That reminds me that I want to open up the API for the colorize_syntax helper and support adding custom colorizers… it’s pretty straightforward but not officially supported:
class DenisColorizer < Nanoc::Filters::ColorizeSyntax::Colorizers::Abstract
identifier :denis
def process(code, language, params = {})
# … return HTML containing highlighted `code` …
end
end filter :colorize_syntax, default_colorizer: :denis
I'm trying out removing the colorize_syntax filter
and then extend the GFM Input to add wiki link finding
@ddfreyne Maybe that's relevant for nanoc upstream, too, since I'm using a custom strip_html replacement to avoid truncating in the middle of a word:
# Custom variant of the built-in
# excerptize(strip_html(@item.compiled_content(snapshot: :pre)), length: threshold)
# because the default also truncated in the middle of a paragraph.
def plain_summary(item, threshold = 300)
sanitizer = Rails::Html::FullSanitizer.new
markdown = Redcarpet::Markdown.new(Redcarpet::Render::StripDown, :space_after_headers => true)
content = sanitizer.sanitize(item.raw_content)
content = markdown.render(content)
content = content.gsub(/\((https?:\/\/|\/).*?\)/, '').gsub(/ ([ \.,;])/, '\1')
"".tap do |summary|
content[0..(threshold+50)].split.each do |word|
summary << word + " "
if summary.length > threshold
summary << "..."
break
end
end
end
end
Could maybe work with https://kramdown.gettalong.org/converter/remove_html_tags.html on the compiled content instead and then drop the Rails HTML helper
So this sorta works. I had to patch the table pipe recognition to ignore wiki links, and this is pretty slow this way
I'm fine with that proviso. The tradeoff is that testing code that moves around the content tree is tricky without some help. I have a couple of sites where I have some degree of interchangeability between
/a.md and /a/index.md and there are a surprising number of edge cases for a concept like "parent" in that situation.
So
nanoc-spec helped a lot.
What brought this to my attention is the capability to display author avatars, which is cool for publications with multiple authors. The overall feed structure mimicks the feature of a Twitter timeline
i'm migrating an old website built on an older version of nanoc, using content_for method... which doesn't seem to exist anymore. can maybe anyone point me out what's the current preferred way ?
@DivineDominion I’ve looked at JSON Feed, but haven’t implemented anything for it… could be good to add
#json_feed to the Blogging helper I suppose!
Not quite Nanoc-related, but I’m rather happy with the horizontal multi-column layout for article I’ve been experimenting with: https://twitter.com/ddfreyne/status/1396578696962744329
@DivineDominion I think it works especially well for the larger Lenses article (https://experiments.denisdefreyne.com/horizontal-layout/index-lenses.html). Still, it required more custom CSS (and even a bit of JavaScript) to get it to work super well.
I’m looking for new career opportunities — anyone here have some good leads? :)
Also see --> https://twitter.com/ddfreyne/status/1399404990783950849
Also see --> https://twitter.com/ddfreyne/status/1399404990783950849
Interesting! Though this is off-putting:
Ruby and JavaScript experience are a must; all Discourse applicants have a 40-80 hour paid trial project (over 2-10 weeks) prior to joining the team. The trial project usually involves adding a feature to Discourse or a Discourse plugin.
:grimacing:
if you have a 50h/wk job already, well that's tough
otherwise it'd boil down to 1 extra work day for 10 weeks at kost, e.g. Saturday. I dunno what more you could offer, excaept more time than 10 weeks, and I bet if you're a single father with a bunch of kids homeschooling, these nice folks might be trying to help
In my book this is as good as it could ever get -- if you want to get to know applicants at all, that is. Other companies might make do without onboarding like that, and it's fine to pick these instead. I don't see how Discourse's approach is problematic by design, though
It’s possibly tricky from a legal/finance point of view in Germany as well, as you’d have to sign up to be a freelancer first (I know I had to)
In good news though, looks like I’ll have a new job soon — if all goes well!
In other news, I looked through the Dart Sass API and found this.
This is good news, because it means that it should now be pretty simple to build a
:dart_sass filter that does proper dependency tracking!
Awesome, was just stepping by to ask what the current state of the sass filter is
I am going now full pro in my web development and use that :-)
(just kidding, still have no clue what I do but I continue to use nanoc, and now I even build a new site)
but I guess, fastest would be to quickly add a own filter that assumes some dart/sass is in path and calls that ?
I am going now full pro in my web development and use that :-)
(just kidding, still have no clue what I do but I continue to use nanoc, and now I even build a new site)
but I guess, fastest would be to quickly add a own filter that assumes some dart/sass is in path and calls that ?