-
-
Notifications
You must be signed in to change notification settings - Fork 15
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
Windows attempt #34
base: main
Are you sure you want to change the base?
Windows attempt #34
Conversation
Hi! This is the friendly automated conda-forge-linting service. I just wanted to let you know that I linted all conda-recipes in your PR ( I do have some suggestions for making it better though... For recipe:
|
@conda-forge-admin, please rerender |
…da-forge-pinning 2020.05.14.17.14.20
@conda-forge-admin, please rerender |
…da-forge-pinning 2020.05.14.17.14.20
Hi! This is the friendly automated conda-forge-linting service. I wanted to let you know that I linted all conda-recipes in your PR ( Here's what I've got... For recipe:
For recipe:
|
Hi! This is the friendly automated conda-forge-linting service. I just wanted to let you know that I linted all conda-recipes in your PR ( I do have some suggestions for making it better though... For recipe:
|
This last commit works (builds) but the tests do not. |
…nda-forge-pinning 2020.05.19.13.30.35
Managed to run the test suite locally, more or less! Some tests hang so you have to These are the results:
Full logs and diffs: |
Yay, that made the test suite run! Now we can actually test on the CI, awesome! And on Azure there's not
|
…nda-forge-pinning 2020.06.07.16.42.19
Hi! This is the friendly automated conda-forge-linting service. I was trying to look for recipes to lint for you, but it appears we have a merge conflict. Please ping the 'conda-forge/core' team (using the @ notation in a comment) if you believe this is a bug. |
Hi! This is the friendly automated conda-forge-linting service. I just wanted to let you know that I linted all conda-recipes in your PR ( I do have some suggestions for making it better though... For recipe:
|
…nda-forge-pinning 2020.06.07.16.42.19 (windows only again)
Tests seem to get stuck at different parts:
Running locally to investigate why. |
Ok, that "works" so at least we can run the full suite reliably now. Summary for now:
Detailed analysis coming soon. |
Checklist
0
(if the version changed)conda-smithy
(Use the phrase@conda-forge-admin, please rerender
in a comment in this PR for automated rerendering)Strategies
Building on Windows will take a bit of effort... Amber team recommends the MinGW compilers, but some dependencies are built with other suites, which is specially complicated in the case of Fortran.
Option A:
mys2
This is the current status for our first line of dependencies:
gfortran
enabled builds are also available.As you can see, there's three main problems:
netcdf-fortran
: I submitted a PR to solve this, but we might need to wait for a new version upstream.fftw
is on Windows but Fortran support is not enabled. I have raised an issue here.Current status
Suboption A
Let CMake choose the bundled dependencies and build the troublesome deps. This includes
boost
,netcdf
,netcdf-fortran
,readline
,fftw
, andarpack
.xblas
static lib is being vendored. We might be able to distribute it like this if everything is made static.Build
FEW
is not available because of some missing Perl libraries.Test
pytraj
cannot be built on Windows+MinGW. Requires MSVC and maybe a separate packaging.resp
won't startrun_test.bat
m2-tcsh
(done, recipe here).bat
or.exe
. We need to provide temporary forwarders so the tests reach the executables.parmed
's bundledsimtk.unit
has a numpy C issue (missing DLL importingnumpy.core._multiarray_umath
).Suboption B
In this case we use
conda-forge
libraries. Blocking:netcdf-fortran
is not available (yet) and compilation fails because of the same issues as in the PR. Once [WIP] Attempt windows netcdf-fortran-feedstock#52 is merged, we might get further. Interesting how this doesn't happen ifnetcdf
is compiled at the same time.arpack
(lacksarsecond
?)fftw
(no Fortran bindings)boost
(no compression support)readline
(not available).Option B: MSVC + flang
With this approach, we would use MSVC compilers for C/C++ and
flang
for Fortran. Flang packages do not always work (scipy example), so it might be tricky.Note: This might change in the future because (A) apparently
ifort
support is coming to CF, and (B) LLVM will provide fullflang
support.Large parts of Amber only expect a MinGW-like compiler, so they won't be able to use another compilers unless the name mangling is done in a MinGW-compatible way.
Given the small margin for success here, I won't even try for now :)