sanitizer.rb

moduleSanitizer
extendself

# Sanitizing of Markdown converted from RDoc of ruby repo
deffrom_repo(path,source)
source
.sub(/\A\#\s*-\*-[^\n]+\n/,'')# First line with settings to editor
.strip
.gsub("ruby-talk:69518\n: ","[ruby-talk:69518] ")# Just broken by rdoc2markdown
.gsub(/\n(.+)\n: {3}(?=\S)/,"\n* \\1: ")# RDoc converts dd/dt this way, no Markdown corresponding syntax...
.gsub(/`(\w[^\n`']*?)'/,'`\1`')# Code pieces wrapped in `foo'
.gsub("`(?``*name*`')`","`(?'`*name*`')`")# Manually fix one case broken by previous line
.gsub("`enor'","'enor'")# And another one...
.gsub(/(\n[^\*\n][^\n]+)\n(\* )/,"\\1\n\n\\2")# List start without empty space after previous paragraph
.gsub(/(\[RFC\d+\])\s+(?=\(http)/,'\1')# Remove wrongful spaces in [RFCxxxx] (http://) for Net:: libs
.split("\n").map(&method(:line))
.join("\n")
.then{per_repo_file(path,_1)}
end

# Sanitizing of Markdown from ruby-lang.org repository
deffrom_site(path,source)
source
.sub(/\A---\n.+?\n---\n/m,'')# YAML frontmatter
.gsub(/\{:.+?\}/,'')# {: .foo} tags
.gsub(/\{% highlight (sh|ruby|c) %\}\n.+?\n\{% endhighlight %\}/m){ |str|
# Shell commands, Ruby code of "old" Jekyll syntax → to just space-formatted
str.gsub(/\{%.+?%\}/,'').gsub(/^/,' ')
}
.gsub(/\{% highlight irb %\}\n.+?\n\{% endhighlight %\}/m){ |str|
# In one place IRB code is too close to the list above, and indent-formatted
# code gets broken.
#
# "^"" is Kramdown's end-of-block marker (to say "the list is ended, this indent
# is NOT continuation of it")
# https://kramdown.gettalong.org/syntax.html#eob-marker
str.gsub(/\{%.+?%\}/,'').gsub(/^/,' ').prepend("\n^\n")
}
.then{per_site_file(path,_1)}
end

private

defline(ln)
returnlnifln.start_with?(' ')# It's code, don't touch it!

ln
.gsub(/(?<=[^`]|^)([A-Z][a-zA-Z:]+[#][a-z_?!\[\]=]+)(?=[^`a-z_?!=\[\]]|$)/,'`\1`')# FooBar#foobar, always method reference
.gsub(/(?<= |^)([#][a-z_?!\[\]=]+)(?=[^`a-z_?!=\[\]]|$)/,'`\1`')# #foobar, always method reference
.gsub(/(?<=[^`])(__[A-Z_]+__)(?=[^`])/,'`\1`')
.gsub(/(?<=[a-z])`(?=[a-z])/,"'")# Fancy apostrophe for typesetting
.gsub('---','—')# Just dash. Probably should be done smarter
end

defper_repo_file(path,content)
casepath
when'doc/keywords.md'
# TODO: Or maybe fancy table?
content
.gsub(/\n\* (\S+):/,"\n* `\\1`:")
when'doc/globals.md'
# TODO: Or maybe fancy table?
content
.gsub(/\n\* (\S+):/,"\n* `\\1`:")
.gsub(/(?<=to the |to )(\$.)/,'`\1`')
.sub('`$s`tderr','`$stderr`')
.sub('`$``'){'<code class="highlighter-rouge">$`</code>'}# Without block, '`' has special meaning in sub
.gsub(/(?<!`)\$(stdin|stdout|DEBUG|VERBOSE)/,'`$\1`')
when'core/Class.md'
content.sub('of the class `Class`.',"of the class `Class`.\n")# To properly render the diagram after
when'core/Float.md'
content.sub("wiki-floats_i\n mprecise\n\n","wiki-floats_imprecise\n")
when'lib/cgi/CGI.md'
content.sub("[cgi]('field_name')","`cgi['field_name']`")
when'lib/irb/IRB.md'
content.sub('[IRB.conf](:IRB_RC)','`IRB.conf[:IRB_RC]`')
when'lib/prime/Prime.md'
content
.sub("e.g.\n Prime","e.g.\n\n Prime")
.sub('`Prime`.instance','`Prime.instance`')
.gsub('`Prime`::`','`Prime::')
when'core/String\.md'
content
.gsub("``!''","`!`")
when%r{ext/date/}
content.sub(/\A.+?\n/,'')# We don't need first header inserted by render script.
when'core/MatchData.md'
# Somebody have broken links to different parts of Regex:
# https://ruby-doc.org/core-2.7.0/MatchData.html#class-MatchData-label-Global+variables+equivalence
content
.sub('Regexp.last_match;','`Regexp.last_match`;')
.sub('[Regexp.last_match](0)','`Regexp.last_match[0]`')
.sub('[Regexp.last_match](i)','`Regexp.last_match[i]`')
.sub('Regexp.last_match`.pre_match`','`Regexp.last_match.pre_match`')
.sub('Regexp.last_match`.post_match`','`Regexp.last_match.post_match`')
.sub('[Regexp.last_match](-1)','`Regexp.last_match[-1]`')

# FIXME: probably could be handled generically (line with a lot of spaces after line without spaces):
when'lib/abbrev/Abbrev.md'
content.gsub("*Generates:*\n","*Generates:*\n\n")
when'lib/logger/Logger.md'
content.gsub(/(Log (format|sample)):\n/,"\\1\n\n")
when'lib/pp/PP.md'
content.gsub("returns this:\n","returns this:\n\n")
when'lib/optparse/OptionParser.md'
content.gsub(/((Used|output):)\n(?= bash-3\.2)/,"\\1\n\n")
when'lib/English/English.md'
content.gsub(/^\* (\$.+?): (\$.+?)$/,'* `\1`: `\2`')
.sub('`$``'){'<code class="highlighter-rouge">$`</code>'}# Without block, '`' has special meaning in sub

# TODO:
# when %r{_special/kernel\.md}
# content.gsub(/ref:\`(.+?)\`/, 'ref:\1')
else
content
end
end

defper_site_file(path,content)
casepath
when'ruby-lang.org/en/documentation/installation/index.md'
content.sub(/\*\[Package Management Systems.+\(\#building-from-source\)/m,'')
else
content
end
end
end