simplified CI setup #163
simplified CI setup #163
Conversation
| after_success: | ||
| - Rscript -e 'install.packages("covr"); Sys.setenv(NOT_CRAN = "true"); covr::codecov("r-pkg/")' | ||
| - .ci/report_to_covr.sh |
There was a problem hiding this comment.
Maybe I'm misunderstanding the diff, but is there a reason you're only running the covr report on 6.2.4? Is it because you only want to run it once?
There was a problem hiding this comment.
Yeah it only needs to run once. So right now we're using testthat to run the tests, and then just reporting to codecov (an external service) one time to power the CI checks around coverage and the badge in the README.
Eventually I'd like to just use covr to run tests (like you and I have done in other projects) but didn't want to pull that into this PR.
.ci/install.sh
Outdated
|
|
||
| if [[ "$TASK" == "pypkg" ]]; then | ||
| PY_PACKAGE_DIR=$(pwd)/py-pkg | ||
| pip install ${R_PACKAGE_DIR} |
There was a problem hiding this comment.
This should be PY_PACKAGE_DIR
I'm not sure if it's transient. Lemme try pushing the other changes I have an we'll see. If 6.1.x and 6.2.x fail for BOTH R and Python, we'll know I got the logic for pulling and starting up ES wrong. |
.ci/seed_es_on_travis.sh
Outdated
|
|
||
| "6.2.4") | ||
| export ES_VERSION=6.2.4; | ||
| export MAPPING_FILE=${ES5_MAPPING_FILE}; |
There was a problem hiding this comment.
ah here's the error (probably?), I used the wrong mapping file.
Python tests didn't fail because there isn't a get_fields() in the Python package yet. Created #164 to document it.
Codecov Report
@@ Coverage Diff @@
## master #163 +/- ##
=========================================
- Coverage 95.14% 92.8% -2.34%
=========================================
Files 8 8
Lines 556 556
=========================================
- Hits 529 516 -13
- Misses 27 40 +13
Continue to review full report at Codecov.
|
|
Ok this is passing! @austin3dickey whenever you have time, could you take one more look? |
Looking at this project for the first time in a bit while working on #161 , I found that there are some places where can reduce duplication. In this PR, I want to propose some reorganization that will removed opportunities for error and make the exact amount of complexity (e.g. all the ES7-specific exceptions in test setup) more obvious.