thanks for your help! @ddfreyne
If you don't mind me asking, how does game.title and game[:title] differ? is it something to do with hash tables? (Still new to ruby)
game is not a hashtable (or a hash as Ruby calls it, or a dictionary/map). It’s its own type that is defined by Nanoc
The
game object implements the [] method
game[:title] is sort of a shorthand for calling the [] method with the :title argument on game… so, kind of like game.[](:title) in a way. (Weird to think of it that way)
class Moo
def [](key)
puts "you requested #{key}"
123
end
def []=(key, value)
puts "setting #{key} to #{value}"
end
end
moo = Moo.new
moo['bloop']
# outputs: you requested bloop
# returns: 123
moo['beep'] = 'bap'
# outputs: setting beep to bap
You can build your own array-like or hashmap-like classes like this :arrow_up:
I created a bug ticket btw: nanoc/nanoc#1446
Will fix it soon
Will fix it soon
yay :D if you need access to the whole website to help you in diagnosing i can (finally) upload it to github and give you access (as its definately not ready to see the world yet)
@ddfreyne do you know how I can make everything stop while I inspect things? right now it just messes the terminal :(
Does anyone have a good working solution for finding related items? The approach I’ve seen used looks up items using the same tags. This approach didn’t really produce good results for me. I was hoping to find something using linking relations, or maybe LSI instead? I thought about porting https://github.com/alfanick/jekyll-related-posts/blob/55768e828c747de31ef77e88475060d1878f8e37/lib/jekyll-related-posts.rb to Nanoc but initial testing suggests it will be reaaally slow.
@ddfreyne hello I have a question for you: I'm still toying with source maps. Let's say
sass filter. Imagine my render() function always compute both the output .css file and the output .map file. Yet for "API" purposes and to fit the concept of reps I have 2 filters: sass and sass_sourcemap. Now I want to cache and make either of the two filters take back what's cached by the other one. What should the key be? pair of item's identifier and checksum? If not, what else?
There are some LSI tools out there but I have not used the Ruby ones
@gpakosz Hmm, I assume you mean caching in memory, i.e. inside the same nanoc process, yeah? You could cache by item identifier (checksum is not needed)
But you’d still have to call
item.compiled_content, as mention before, to make sure the dependnecies are working as intended
@giodestone The bug you discovered (nanoc/nanoc#1446) is fixed in
master, but there’ll be a proper Nanoc release with the fix soon!
Nanoc 4.11.8 is out!
@da2x it’s either-or at the moment. Do you have a use case for having attributes in two locations?
I started working from the Jekyll related post plugin, but it was incredibly slow going. It uses the ClassifierReborn gem and that thing is reeeaaaly slow. Made a small thing in R with the LSA and LSAfun libraries and it processes 500 ~1500-word documents in seconds rather than six hours. Outputs a list of related items to a CSV file, and I then process that file in the Nanoc pre-processor and set them attributes for items (e.g.
related_items: [a.md, b.md]) and pull them in my layout like so: @items.find_all(@item[:related_items]).
@ddfreyne so what I did before was the above but output
related_items: [a.md, b.md] in c.yaml instead of doing it in the pre-processor. It’s a corner case for sure, but I also expected it to just work. The .yaml file took precedence over the in-file meta data which was just silently dropped.
@da2x You could try to use https://github.com/bobthecow/forematter — it’s not super-maintained, but I think it might fit your use case
@ddfreyne I wonder if I'm the one who introduced #1451. Or at least my refactor revealed it. It's a bit hard to read the history since files moved around in the repo though
@ddfreyne In any case I'm sorry!
Let me know if that works