository docker.io/dockerswarm/dind
Tag 1.10.2-rc1 not found in repository docker.io/dockerswarm/dind

[dnephin]

Edit: Oops, nevermind, this is different

I'm fixing the dind image build now

[simonvanderveldt]

I noticed there were markdown files documenting the commands as well, so I've updated docs/reference/run.md to reflect the change in this PR.

The test on Janky now fails on the flake8 test, which works fine locally. Not sure why this is happening.
Note the test on Janky passed before I added the markdown file, so I guess something is broken/has changed in the test?

[dnephin]

I've got a fix for the max-complexity error in this commit: dnephin@8246c2a

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

[simonvanderveldt]

I've got a fix for the max-complexity error in this commit: dnephin/compose@8246c2a

Thanks!
Do you think the build_container_options() could be mistaken for docker build options?

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

Done, works fine locally, hope Janky is happy this time :)

[dnephin]

I just noticed, this should match what docker run --help says "Working directory inside the container"

[simonvanderveldt]

I just noticed, this should match what docker run --help says "Working directory inside the container"

Done, thx for the hint!

[dnephin]

With the fix to the help text LGTM

[dnephin]

LGTM

[simonvanderveldt]

Janky errors seem to be unrelated?

____________________________ CLITestCase.test_kill _____________________________
tests/acceptance/cli_test.py:1163: in test_kill
self.dispatch(['up', '-d'], None)
tests/acceptance/cli_test.py:128: in dispatch
return wait_on_process(proc, returncode=returncode)
tests/acceptance/cli_test.py:48: in wait_on_process
assert proc.returncode == returncode
E AssertionError: assert 1 == 0
E + where 1 = <subprocess.Popen object at 0x7f30c77aacd0>.returncode
----------------------------- Captured stdout call -----------------------------
Running process: 205
Stderr: Creating simplecomposefile_simple_1
Starting simplecomposefile_another_1
Container is marked for removal and cannot be started.
Stdout: ----------------------------- Captured stderr call -----------------------------
Killing simplecomposefile_simple_1 ... Killing simplecomposefile_simple_1 ... done
Removing simplecomposefile_simple_1 ... Removing simplecomposefile_another_1 ... Removing simplecomposefile_simple_1 ... done
Removing simplecomposefile_another_1 ... error
ERROR: for simplecomposefile_another_1 Driver aufs failed to remove root filesystem f75f290d301456234f1d37c18bd63254a3f6bd48c3efcc2b099768760c071b3c: rename /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8 /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8-removing: device or resource busy 

[dnephin]

yup, unrelated

[dnephin]

Needs a rebase to be merged

[simonvanderveldt]

Needs a rebase to be merged

@dnephin When will you need it to be rebased? I'm on holidays till the 25th of march :)

[dnephin]

Ok, don't worry about it. I will pick this up and carry the commits for the release. Code freeze is this week.

[dnephin]

Carried in #3136

[simonvanderveldt]

Awesome, thanks!

[simonvanderveldt]

This adds the possibility to override the workdir from the command line, using either -w or --workdir just like docker run.
In essence it does the same as #332, though that code couldn't really be rebased so I wrote my own. Do we want to attribute anything of this to @scottynomad?

Also not really sure about the acceptance tests, there seem to be multiple different ways of doing them within tests/acceptance/cli_test.py. I basically mirrored the -u / --user tests.

Relevant for #363
Fixes#2811

[simonvanderveldt]

@GordonTheTurtle doesn't seem to be happy because the DinD tag used in the test is missing

++ docker run -d --name compose-dind-1.10.2-rc1-2391 --privileged --volume=/var/lib/docker dockerswarm/dind:1.10.2-rc1 docker daemon -H tcp://0.0.0.0:2375 --storage-driver=aufs
Unable to find image 'dockerswarm/dind:1.10.2-rc1' locally
Pulling repository docker.io/dockerswarm/dind
Tag 1.10.2-rc1 not found in repository docker.io/dockerswarm/dind

[dnephin]

Edit: Oops, nevermind, this is different

I'm fixing the dind image build now

[simonvanderveldt]

I noticed there were markdown files documenting the commands as well, so I've updated docs/reference/run.md to reflect the change in this PR.

The test on Janky now fails on the flake8 test, which works fine locally. Not sure why this is happening.
Note the test on Janky passed before I added the markdown file, so I guess something is broken/has changed in the test?

[dnephin]

I've got a fix for the max-complexity error in this commit: dnephin@8246c2a

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

[simonvanderveldt]

I've got a fix for the max-complexity error in this commit: dnephin/compose@8246c2a

Thanks!
Do you think the build_container_options() could be mistaken for docker build options?

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

Done, works fine locally, hope Janky is happy this time :)

[dnephin]

I just noticed, this should match what docker run --help says "Working directory inside the container"

[simonvanderveldt]

I just noticed, this should match what docker run --help says "Working directory inside the container"

Done, thx for the hint!

[dnephin]

With the fix to the help text LGTM

[dnephin]

LGTM

[simonvanderveldt]

Janky errors seem to be unrelated?

____________________________ CLITestCase.test_kill _____________________________
tests/acceptance/cli_test.py:1163: in test_kill
self.dispatch(['up', '-d'], None)
tests/acceptance/cli_test.py:128: in dispatch
return wait_on_process(proc, returncode=returncode)
tests/acceptance/cli_test.py:48: in wait_on_process
assert proc.returncode == returncode
E AssertionError: assert 1 == 0
E + where 1 = <subprocess.Popen object at 0x7f30c77aacd0>.returncode
----------------------------- Captured stdout call -----------------------------
Running process: 205
Stderr: Creating simplecomposefile_simple_1
Starting simplecomposefile_another_1
Container is marked for removal and cannot be started.
Stdout: ----------------------------- Captured stderr call -----------------------------
Killing simplecomposefile_simple_1 ... Killing simplecomposefile_simple_1 ... done
Removing simplecomposefile_simple_1 ... Removing simplecomposefile_another_1 ... Removing simplecomposefile_simple_1 ... done
Removing simplecomposefile_another_1 ... error
ERROR: for simplecomposefile_another_1 Driver aufs failed to remove root filesystem f75f290d301456234f1d37c18bd63254a3f6bd48c3efcc2b099768760c071b3c: rename /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8 /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8-removing: device or resource busy 

[dnephin]

yup, unrelated

[dnephin]

Needs a rebase to be merged

[simonvanderveldt]

Needs a rebase to be merged

@dnephin When will you need it to be rebased? I'm on holidays till the 25th of march :)

[dnephin]

Ok, don't worry about it. I will pick this up and carry the commits for the release. Code freeze is this week.

[dnephin]

Carried in #3136

[simonvanderveldt]

Awesome, thanks!

[simonvanderveldt]

This adds the possibility to override the workdir from the command line, using either -w or --workdir just like docker run.
In essence it does the same as #332, though that code couldn't really be rebased so I wrote my own. Do we want to attribute anything of this to @scottynomad?

Also not really sure about the acceptance tests, there seem to be multiple different ways of doing them within tests/acceptance/cli_test.py. I basically mirrored the -u / --user tests.

Relevant for #363
Fixes#2811

[simonvanderveldt]

@GordonTheTurtle doesn't seem to be happy because the DinD tag used in the test is missing

++ docker run -d --name compose-dind-1.10.2-rc1-2391 --privileged --volume=/var/lib/docker dockerswarm/dind:1.10.2-rc1 docker daemon -H tcp://0.0.0.0:2375 --storage-driver=aufs
Unable to find image 'dockerswarm/dind:1.10.2-rc1' locally
Pulling repository docker.io/dockerswarm/dind
Tag 1.10.2-rc1 not found in repository docker.io/dockerswarm/dind

[dnephin]

Edit: Oops, nevermind, this is different

I'm fixing the dind image build now

[simonvanderveldt]

I noticed there were markdown files documenting the commands as well, so I've updated docs/reference/run.md to reflect the change in this PR.

The test on Janky now fails on the flake8 test, which works fine locally. Not sure why this is happening.
Note the test on Janky passed before I added the markdown file, so I guess something is broken/has changed in the test?

[dnephin]

I've got a fix for the max-complexity error in this commit: dnephin@8246c2a

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

[simonvanderveldt]

I've got a fix for the max-complexity error in this commit: dnephin/compose@8246c2a

Thanks!
Do you think the build_container_options() could be mistaken for docker build options?

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

Done, works fine locally, hope Janky is happy this time :)

[dnephin]

I just noticed, this should match what docker run --help says "Working directory inside the container"

[simonvanderveldt]

I just noticed, this should match what docker run --help says "Working directory inside the container"

Done, thx for the hint!

[dnephin]

With the fix to the help text LGTM

[dnephin]

LGTM

[simonvanderveldt]

Janky errors seem to be unrelated?

____________________________ CLITestCase.test_kill _____________________________
tests/acceptance/cli_test.py:1163: in test_kill
self.dispatch(['up', '-d'], None)
tests/acceptance/cli_test.py:128: in dispatch
return wait_on_process(proc, returncode=returncode)
tests/acceptance/cli_test.py:48: in wait_on_process
assert proc.returncode == returncode
E AssertionError: assert 1 == 0
E + where 1 = <subprocess.Popen object at 0x7f30c77aacd0>.returncode
----------------------------- Captured stdout call -----------------------------
Running process: 205
Stderr: Creating simplecomposefile_simple_1
Starting simplecomposefile_another_1
Container is marked for removal and cannot be started.
Stdout: ----------------------------- Captured stderr call -----------------------------
Killing simplecomposefile_simple_1 ... Killing simplecomposefile_simple_1 ... done
Removing simplecomposefile_simple_1 ... Removing simplecomposefile_another_1 ... Removing simplecomposefile_simple_1 ... done
Removing simplecomposefile_another_1 ... error
ERROR: for simplecomposefile_another_1 Driver aufs failed to remove root filesystem f75f290d301456234f1d37c18bd63254a3f6bd48c3efcc2b099768760c071b3c: rename /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8 /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8-removing: device or resource busy 

[dnephin]

yup, unrelated

[dnephin]

Needs a rebase to be merged

[simonvanderveldt]

Needs a rebase to be merged

@dnephin When will you need it to be rebased? I'm on holidays till the 25th of march :)

[dnephin]

Ok, don't worry about it. I will pick this up and carry the commits for the release. Code freeze is this week.

[dnephin]

Carried in #3136

[simonvanderveldt]

Awesome, thanks!

[simonvanderveldt]

This adds the possibility to override the workdir from the command line, using either -w or --workdir just like docker run.
In essence it does the same as #332, though that code couldn't really be rebased so I wrote my own. Do we want to attribute anything of this to @scottynomad?

Also not really sure about the acceptance tests, there seem to be multiple different ways of doing them within tests/acceptance/cli_test.py. I basically mirrored the -u / --user tests.

Relevant for #363
Fixes#2811

[simonvanderveldt]

@GordonTheTurtle doesn't seem to be happy because the DinD tag used in the test is missing

++ docker run -d --name compose-dind-1.10.2-rc1-2391 --privileged --volume=/var/lib/docker dockerswarm/dind:1.10.2-rc1 docker daemon -H tcp://0.0.0.0:2375 --storage-driver=aufs
Unable to find image 'dockerswarm/dind:1.10.2-rc1' locally
Pulling repository docker.io/dockerswarm/dind
Tag 1.10.2-rc1 not found in repository docker.io/dockerswarm/dind

[dnephin]

Edit: Oops, nevermind, this is different

I'm fixing the dind image build now

[simonvanderveldt]

I noticed there were markdown files documenting the commands as well, so I've updated docs/reference/run.md to reflect the change in this PR.

The test on Janky now fails on the flake8 test, which works fine locally. Not sure why this is happening.
Note the test on Janky passed before I added the markdown file, so I guess something is broken/has changed in the test?

[dnephin]

I've got a fix for the max-complexity error in this commit: dnephin@8246c2a

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

[simonvanderveldt]

I've got a fix for the max-complexity error in this commit: dnephin/compose@8246c2a

Thanks!
Do you think the build_container_options() could be mistaken for docker build options?

I guess I rebased on master, so it doesn't merge cleanly on your branch, but you can cherry-pick this commit if you add my fork as a remote.

Done, works fine locally, hope Janky is happy this time :)

[dnephin]

I just noticed, this should match what docker run --help says "Working directory inside the container"

[simonvanderveldt]

I just noticed, this should match what docker run --help says "Working directory inside the container"

Done, thx for the hint!

[dnephin]

With the fix to the help text LGTM

[dnephin]

LGTM

[simonvanderveldt]

Janky errors seem to be unrelated?

____________________________ CLITestCase.test_kill _____________________________
tests/acceptance/cli_test.py:1163: in test_kill
self.dispatch(['up', '-d'], None)
tests/acceptance/cli_test.py:128: in dispatch
return wait_on_process(proc, returncode=returncode)
tests/acceptance/cli_test.py:48: in wait_on_process
assert proc.returncode == returncode
E AssertionError: assert 1 == 0
E + where 1 = <subprocess.Popen object at 0x7f30c77aacd0>.returncode
----------------------------- Captured stdout call -----------------------------
Running process: 205
Stderr: Creating simplecomposefile_simple_1
Starting simplecomposefile_another_1
Container is marked for removal and cannot be started.
Stdout: ----------------------------- Captured stderr call -----------------------------
Killing simplecomposefile_simple_1 ... Killing simplecomposefile_simple_1 ... done
Removing simplecomposefile_simple_1 ... Removing simplecomposefile_another_1 ... Removing simplecomposefile_simple_1 ... done
Removing simplecomposefile_another_1 ... error
ERROR: for simplecomposefile_another_1 Driver aufs failed to remove root filesystem f75f290d301456234f1d37c18bd63254a3f6bd48c3efcc2b099768760c071b3c: rename /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8 /var/lib/docker/aufs/mnt/0cb124ee817b6072798a141ba784ce4c79f427269198d412424bd8117fdd58e8-removing: device or resource busy 

[dnephin]

yup, unrelated

[dnephin]

Needs a rebase to be merged

[simonvanderveldt]

Needs a rebase to be merged

@dnephin When will you need it to be rebased? I'm on holidays till the 25th of march :)

[dnephin]

Ok, don't worry about it. I will pick this up and carry the commits for the release. Code freeze is this week.

[dnephin]

Carried in #3136

[simonvanderveldt]

Awesome, thanks!