-
Notifications
You must be signed in to change notification settings - Fork 262
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add AppVeyor testing for Windows #347
Conversation
It seems that the https://ci.appveyor.com/project/ocefpaf/netcdf-c/build/1.0.13 This can be improved to run some tests too, right now it only builds the cc @dopplershift who may be interested in this PR. |
I added netcdf-c to Unidata's AppVeyor, and have given Unidata's netCDF team what should be proper permissions. Power-cycling the PR to hopefully get AppVeyor to run. |
Thanks @dopplershift for jumping in on this! |
Well, the infrastructure works--but the build failed. |
I'll dig into it but at a glance it looks like it is using visual studio 9, which is not a version we have tested against. I know that a lot of effort goes into supporting versions 11 and 12, as there are differences from 14 (the latest, I think). I'll look and figure out the specifics. |
What's strange it that it worked for @ocefpaf's. |
It may be environmental; I see the following:
This suggests a header isn't being seen for some reason. It could be environmental, or maybe something in cmake? I'm not certain, yet, I'll have to look. It could also be due to missing tools, potentially. I'm not sure what this config |
Remember that ncgen.l and ncgen.y should not be scanned because |
The test passed up until ocefpaf@fd2911a See https://ci.appveyor.com/project/ocefpaf/netcdf-c/history PS: Note that this PR implementes only a "can we build this" kind of test. Additional real tests should be added later. |
conda.recipe/meta.yaml
Outdated
test: | ||
commands: | ||
- ncdump -h http://geoport-dev.whoi.edu/thredds/dodsC/estofs/atlantic | ||
# FIXME: the next two are failing at the moment. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thes URLS are either not reliable (geoport-dev.whoi.edu) or cycle off (the GFS URL with a specific date).
This could be a good one as:
- It's at Unidata
- It doesn't reference a model resolution that may disappear someday (e.g. 0.5=>0.25) and it doesn't specify a particular date
ncdump -h http://thredds.ucar.edu/thredds/dodsC/grib/NCEP/WW3/Global/Best
IMHO, it's not useful to do more than one of these.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Indeed we need reliable endpoints for a robust test. Here I only gather the URLs I knew were failing. Before we merge this we can try to find, or create, some stable ones.
conda.recipe/meta.yaml
Outdated
test: | ||
commands: | ||
- ncdump -h http://geoport-dev.whoi.edu/thredds/dodsC/estofs/atlantic | ||
# FIXME: the next two are failing at the moment. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thes URLS are either not reliable (geoport-dev.whoi.edu) or cycle off (the GFS URL with a specific date).
This could be a good one as:
- It's at Unidata
- It doesn't reference a model resolution that may disappear someday (e.g. 0.5=>0.25) and it doesn't specify a particular date
ncdump -h http://thredds.ucar.edu/thredds/dodsC/grib/NCEP/WW3/Global/Best
IMHO, it's not useful to do more than one of these.
@rsignell-usgs But is it currently broken on windows with 4.4.1.1? |
Okay, actually, the latest
Here's my config:
|
I don't think you finished that thought... |
Where is that geoport reference coming from? The only occurrence |
@DennisHeimbigner that was just a test for the conda recipe. |
Latest source is no longer build on Windows for Python 2.7 (vc9) and Python 3.4 (vc10). This is currently a blocker to build See https://ci.appveyor.com/project/ocefpaf/netcdf-c/build/1.0.23 |
Latest source broke all https://ci.appveyor.com/project/ocefpaf/netcdf-c/build/1.0.25 |
This PR is stale and I do not have the bandwidth to keep updating it. I guess that Windows will always be a second class citizen in most projects :-( |
To help testing on Windows I put together this
appveyor.yml
file that buildsnetcfd-c
withconda
andcmake
. The reason I choseconda
is to easily get all the dependencies (curl
,hdf4
,hdf5
, andzlib
).However, in order to get this PR working, we need to figure out a way to use AppVeyor'scygwin
macro language (m4
) to provide the auto-generated source files. (Or those files could be committed into the repo?)PS: The repo adims must activate AppVeyor. Here is the first run on my account: https://ci.appveyor.com/project/ocefpaf/netcdf-c/build/1.0.8