A C extension module that compiles fine on Travis-CI without anaconda fails when installed with anaconda. It appears to install just fine, but when I try to import it, I get the following error:
ImportError: /home/travis/anaconda/lib/python2.7/site-packages/quaternion/numpy_quaternion.so: undefined symbol: __log_finite
The full error can be seen here. Obviously, this looks like a linker error, where it can't find glibc (which I believe is where __log_finite is found). But why should it fail to find glibc?
When I run nm on that .so file (through Travis), it shows that __log_finite is indeed undefined, but shouldn't it find it through the usual process?
I've tried installing quaternion through pip and I've tried installing it by directly downloading it and running python setup.py install. Both seem to work, in the sense that it looks like all the files are where they should be. But both fail on import because they can't find that symbol.
I've even tried installing the full version of anaconda (rather than just miniconda, which is recommended). Nothing seems to work. How can I make Travis find that symbol, and is this something I'll have to worry about ordinarily with my distribution?
It appears to be a problem with a -ffast-math flag in my quaternion package. One thing that flag does is make the code assume that the numbers are finite, so that instead of using the log function, it uses some log_finite function, which for some reason Travis doesn't have --- or something. Anyway, I have my numba package set an environment variable in Travis builds, which the quaternion package then looks for on installation, and turns off fast-math. This is unfortunate, because it means I'm not actually testing the code as it's actually used. But it means my code builds and tests pass.
There seems to be about one mention of this on the internet. Or not; I can't tell.
Related
I am trying to run the AdaIN-style model in Google Colab. In order to do this, I need to install Lua. I saw this question and tried the solution given (reproduced below for convenience), both with the given distro and the official torch distro.
%cd /content/
!git clone https://github.com/nagadomi/distro.git torch --recursive
import os
os.chdir('./torch/')
!bash install-deps
!./install.sh
!. ./install/bin/torch-activate
Now it should work using absolute path to th:
!/content/torch/install/bin/th neural_style.lua -style_image examples/inputs/picasso_selfport1907.jpg -content_image examples/inputs/brad_pitt.jpg
However, the last line fails, giving me the below error
/bin/bash: . ./install/bin/torch-activate: No such file or directory
However, the binary for th seems to exist at the given filepath. Thus, I try to run the given command
th test.lua -content input/content/cornell.jpg -style input/style/woman_with_hat_matisse.jpg
but I get the following error
/content/torch/install/bin/luajit: /content/torch/install/share/lua/5.1/trepl/init.lua:389: module 'cudnn' not found:No LuaRocks module found for cudnn
I tried looking this up further, but wasn't able to make much progress.
I also tried this method, which seems similar to the above, but I get the same errors.
Any pointers would be appreciated.
Problem summary
When running mypy on my code, I keep getting many Library stubs not installed errors.
Few examples below:
/opt/conda/envs/my_ci/lib/python3.9/site-packages/ray/core/generated/agent_manager_pb2.py:5: error: Library stubs not installed for "google.protobuf.internal.enum_type_wrapper" (or incompatible with Python 3.9)
/opt/conda/envs/my_ci/lib/python3.9/site-packages/torch/utils/tensorboard/summary.py:8: error: Library stubs not installed for "six.moves" (or incompatible with Python 3.9)
/opt/conda/envs/my_ci/lib/python3.9/site-packages/ray/rllib/algorithms/algorithm.py:28: error: Library stubs not installed for "pkg_resources" (or incompatible with Python 3.9)
Currently I have to use the command
mypy --install-types --non-interactive my_folder --config-file=mypy.ini.
And although it solves the issue, the problem now is that it takes at least 2 min for installing missing types. This is very long for our CI/CD pipeline.
Question
What are alternative ways of addressing missing library stubs? E.g., such that I could maybe 'split' mypy install types (or other solution that is more time-consuming and potentially can be put as part of docker image), from pure run mypy command (that takes less time and runs as part of gitlab ci/cd pipeline).
I tried running mypy --install-types command first, and then run mypy without success. Could it be that I am doing something wrong?
I will appreciate any help and ideas!
I have been using a google colab template for iterative LQR that uses the Pydrake, however, it seems like the code repository is removed and I can't reinstall it on google Colab:
try:
import pydrake
import underactuated
except ImportError:
!curl -s https://raw.githubusercontent.com/RussTedrake/underactuated/master/scripts/setup/jupyter_setup.py > jupyter_setup.py
from jupyter_setup import setup_underactuated
setup_underactuated()
# Setup matplotlib backend (to notebook, if possible, or inline).
from underactuated.jupyter import setup_matplotlib_backend
plt_is_interactive = setup_matplotlib_backend()
File "/content/jupyter_setup.py", line 1
404: Not Found
^
SyntaxError: invalid syntax
I tried clicking this link https://raw.githubusercontent.com/RussTedrake/underactuated/master/scripts/setup/jupyter_setup.py, and the page is not found... everything was working fine yesterday
Sorry. You're correct... I updated it this morning, and don't have a good deprecation policy in place on that repo, and this setup script is two versions ago. The path you want is https://raw.githubusercontent.com/RussTedrake/underactuated/master/setup/jupyter_setup.py
(remove the script from the directory). But if you look at that file, you'll see that even that is pointing to an updated setup script which you might want to point to.
This is actually all good news... we are on the path to a much better solution. You can now just pip install drake on colab (see the drake installation guide). Once I land the pip install underactuated (probably in time for my Spring offering of the class), then all of that nasty setup will be gone.
I have looked at many other posts related to this issue and have tried each solution. None have worked in my case, including copying over the makevars from Rcpp. Anyhow, when building on Travis I get the following error
undefined symbol: dpotrf_’
The interesting note is that the package installs fine on windows, macOS, and linux.
here is my repo R package
I can reproduce the failure on a very standard Debian testing system (which I use for the extensive reverse dependency checks on Rcpp and RcppArmadillo).
After installing packages bain and BFpack (I had the rest) I attempted to build the tar.gz from your pristine just-checked-out sources. And I get:
*** installing help indices
*** copying figures
** building package indices
** testing if installed package can be loaded from temporary location
Error: package or namespace load failed for ‘BGGM’ in dyn.load(file, DLLpath = DLLpath, ...):
unable to load shared object '/tmp/Rinst106c6ed5251a/00LOCK-BGGM/00new/BGGM/libs/BGGM.so':
/tmp/Rinst106c6ed5251a/00LOCK-BGGM/00new/BGGM/libs/BGGM.so: undefined symbol: dpotrf_
Error: loading failed
Execution halted
ERROR: loading failed
* removing ‘/tmp/Rinst106c6ed5251a/BGGM’
-----------------------------------
ERROR: package installation failed
This appears to be a moderately complex and large enough package so please pardon me for not diving in and debugging. I would suggest you simplify with smaller mock packages to see what may be wrong. (dpotrf is a fairly standard LAPACK routine so something somewhere calls it. Maybe you call it explicitly. Maybe you did a Fortran-to-C mapping wrong. Maybe you have something wrong in how you interface with RcppArmadillo. Hard to tell...)
Edit: You committed compiled code and a Windows library. "Don't do that." When Travis builds it also starts from a git checkout as I did. That may be the difference.
Edit 2: It wasn't, but your R code mixes .Call() with generated entry points (ie via RcppExports.cpp and RcppExports.R). I have seen that blow up for other people. That may be something to look into.
Disclaimer: I work with D_Williams, but I figured out the problem, and others may find it useful.
A functioning configure.ac was present, and a Makevars.in is present.
The problem is that the configure file was not yet generated. This is an autotools/autoconf setup. To resolve it, I ran autoconf in the package directory, which generated the configure file. That configure file is then executed when R builds the package. The configure file modifies the Makevars.in and creates Makevars. That Makevars file ultimately defines where to find libraries, includes, compilers, compiler options, etc.
If you do not generate the configure file from configure.ac using autoconf, then there is no configure file to be executed, and no Makevars to define the needed options at compile time. Therefore, the compiler is not fully configured, and it will fail.
TLDR: If you have an configure.ac, you must run autoconf on it, and commit that configure file to your repo. R needs to execute it to have a functioning Makevars.
I'm having trouble running Agda on my windows 7 64-bit pc. I tried running the following commands:
cabal install agda
and
cabal install agda-executable
which both work, but I still can't seem to get it going with emacs, can someone help me? I've tried the one click installer from here but it doesn't seem to work, I run into this complaint:
C:\agda2\bin\agda2-install.cmd failed.
Code = 1
Incorrect Function
Setup was not completed.
Please correct the problem and run setup again
Install Agda (cabal install agda agda-executable), make sure agda-mode.exe is on search path. This required because installer will put something like (shell-command-to-string "agda-mode.exe locate") to .emacs file.
run "agda-mode setup".
My Emacs bin directory is also on the search path but I'm not sure if it required or not. Anyway, you will get error messages if you'll do something wrong.
You will also need Haskell mode (can be obtained from here https://github.com/haskell/haskell-mode) and you will have to add (load "c:\\haskell\\dev\\haskell-mode\\haskell-site-file.el") to your .emacs file. I load it prior to Agda stuff, I do not really know does load order matter or not.
I have GHC 7.0.4 (with Haskell Platform 2011.4.0.0) and I could install Agda 2.3.0 installer in Win7 64b without problems.
If you have GHC 7.4, you should try Agda 2.3.0.1