Re: ANN: scikit-image 0.10.0
Hi, On Monday, June 2, 2014 1:35:11 PM UTC-7, Stéfan van der Walt wrote:
Hi Matthew
I guess it would be good to build them automatically next time ... so maybe yes to include the scripts - should be only a few lines of shell I
On Mon, Jun 2, 2014 at 7:29 PM, Matthew Brett <matthe...@gmail.com <javascript:>> wrote: think,
once you've got the MacPythons installed.
I admit that I am out of my depth here. Would you be willing to help us on this one? Is it, e.g., possible to add a target to setup.py so that anyone on OSX can build these?
Er - I am not sure how that could sensibly be done in the setup.py, because you'll need to follow something like the recipe on the Spinning Wheels page, installing MacPythons, pip and packages, then running ``$PYTHON setup.py bdist_wheel`` for each Python. You're welcome to a login to one of our Berkeley machines where the basic setup is already done. In the Travis-CI example, how does Travis get triggered? (I presume
you don't commit to that repository each time you want to test MPL?)
Actually, yes, that is how I / we trigger builds at the moment, because we just got going with these megatesting things. There seem to be various options for triggering builds, including (I suppose) making fake commits, but I haven't investigated very far yet. See you, Matthew
On Monday, June 2, 2014 3:11:20 PM UTC-7, Matthew Brett wrote:
Hi,
On Monday, June 2, 2014 1:35:11 PM UTC-7, Stéfan van der Walt wrote:
Hi Matthew
I guess it would be good to build them automatically next time ... so maybe yes to include the scripts - should be only a few lines of shell I
On Mon, Jun 2, 2014 at 7:29 PM, Matthew Brett <matthe...@gmail.com> wrote: think,
once you've got the MacPythons installed.
I admit that I am out of my depth here. Would you be willing to help us on this one? Is it, e.g., possible to add a target to setup.py so that anyone on OSX can build these?
Er - I am not sure how that could sensibly be done in the setup.py, because you'll need to follow something like the recipe on the Spinning Wheels page, installing MacPythons, pip and packages, then running ``$PYTHON setup.py bdist_wheel`` for each Python.
You're welcome to a login to one of our Berkeley machines where the basic setup is already done.
In the Travis-CI example, how does Travis get triggered? (I presume
you don't commit to that repository each time you want to test MPL?)
Actually, yes, that is how I / we trigger builds at the moment, because we just got going with these megatesting things. There seem to be various options for triggering builds, including (I suppose) making fake commits, but I haven't investigated very far yet.
I just went through a quite painful learning process with all this,
because the docs are quite scattered, with lots of old stuff mixed in. So I'd like to comment, in case you don't insist on the Travis build machinery to be MacPython (is that even possible, as Travis runs on Linux?), to speed things up IMMENSELY over using pip for requirements install, one can use conda for the env setup in Travis. It speeds things up, because pip doesn't store binaries and instead spends a lot of time compiling things (or so it says here: http://sburns.org/2014/03/28/faster-travis-builds.html). Find the instructions on how to setup Travis with conda in that link, I tried it and it works fine here. I did one thing different compared to that blog entry though: In the travis yaml I prefer to "script: python setup.py develop && python setup.py test", because that way I don't get a useless test run in case the develop install fails due to a bug.
On Tue, Jun 3, 2014 at 1:01 AM, Michael Aye <kmichael.aye@gmail.com> wrote:
I just went through a quite painful learning process with all this, because the docs are quite scattered, with lots of old stuff mixed in. So I'd like to comment, in case you don't insist on the Travis build machinery to be MacPython (is that even possible, as Travis runs on Linux?), to speed things up IMMENSELY over using pip for requirements install, one can use conda for the env setup in Travis. It speeds things up, because pip doesn't store binaries and instead spends a lot of time compiling things (or so it says here: http://sburns.org/2014/03/28/faster-travis-builds.html). Find the instructions on how to setup Travis with conda in that link, I tried it and it works fine here. I did one thing different compared to that blog entry though: In the travis yaml I prefer to "script: python setup.py develop && python setup.py test", because that way I don't get a useless test run in case the develop install fails due to a bug.
This works great--I've used that approach for our Windows buildbot too. On Linux, since we have apt-get, it's less of an issue, even though we currently build Matplotlib from source to get the latest version. Stéfan
On 06/02/2014 06:01 PM, Michael Aye wrote:
On Monday, June 2, 2014 3:11:20 PM UTC-7, Matthew Brett wrote:
Hi,
On Monday, June 2, 2014 1:35:11 PM UTC-7, Stéfan van der Walt wrote:
Hi Matthew
On Mon, Jun 2, 2014 at 7:29 PM, Matthew Brett <matthe...@gmail.com> wrote: > I guess it would be good to build them automatically next time ... so maybe > yes to include the scripts - should be only a few lines of shell I think, > once you've got the MacPythons installed.
I admit that I am out of my depth here. Would you be willing to help us on this one? Is it, e.g., possible to add a target to setup.py so that anyone on OSX can build these?
Er - I am not sure how that could sensibly be done in the setup.py, because you'll need to follow something like the recipe on the Spinning Wheels page, installing MacPythons, pip and packages, then running ``$PYTHON setup.py bdist_wheel`` for each Python.
You're welcome to a login to one of our Berkeley machines where the basic setup is already done.
In the Travis-CI example, how does Travis get triggered? (I presume you don't commit to that repository each time you want to test MPL?)
Actually, yes, that is how I / we trigger builds at the moment, because we just got going with these megatesting things. There seem to be various options for triggering builds, including (I suppose) making fake commits, but I haven't investigated very far yet.
I just went through a quite painful learning process with all this, because the docs are quite scattered, with lots of old stuff mixed in. So I'd like to comment, in case you don't insist on the Travis build machinery to be MacPython (is that even possible, as Travis runs on Linux?), to speed things up IMMENSELY over using pip for requirements install, one can use conda for the env setup in Travis. It speeds things up, because pip doesn't store binaries and instead spends a lot of time compiling things (or so it says here: http://sburns.org/2014/03/28/faster-travis-builds.html). Find the instructions on how to setup Travis with conda in that link, I tried it and it works fine here. I did one thing different compared to that blog entry though: In the travis yaml I prefer to "script: python setup.py develop && python setup.py test", because that way I don't get a useless test run in case the develop install fails due to a bug.
Another option if you don't like making fake commits to trigger builds is to use Travis CI's "Restart Build" or the Test Hook button from the repositories Service Hooks page. Only the repositories admin can do this but it might be preferable to littering the repo with fake commits. - Jonathan Helmus
Hi, On Monday, June 2, 2014 4:01:34 PM UTC-7, Michael Aye wrote:
On Monday, June 2, 2014 3:11:20 PM UTC-7, Matthew Brett wrote:
Hi,
On Monday, June 2, 2014 1:35:11 PM UTC-7, Stéfan van der Walt wrote:
Hi Matthew
I guess it would be good to build them automatically next time ... so maybe yes to include the scripts - should be only a few lines of shell I
On Mon, Jun 2, 2014 at 7:29 PM, Matthew Brett <matthe...@gmail.com> wrote: think,
once you've got the MacPythons installed.
I admit that I am out of my depth here. Would you be willing to help us on this one? Is it, e.g., possible to add a target to setup.py so that anyone on OSX can build these?
Er - I am not sure how that could sensibly be done in the setup.py, because you'll need to follow something like the recipe on the Spinning Wheels page, installing MacPythons, pip and packages, then running ``$PYTHON setup.py bdist_wheel`` for each Python.
You're welcome to a login to one of our Berkeley machines where the basic setup is already done.
In the Travis-CI example, how does Travis get triggered? (I presume
you don't commit to that repository each time you want to test MPL?)
Actually, yes, that is how I / we trigger builds at the moment, because we just got going with these megatesting things. There seem to be various options for triggering builds, including (I suppose) making fake commits, but I haven't investigated very far yet.
I just went through a quite painful learning process with all this,
because the docs are quite scattered, with lots of old stuff mixed in. So I'd like to comment, in case you don't insist on the Travis build machinery to be MacPython (is that even possible, as Travis runs on Linux?), to speed things up IMMENSELY over using pip for requirements install
You can run travis tests on OSX virtual machines using the 'objective-c' language (that's what I'm doing). You can also use the multi-os feature in travis: http://blog.travis-ci.com/2014-05-13-multi-os-feature-available/ - but this needs the travis team to turn on that feature for you. Luckily, with wheels, installing with a requirements file just got way faster and easier. For example, here's me doing a test of the scipy stack install on first generation MacBook Air: http://nipy.bic.berkeley.edu/builders/scipy-stack-2.7.6-wheel-requires/build... Even on this very slow machine on a home network, download and installation of the scipy stack is 3.5 minutes. Cheers, Matthew
participants (4)
-
Jonathan Helmus -
Matthew Brett -
Michael Aye -
Stéfan van der Walt