Case sensitivity causes confusing archive generation

[gboone]

The problem: You have both tags "USA" and "usa" and the following jekyll-archives configuration:

jekyll-archives:
  enabled:
    - 'tags'
  layout: tag_index
  permalinks:
    tag: '/tags/:name/'

USA has 4 posts
usa has 1 post

The result:
On the page generated at /tags/usa will include only the 4 posts tagged in caps.

Interim solution:
Users should build a habit of consistent capitalization when tagging.

[parkr]

@alfredxing, have a minute?

[alfredxing]

I've been super busy this term so I've barely been able to keep up!
Just added this to my to-do list so I remember.

Reminds me of a previous discussion about whether tags should be case-sensitive or not. What are your thoughts on that?

We can either make tags case-insensitive, or make the generated URLs case-sensitive...

[gboone]

I'd vote making tags case-insensitive so, in the example above, USA and usa are treated as the same tag with the same url tags/usa/.

[hazzik]

The real problem, that one of the tag pages will be missed.

[hazzik]

Same as #24

[ghost]

A fix for this is here: fc8af85e/lib/jekyll-archives.rb and fc8af85e/lib/jekyll-archives/archive.rb and an alternate fix is in #82.

Tags and category hashes are sorted by default as a, A so uppercased tags/cats have priority over lowercase. (So they end up overwriting the lowercase versions) Suppose we have 4 posts with 4 different versions of a tag:

- The Tag
- The tag
- the Tag
- the tag

Here's the default output of @site.post_attr_hash('tags')

{"The tag"=>[#<Jekyll::Document _posts/2016-05-09-The-tag.md collection=posts>], "The Tag"=>[#<Jekyll::Document _posts/2016-05-12-the-tag.md collection=posts>, #<Jekyll::Document _posts/2016-05-10-The-Tag.md collection=posts>], "the Tag"=>[#<Jekyll::Document _posts/2016-05-11-the-Tag.md collection=posts>]}

As you can see Jekyll sorts The Tag and the tag into a single array, but puts The tag and the Tag into separate arrays, for a total of three arrays.

However, jekyll-archive doesn't take the sort into account when it uses this hash to build archive pages. Only a single array will be used for the archive page, and the other arrays will be completely dropped. This creates missing posts on archive pages.

So I'm using a method for case-insensitive tag and categories. I borrowed it from the date def block. This code yields the following hash for the previous 4 versions of the tag:

{"The Tag"=>[#<Jekyll::Document _posts/2016-05-12-the-tag.md collection=posts>, #<Jekyll::Document _posts/2016-05-11-the-Tag.md collection=posts>, #<Jekyll::Document _posts/2016-05-10-The-Tag.md collection=posts>, #<Jekyll::Document _posts/2016-05-09-The-tag.md collection=posts>]}

So it works nicely to create case-insensitive tag & category archive page generation.

The generated page titles default to downcased, but you can alter that to uppercase, or capitalized using some lines in _config.yml.

[wsdjeg]

@alfredxing no please do not make tag case insensitive, it is same as url, by default USA and usa is not a same tag. sometimes we usually use uppercase for long tag abbreviations。