shajra-simspace
commented
4 years ago I'm not on the right computer to set up a PR, but I had a moment to write a version of iso_usls_update.rb that seems to fix the break. I switched from regex parsing to XML parsing, which hopefully is more robust.
#!/usr/bin/env nix-shell
#!nix-shell -i ruby
#
# Heuristic to update the ISO urls
#
require 'open-uri'
require 'json'
require 'rexml/document'
include REXML
isos = {}
xml = Document.new(open("https://nixos.org/nixos/download.html"))
xml.root.elements.each("body/*/*/*/ul/li") { |li|
hrefs = li.get_elements("a").map { |a| a.attribute("href") }
if hrefs.size == 2 and /.*minimal.*/i =~ hrefs[0].value
then
iso_url = hrefs[0].value
arch_re = /https:\/\/.+-([^-]+)-linux.iso/
arch = arch_re.match(iso_url).captures
iso_sha256 = open(hrefs[1].value).read.strip.split.first
if arch.size == 1
then
isos[arch[0]] = {
iso_url: iso_url,
iso_sha256: iso_sha256
}
end
end
}
out = JSON.pretty_generate(isos)
puts out
File.write("iso_urls.json", out)
cdituri
zimbatm
I think the web site put in a newline that breaks the assumption of the regex parse. I haven't had a moment to sort out the way I'd prefer to fix it, but I thought I'd issue the bug here now. I may put in a PR to fix it when I get the time.
In the meantime, it's easy enough to edit the generated JSON file manually.