Closed jbrown-xentity closed 2 years ago
Thanks to @mogul's help yesterday untangling the rebase issue, the builds relatively correctly but with many conflicts.
Here is one, for example:
Conflict: The following destination is shared by multiple files.[0m
2022-03-16 18:04:32 INFO [build-jekyll] [33m The written file may end up with unexpected contents.[0m
2022-03-16 18:04:32 INFO [build-jekyll] [33m [36m/tmp/work/site_repo/_site/food/dr-woteki-on-scientific-integrity/index.html[0m[0m
2022-03-16 18:04:32 INFO [build-jekyll] [33m - /tmp/work/site_repo/www.data.gov/food/dr-woteki-on-scientific-integrity/index.html[0m
2022-03-16 18:04:32 INFO [build-jekyll] [33m - /tmp/work/site_repo/www.data.gov/_posts/2013-07-25-dr-woteki-on-scientific-integrity.md[0m
2022-03-16 18:04:32 INFO [build-jekyll] [33m [0m
See below for the full conflict error message:
And a cleaned up list of those html files (grabbed with cat conflicts | rg -o --trim '^.*?(/Users/rob/work/repos/datagov-website/www.data.gov/.*\\.html).*?$' | cut -c 3-
(I put the list above in a file called conflicts)):
This removes the conflict warning in build but unearths a larger problem. Those destination urls are using the html version and not the md version. When the html is removed, the md is used.
The md versions are unfortunately not at all ready to be used:
vs
Given that, I think we need to go with the html files for now and remove the md. I understand that this somewhat negates our hopes that these files would serve as a springboard into using layouts for the rest of the site, and maybe they will but in the short term it will take a lot of time and effort to make sure these md files are all complete with their intended layouts.
cat conflicts | rg -o --trim '^.*?(\.md).*?$' | cut -c 3-
for list of just the md files described above:
With those files removed there are no conflict warnings and here is the Federalist build and still the same PR
Down to
out of the original
The larger groups of issues are done, no on to the slow individual ones..
At long last:
One very large caveat with this: html proofer ignores files in the .htmlproofer-ignore file. I'm not sure how this file was populated, but currently lists some 900+ files. Some/all of these files may contain issues. I'm not sure how far down this rabbit hole we want to go, but at a glance there are potentially important html files not being checked like /index.html
, /ocean.html
, and other top-level files.
From my understanding of the static site discussion yesterday, this is ready for review with approval of PR. See Federalist build.
User Story
In order to manage datagov-website easily, data.gov admins want all raw HTML files replaced with template versions and Jekyll layouts.
Acceptance Criteria
[ACs should be clearly demoable/verifiable whenever possible. Try specifying them using BDD.]
Background
This is coming from https://github.com/GSA/datagov-deploy/issues/2873. While the AC is complete, and some files were transitioned to templates here.
Styling the jekyll site will be the bulk of the work; this will be a manual process. The whole site will share some common header/footer element. Other than root level pages, all contents can be grouped into different categories and each category will have some different UI elements that need custom layout. There are 20-30 categories, depending on what tool is used to do the counting, so there are 20-30 layouts needed to be coded in Jekyll. For reference, here are the templates that existed in the Wordpress era.
Security Considerations (required)
[Any security concerns that might be implicated in the change. "None" is OK, just be explicit here!]
Sketch
Next steps from basic build: