rbuchberger
commented
5 years ago PictureTag.page pulls its data from the context object jekyll passes it (context.registers[:page]), and the markdown check just looks at the filename extension of the page name. It sounds like the tag is getting called in a way where that doesn't work, but I'm having a hard time figuring out how.
Can you narrow down where it happens? Includes were my first thought, but they seem to work on my end.
Here's some troubleshooting code:
In lib/jekyll-picture-tag.rb, add the following to the very beginning of the render function:
puts context.registers[:page].reject { |k, _v| k == 'content' }
This will spit out all the pages where it's called from, stripping out the content so it's actually readable. The very last one it spits out before crashing will tell us something. Can you tell me what it says?
If it's helpful, I can make a branch to do it for you, so all you have to change is the gemfile. Thanks!
user-none
In utils.rb a function
markdown_pagewas added to check if a page is a markdown file. This doesn't seem to be working becausepage['name']isnil. I added a for loop to print all of the kvs for a page when the function is called.I get two different types of output. A "post" and a "fragment" I'll call them.
Page
Fragment
I added a
nilcheck forpageandpage['name']to skip this and aputsto printpage['name']when it exists. Doing this I eventually getindex.htmlas the only output. I'm not sure what's going on exactly but it looks like it's running over the generated html as well as the markdown source files. But the markdown files don't have a name attribute.