Closed hlapp closed 8 years ago
This seems to pass, so I'm going to merge it.
aaand, it doesn't work: https://circleci.com/gh/NESCent/popgenInfo/430
When I inspect the directories, all of the write permissions appear to be the same:
12:29:36~/Documents/popgenInfo (master)> ls -larth build/libs/highlight/
total 40
-rw-r--r-- 1 zhian staff 320B Mar 1 12:29 textmate.css
-rw-r--r-- 1 zhian staff 10K Mar 1 12:29 highlight.js
-rw-r--r-- 1 zhian staff 299B Mar 1 12:29 default.css
drwxr-xr-x 5 zhian staff 170B Mar 1 12:29 ..
drwxr-xr-x 5 zhian staff 170B Mar 1 12:29 .
12:29:45~/Documents/popgenInfo (master)> ls -larth use
total 136
-rw-r--r-- 1 zhian staff 3.1K Mar 1 12:29 startMicrosatellite.Rmd
-rw-r--r-- 1 zhian staff 5.6K Mar 1 12:29 StartSNP.Rmd
-rw-r--r-- 1 zhian staff 0B Mar 1 12:29 README.md
-rw-r--r-- 1 zhian staff 5.2K Mar 1 12:29 PopDiffSequenceData.Rmd
-rw-r--r-- 1 zhian staff 162B Mar 1 12:29 Makefile
-rw-r--r-- 1 zhian staff 1.3K Mar 1 12:29 LinksToResources.md
-rw-r--r-- 1 zhian staff 6.4K Mar 1 12:29 DifferentiationSNP.Rmd
-rw-r--r-- 1 zhian staff 9.2K Mar 1 12:29 2016-01-26-SNP-selection.Rmd
-rw-r--r-- 1 zhian staff 11K Mar 1 12:29 2015-12-15-microsatellite-differentiation.Rmd
-rw-r--r-- 1 zhian staff 5.9K Mar 1 12:29 2015-05-18-Dist-SNP.Rmd
drwxr-xr-x 38 zhian staff 1.3K Mar 1 12:29 ..
drwxr-xr-x 12 zhian staff 408B Mar 1 12:29 .
Apparently the bootstrap and jQuery libraries have been updated for Rmarkdown/pandoc. The following is after issuing make deploy
:
hlapp@host:~/popgenInfo$ ls -l build/libs/
total 24
drwxr-xr-x 6 hlapp adm 4096 Mar 1 19:55 bootstrap-3.3.1
drwxr-xr-x 6 root root 4096 Mar 1 20:05 bootstrap-3.3.5
drwxr-xr-x 2 hlapp adm 4096 Mar 1 19:55 highlight
drwxr-xr-x 2 hlapp adm 4096 Mar 1 19:55 jquery-1.11.0
drwxr-xr-x 2 root root 4096 Mar 1 20:05 jquery-1.11.3
drwxr-xr-x 2 root root 4096 Mar 1 20:05 navigation-1.0
hlapp@host:~/popgenInfo$
The directories owned by root were newly created from within the Docker container, so must have come from running the rendering through Rmarkdown and pandoc.
@zkamvar: Is there any reason we need to keep anything under build/lib
in the git repo if the rendering generates these anyway? If not, I can just delete the whole build/lib
directory from within the container after the rendered site archive is created.
I should of course also look into controlling the userID under which the processes run within the container. I think newer versions of Docker allow this (unlike at the time of the hackathon).
@zkamvar: Is there any reason we need to keep anything under build/lib in the git repo if the rendering generates these anyway? If not, I can just delete the whole build/lib directory from within the container after the rendered site archive is created.
No, I don't think we need to keep anything under build/lib
if Rmarkdown/pandoc is auto-generating it upon absence. I've created a branch that removes these.
More precisely, fix the deploy step timeout. The build actually succeeds, which presumably is why there wasn't any issue with the test on PR #143. The build with failed deploy step is here: https://circleci.com/gh/NESCent/popgenInfo/428
The odd thing here is that nobody changed the jquery copy, so it's not clear why this would be write-protected now if it wasn't before.
The integration test for this PR is somewhat bogus, because the failing step (deployment of rebuilt website) is not (and, until we create a development server, arguably cannot be) actually part of the tests conducted for a PR.