@a4z You can integrate a spell checker with Nanoc if you want: https://github.com/nanoc/nanoc.ws/blob/main/lib/checks.rb
That is how I do it with nanoc.ws
apropos check, I get 2 fails positives from the ilink checker, both to the same site , one time from the blog index and 1 time from an article linking to the same page, in both case the html is
<a href="/blog/2017/10/14/Using-SQLite3-in-your-C++-application.html">Using SQLite3 in your C++ application</a>both links work... maybe the ++ in the URL?
4 replies
Hey hey @ddfreyne. Do you still consider https://groups.google.com/d/topic/nanoc/aSG7QC_oqZQ/discussion to be the ideal way to combine multiple items into one?
2 replies
can I create a ignore list for various checkers ?
I have 2 problems, one is that thi the url checker, reporting an URL that clearly works
the second is the html validator, since google wants to put me a html page into my root so it knows I am the owner of the page. And this is of course no valid html5
It would be nice if I could automate the checkers, but I need a way to exclude such false positives
I have 2 problems, one is that thi the url checker, reporting an URL that clearly works
wget https://a4z.gitlab.io/blog/2017/10/14/Using-SQLite3-in-your-C++-application.html as being wrong,the second is the html validator, since google wants to put me a html page into my root so it knows I am the owner of the page. And this is of course no valid html5
It would be nice if I could automate the checkers, but I need a way to exclude such false positives
I’ve not needed to use snapshots before and am somewhat confused about their limitations.
I’ve basically got one item that needs to write twice. I can’t regenerate the item from the beginning. E.g. https://gist.githubusercontent.com/da2x/bf51169d5dd847ebd68f324285af6016/raw/3ea397581185fe1bb28e3f0566991aeb7e0381b3/some.rb
Not quite ideal.
Hmm, hold on… it might work if you create a filter e.g.
:read_snapshot, with the two arguments being the rep and snapshot name, which then returns @item.compiled_content(rep: rep_name, snapshot: snapshot_name)
I haven’t tried it and I think it might cause issues with cyclical dependencies
btw, I am deeply impressed by the live reload injection nanoc does, is this somewhere available as a stand alone thing?
( I have some slides with ascidoctor-revealjs and a rake file that rebulids on change, now I would like to have this live reload that nanoc has to reload the presentation in the browser)
( I have some slides with ascidoctor-revealjs and a rake file that rebulids on change, now I would like to have this live reload that nanoc has to reload the presentation in the browser)
@a4z You can also install the
adsf and adsf-live gems, and then use the adsf --live-reload command to run a live-reloading web server, separately from Nanoc
Thanks! will check adsf ,
want to put this into my rake file and replace my WEBrick call so the live reload script becomes injected,
do I understand that correct, just with using
want to put this into my rake file and replace my WEBrick call so the live reload script becomes injected,
do I understand that correct, just with using
asdf-live + asdf I can remove my existing guard live reload rule ? An this becomes independent from the live reload browser extensions?
@ddfreyne GitHub doesn’t let me open a pull request at the moment. do you want this patch? da2x/nanoc@7b81707 It implements copy-on-write reflink copying on Btrfs and XFS on Linux. Like a hardlink, except the files are dereferenced if either copies are modified. Requires my Linux-only reflink.gem, which is loaded optimistically. Falls back to old behaviour if the gem isn’t installed. It’s faster on Linux and mitigates the risky business of involving a hardlink. I’m not sure if this way of loading the gem makes it slower on other OSes, though.
My upstream bandwidth is pretty sucky though (quite old domestic ADSL).
I'm starting with a question :) According to the docs, I'm under the impression that
<%=link_to_unless_current("hello", "/.*")%> should link to the page root. I get <a href="/.*">, the literal link, instead. Am on the latest nanoc release and don't have legacy identifiers enabled 🤔
paths with trailing slashes work, e.g.
<%=link_to_unless_current("hello", "/")%>, but only in the page content, not in the main menu partial (that itself invokes a ruby helper method to render the whole menu). I wonder if that's because @item is not available in the helper module to check the context?
@DivineDominion Hey there! The second argument to link_to/link_to_unless_current is going to be exactly what ends up in the href attribute.
If you want to link to a specifc item, you can use e.g. <%= link_to 'Home', @items['/home.*'] %> — i.e. the second argument is the item you want to link to
Uh oh, I think the documentation might be wrong. Checking…
Yep, documentation is wrong — the docs say
link_to_unless_current('Blog', '/blog.*') but that should be link_to_unless_current('Blog', @items['/blog.*'])
@ddfreyne I am still not certain why the
link_to_unless_current helper produces a span when I use it in e.g. overview/index.md and it's part of the content, but the same call won't work in the nav bar (default calls _navigation.erb.html which uses lib/helpers/menu.rb's render_menu). The code to link_to_unless_current is rather simple, too. My best guess would be that the partial is cached, but using puts @item.path + " -- " + @items["/overview.*"].path reveals that this line is called for each item indeed, and that the == comparison inside link_to_unless_current should be true. :/
Thanks for the hint! I think I created the problem by using
Struct.new(...) to represent menu item configurations, and have a render method inside, and call link_to_unless_current from there. @item_rep is nil there. I think this is because of the way use_helper includes the module itself (works from there), but not the module's inner Struct definitions
I know I'm not a great help to the nanoc community myself even though I use it for what now, 7, 8 years? But there aren't very many people writing about their setup
I have, though, created aliases for some hosts that were in nanoc.yaml, to create some obscurity… but it generally isn’t that useful
You could also use dotenv (https://github.com/bkeepers/dotenv) to load sensitive information