Conversation
kokarez
force-pushed
the
fix_auto_test
branch
4 times, most recently
from
5d9cfe4 to
b53a197
Compare
| - conda update -yq conda | ||
| - conda install -y cmake -c conda-forge | ||
| - conda install -y xtl==0.6.7 nlohmann_json -c conda-forge | ||
| - conda install -y xtensor=0.20.10 -c conda-forge |
There was a problem hiding this comment.
Any reason why you use three separate commands here? I'd prefer just one command.
There was a problem hiding this comment.
Really sorry about this long delay - I did not receive any notif :( and sadly I got distract but Im back to finish this push!
Done
| cmake . -DCMAKE_INSTALL_PREFIX=$HOME/miniconda | ||
| make install | ||
| cd $TRAVIS_BUILD_DIR | ||
| fi |
There was a problem hiding this comment.
I forgot why I had this failsafe thing anyway, so good riddance, thanks.
kokarez
force-pushed
the
fix_auto_test
branch
3 times, most recently
from
77ecef6 to
3e8b6ff
Compare
|
Interesting enough the -march=native was creating incompatible assembly code on the test system, Im not sure why this was added in the first place. I digged around, could not find why, also the xtensor is adding only for c++ (from what I saw) and add_compile_options will add for all. Here for more context: https://stackoverflow.com/questions/39501481/difference-between-add-compile-options-and-setcmake-cxx-flags |
|
@kokarez thanks for all the work! One small request, if it's not too much trouble for you: I would prefer if you did regular pushes instead of force pushes so I can keep track of what happened between commits instead of having to compare the whole thing after each commit. |
|
I do force to keep my list of commit clean an meaning full - I don't really like to have commit to fix comment that end up in master. But yea I also understand that diffing is annoying. Will do simple push starting from now |
|
Thanks, much appreciated! |
3 participants
Here what was wrong but here is the main summary:
So overall this is cleaning/simplification of the test conf.