ferrum_common 0.2.0 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e128a92c5b106f583d12cf8cc772b95637272dae11711e1095916e8485ccd9c5
-  data.tar.gz: c52cb59ea98be841986e6930e6a21456636491684cbe2d812f474dff996bb105
+  metadata.gz: '05923a936f2ca3b407d7313fbf4ad661ecf51eb1d1bab8870204d48c26e85564'
+  data.tar.gz: 1523dd2d0b471909b4424e61e314ee8b4def0ab9ec7e6894a8be2894b94d2ba2
 SHA512:
-  metadata.gz: aee22d614c6b08f217fcf0786083ef0b7a4aa0c154f79948916fe3433796c1255168df736ae6fdac82e3e0fd4d91ce309aaddadd0fc80e93087856e4cc355789
-  data.tar.gz: 6a2939158a8d3dc9da7938d63cad598fb005c88ef0307418ea9dd8a6f5eb3a8af34a1ad5c59257e23d7a99150a7c627520f9d26e1abe9c69ca4c3a0292f52546
+  metadata.gz: 0c3b3a806de76805483ae3bd4d129b59afd12e4267096ec95221a012e6935140fafb1cf0df2965a1756cd649c83ef725e1e8785091afd130f74da5a8ecd3f6af
+  data.tar.gz: 91037fe48c613a829a95deae97d7a6658f3b4a46f31a7301d4a43d6b79346c424b0e47c05ef63363af01b70fbdbccaedef1a35d365676773050d8655d9d21c7d

data/ferrum_common.gemspec

@@ -1,6 +1,6 @@
 Gem::Specification.new do |spec|
   spec.name         = "ferrum_common"
-  spec.version      = "0.2.0"
+  spec.version      = "0.4.0"
   spec.summary      = "[WIP] common useful extensions for ferrum or cuprite"
 
   spec.author       = "Victor Maslov aka Nakilon"
@@ -9,8 +9,8 @@ Gem::Specification.new do |spec|
   spec.metadata     = {"source_code_uri" => "https://github.com/nakilon/ferrum_common"}
 
   spec.add_dependency "ferrum"
-  spec.add_dependency "browser_reposition"
   spec.add_dependency "nokogiri"
+  spec.add_dependency "ffi"
   spec.required_ruby_version = ">=2.5"  # why?
 
   spec.files        = %w{ LICENSE ferrum_common.gemspec lib/ferrum_common.rb }

data/lib/ferrum_common.rb

@@ -1,44 +1,78 @@
 require "ferrum"
+::Ferrum::Node.module_eval do
+  def rect
+    ::Struct.new(:x, :y, :width, :height).new(*::JSON.load(page.evaluate(<<~HEREDOC, self)))
+      ( function(node) {
+        var x = scrollX, y = scrollY;
+        var rect = JSON.parse(JSON.stringify(node.getBoundingClientRect()));
+        rect.top += scrollY;
+        rect.left += scrollX;
+        var t = JSON.stringify( [rect.left, rect.top, rect.width, rect.height, ] );
+        return t;
+      } )(arguments[0])
+    HEREDOC
+  end
+end
+
 module FerrumCommon
 
   module Common
 
-    private def mhtml browser, timeout, mtd, msg = nil
-      Timeout.timeout(timeout){ yield }
-    rescue Timeout::Error
+    private def yield_with_timeout browser, timeout, mtd, msg = nil
+      ::Timeout.timeout(timeout){ yield }
+    rescue ::Timeout::Error
+      begin
       browser.mhtml path: "temp.mhtml"
-      STDERR.puts "dumped to ./temp.mhtml"
+      rescue ::Ferrum::BrowserError
+        sleep 1
+        browser.mhtml path: "temp.mhtml"
+      end
+      ::STDERR.puts "dumped to ./temp.mhtml"
       $!.backtrace.reject!{ |_| _[/\/gems\/concurrent-ruby-/] }
       $!.backtrace.reject!{ |_| _[/\/gems\/ferrum-/] }
-      raise Timeout::Error, "#{$!.to_s} after #{timeout} sec in #{mtd}#{" (#{msg.respond_to?(:call) ? msg.call : msg})" if msg}"
+      raise ::Timeout::Error, "#{$!.to_s} after #{timeout} sec in #{mtd}#{" (#{msg.respond_to?(:call) ? msg.call : msg})" if msg}"
     end
 
     def until_true timeout, msg = nil
-      mhtml self, timeout, __method__, msg do
+      yield_with_timeout self, timeout, __method__, msg do
         begin
           yield
         rescue Ferrum::NodeNotFoundError
           redo
-        end or (sleep timeout*0.1; redo)
+        end or (sleep timeout * 0.1; redo)
       end
     end
 
-    def until_one type, selector, timeout
+    def find_any type, selector, timeout
+      Timeout.timeout timeout do
+        t = public_method(type).call selector
+        return t unless t.empty?
+        sleep timeout * 0.1
+        redo
+      end
+    rescue Timeout::Error
+    end
+
+    def until_n type, selector, timeout, n, node = nil
       t = nil
-      mhtml self, timeout, __method__, ->{ "expected exactly one node for #{type} #{selector.inspect}, got #{t ? t.size : "none"}" } do
+      yield_with_timeout self, timeout, __method__, ->{ "expected exactly #{n} nodes for #{type} #{selector.inspect}, got #{t ? t.size : "none"}" } do
         t = begin
-          public_method(type).call selector
+          public_method(type).call selector, within: node
         end
-        unless 1 == t.size
+        unless n == t.size
           sleep timeout * 0.1
           redo
         end
       end
-      t.first
+      t
+    end
+
+    def until_one type, selector, timeout, node = nil
+      until_n(type, selector, timeout, 1, node).first
     end
 
     def abort msg_or_cause
-      # puts (msg_or_cause.respond_to?(:backtrace) ? msg_or_cause : Thread.current).backtrace
+      puts msg_or_cause
       puts (msg_or_cause.respond_to?(:full_message) ? msg_or_cause.full_message : Thread.current.backtrace)
       mhtml path: "temp.mhtml"
       STDERR.puts "dumped to ./temp.mhtml"
@@ -49,28 +83,100 @@ module FerrumCommon
   Ferrum::Page.include Common
   Ferrum::Frame.include Common
 
-  if "darwin" == Gem::Platform.local.os
-    require "browser_reposition"
-    Ferrum::Browser.include Common, BrowserReposition
-    def self.new **_
-      Ferrum::Browser.new(**_).tap(&:reposition)
+  module CoreGraphics
+    require "ffi"
+    extend ::FFI::Library
+    ffi_lib "/System/Library/Frameworks/CoreGraphics.framework/CoreGraphics"
+    attach_function :CGGetActiveDisplayList, [:uint32, :pointer, :pointer], :uint32
+    class CGRect < ::FFI::Struct
+      layout :x, :double,
+             :y, :double,
+             :width, :double,
+             :height, :double
+    end
+    attach_function :CGDisplayBounds, [:uint32], CGRect.by_value
+  end if "Darwin" == ::Etc.uname[:sysname]
+  def self.browser_reposition ferrum_browser, mtd, dx = 0, dy = 0
+    case ::Etc.uname[:sysname]
+    when "Darwin"
+      count_ptr = ::FFI::MemoryPointer.new :uint32
+      CoreGraphics.CGGetActiveDisplayList 0, nil, count_ptr
+      display_ids = ::FFI::MemoryPointer.new :uint32, count_ptr.read_uint32
+      CoreGraphics.CGGetActiveDisplayList count_ptr.read_uint32, display_ids, nil
+      displays = display_ids.read_array_of_uint32(count_ptr.read_uint32).map do |display_id|
+        CoreGraphics.CGDisplayBounds display_id
+      end
+      best = displays.map{ |_| _[:x] }.method(mtd).call
+      display = displays.find{ |_| best == _[:x] }
+      ferrum_browser.position = {left: display[:x] + dx, top: display[:y] + dy}
+    else
+      "#{::Etc.uname[:sysname].inspect} isn't supported yet"
     end
-  else
-    Ferrum::Browser.include Common
-    def self.new **_
-      Ferrum::Browser.new **_
+  end
+
+  def self.parse_mhtml mhtml_string
+    scanner = ::StringScanner.new mhtml_string
+    fail scanner.peek(400).inspect unless scanner.scan(/\AFrom: <Saved by Blink>\r
+Snapshot-Content-Location: \S+\r
+Subject:(?: \S.*\r\n)+Date: [A-Z][a-z][a-z], \d\d? [A-Z][a-z][a-z] 20\d\d \d\d:\d\d:\d\d [-+]0\d00\r
+MIME-Version: 1\.0\r
+Content-Type: multipart\/related;\r
+\ttype="text\/html";\r
+\tboundary="(----MultipartBoundary--[a-zA-Z0-9]{42}----)"\r\n\r\n\r\n--\1/)
+    delimeter = scanner[1]
+    fail unless scanner.charpos == prev = scanner.pos
+    while s = scanner.search_full(::Regexp.new(delimeter), true, true)
+      case doc = s[0...-delimeter.size-4]
+      when /\A\r\nContent-Type: text\/html\r
+Content-ID: <frame-[A-Z0-9]{32}@mhtml\.blink>\r
+Content-Transfer-Encoding: quoted-printable\r
+Content-Location: chrome-error:\/\/chromewebdata\/\r\n\r\n/,
+           /\A\r\nContent-Type: text\/html\r
+Content-ID: <frame-[A-Z0-9]{32}@mhtml\.blink>\r
+Content-Transfer-Encoding: quoted-printable\r\n\r\n/
+        # trash
+      when /\A\r\nContent-Type: text\/html\r
+Content-ID: <frame-[A-Z0-9]{32}@mhtml\.blink>\r
+Content-Transfer-Encoding: quoted-printable\r
+Content-Location: \S+\r\n\r\n/
+        # html
+      when /\A\r\nContent-Type: text\/css\r
+Content-Transfer-Encoding: quoted-printable\r
+Content-Location: \S+\r\n\r\n/
+        # css
+      when /\A\r\nContent-Type: image\/(webp|png|gif|jpeg)\r
+Content-Transfer-Encoding: base64\r
+Content-Location: (https?:\S+)\r\n\r\n(([+\/0-9A-Za-z=]{,76}\r\n)+)\z/
+        yield $1, $2, $3
+      when /\A\r\nContent-Type: binary\/octet-stream\r
+Content-Transfer-Encoding: base64\r
+Content-Location: https:\/\/\S+\r\n\r\n/
+        # binary
+      when /\A\r\nContent-Type: image\/svg\+xml\r
+Content-Transfer-Encoding: quoted-printable\r
+Content-Location: https?:\S+\r\n\r\n/
+        # svg
+      when /\A\r\nContent-Type: image\/gif\r
+Content-ID: <frame-[0-9A-F]{32}@mhtml\.blink>\r
+Content-Transfer-Encoding: base64\r
+Content-Location: https?:\S+\r\n\r\n/
+        # gif
+      else
+        STDERR.puts s[0..300]
+        fail
+      end
+      fail unless scanner.charpos == prev = scanner.pos
     end
   end
 
   # https://datatracker.ietf.org/doc/html/rfc2557
   # https://en.wikipedia.org/wiki/Quoted-printable
-  # require "strscan"
   require "nokogiri"  # Oga crashes on vk charset
-  def self.process_mhtml mht
-    scanner = ::StringScanner.new mht
-    fail scanner.peek(100).inspect unless scanner.scan(/\AFrom: <Saved by Blink>\r
+  def self.process_mhtml
+    scanner = ::StringScanner.new(mht = ARGF.read)
+    fail scanner.peek(400).inspect unless scanner.scan(/\AFrom: <Saved by Blink>\r
 Snapshot-Content-Location: \S+\r
-Subject:(?: \S+\r\n)+Date: [A-Z][a-z][a-z], \d\d? [A-Z][a-z][a-z] 20\d\d \d\d:\d\d:\d\d -0000\r
+Subject:(?: \S.*\r\n)+Date: [A-Z][a-z][a-z], \d\d? [A-Z][a-z][a-z] 20\d\d \d\d:\d\d:\d\d [-+]0\d00\r
 MIME-Version: 1\.0\r
 Content-Type: multipart\/related;\r
 \ttype="text\/html";\r
@@ -88,66 +194,77 @@ Content-Location: chrome-error:\/\/chromewebdata\/\r\n\r\n/,
            /\A\r\nContent-Type: text\/html\r
 Content-ID: <frame-[A-Z0-9]{32}@mhtml\.blink>\r
 Content-Transfer-Encoding: quoted-printable\r\n\r\n/
-        puts "trash #{$'.size}"
+        STDERR.puts "trash #{$'.size}"
         reps.push [prev-delimeter.size-2, scanner.pos-delimeter.size-4, "", ""]
       when /\A\r\nContent-Type: text\/html\r
 Content-ID: <frame-[A-Z0-9]{32}@mhtml\.blink>\r
 Content-Transfer-Encoding: quoted-printable\r
 Content-Location: \S+\r\n\r\n/
-        puts "html #{$'.size}"
+        STDERR.puts "html #{$'.size}"
         header = $&
-        t = $'.gsub(/=([0-9A-F][0-9A-F])/){ fail $1 unless "3D" == $1 || "20" == $1 || "0A" == $1 unless "80" <= $1; $1.hex.chr }.gsub("=\r\n", "")
-        puts "unpacked #{t.size}"
+        t = $'.gsub(/=([0-9A-F][0-9A-F])/){ fail $1 unless "3D" == $1 || "20" == $1 || "0A" == $1 || "09" == $1 unless "80" <= $1; $1.hex.chr }.gsub("=\r\n", "")
+        STDERR.puts "unpacked #{t.size}"
         html = ::Nokogiri::HTML t#.force_encoding "utf-8"
 
-        puts ".to_s.size #{html.to_s.size}"
+        STDERR.puts ".to_s.size #{html.to_s.size}"
 
         html.xpath("//*[not(*)]").group_by(&:name).
           map{ |_, g| [_, g.map(&:to_s).map(&:size).reduce(:+)] }.
-          sort_by(&:last).reverse.take(5).each &method(:p)
+          sort_by(&:last).reverse.take(5).each{ |_| STDERR.puts _.inspect }
 
         if block_given?
           yield html
-          puts "yielded"
-          puts "yield #{html.to_s.size}"
+          STDERR.puts "yielded"
+          STDERR.puts "yield #{html.to_s.size}"
         end
 
         reps.push [prev, scanner.pos-delimeter.size-4, header, html.to_s, true, :html]
       when /\A\r\nContent-Type: text\/css\r
 Content-Transfer-Encoding: quoted-printable\r
 Content-Location: \S+\r\n\r\n/
-        puts "css > #{$'.size}"
+        STDERR.puts "css > #{$'.size}"
         header = $&
         css = $'.gsub(/=([0-9A-F][0-9A-F])/){ fail $1 unless "3D" == $1 || "20" == $1 || "0A" == $1 unless "80" <= $1; $1.hex.chr }.gsub("=\r\n", "")
         css.gsub!(/[\r\n]+/, "\n")
 
-        puts "css < #{css.size}"
+        STDERR.puts "css < #{css.size}"
         reps.push [prev, scanner.pos-delimeter.size-4, header, css, true, :css]
 
       when /\A\r\nContent-Type: image\/(webp|png|gif|jpeg)\r
 Content-Transfer-Encoding: base64\r
-Content-Location: \S+\r\n\r\n/
-        puts "#{$1} #{$'.size}"
+Content-Location: https?:\S+\r\n\r\n/
+        STDERR.puts "#{$1} #{$'.size}"
+      when /\A\r\nContent-Type: binary\/octet-stream\r
+Content-Transfer-Encoding: base64\r
+Content-Location: https:\/\/\S+\r\n\r\n/
+        STDERR.puts "binary #{$'.size}"
       when /\A\r\nContent-Type: image\/svg\+xml\r
 Content-Transfer-Encoding: quoted-printable\r
-Content-Location: \S+\r\n\r\n/
-        puts "svg #{$'.size}"
+Content-Location: https?:\S+\r\n\r\n/
+        STDERR.puts "svg #{$'.size}"
+      when /\A\r\nContent-Type: image\/gif\r
+Content-ID: <frame-[0-9A-F]{32}@mhtml\.blink>\r
+Content-Transfer-Encoding: base64\r
+Content-Location: https?:\S+\r\n\r\n/
+        STDERR.puts "gif #{$'.size}"
       else
-        puts doc[0..300]
+        STDERR.puts doc[0..300]
         fail
       end
       fail unless scanner.charpos == prev = scanner.pos
     end
 
-    p is = reps.map.with_index{ |(_, _, _, _, _, type), i| i if :html == type }.compact
-    p cs = reps.map.with_index{ |(_, _, _, _, _, type), i| i if :css == type }.compact
+    is = reps.map.with_index{ |(_, _, _, _, _, type), i| i if :html == type }.compact
+    STDERR.puts is.inspect
+    cs = reps.map.with_index{ |(_, _, _, _, _, type), i| i if :css == type }.compact
+    STDERR.puts cs.inspect
     cs.each_cons(2){ |i,j| fail unless i+1==j }
     fail unless is == [cs[0]-1]
     File.write "temp.htm", reps[is[0]][3]
-    puts "css > #{File.size "temp.css"}"
     File.open("temp.css", "w"){ |f| cs.each{ |i| f.puts reps[i][3] } }
+    STDERR.puts "css > #{File.size "temp.css"}"
     system "uncss temp.htm -s temp.css -o out.css"
-    puts "css < #{File.size "out.css"}"
+    STDERR.puts "css < #{File.size "out.css"}"
     reps[cs[0]][1] = reps[cs[-1]][1]
     reps[cs[0]+1..cs[-1]] = []
     reps[cs[0]][3] = File.read "out.css"
@@ -158,11 +275,11 @@ Content-Location: \S+\r\n\r\n/
           b.gsub(/[\x80-\xFF]/n){ |_| "=%02X" % _.ord }.
           gsub(/.{73}[^=][^=](?=.)/, "\\0=\r\n") :
         header + str.gsub("\n", "\r\n")
-      p [str.size, "to - from = #{to - from}"]
+      STDERR.puts [str.size, "to - from = #{to - from}"].inspect
       mht[from...to] = str
     end
-    p ::File.write "temp.mht", mht
-    puts "OK"
+    puts mht
+    STDERR.puts "OK"
   end
 
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: ferrum_common
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.4.0
 platform: ruby
 authors:
 - Victor Maslov aka Nakilon
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-09-16 00:00:00.000000000 Z
+date: 2026-07-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ferrum
@@ -25,7 +25,7 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
-  name: browser_reposition
+  name: nokogiri
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
@@ -39,7 +39,7 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
-  name: nokogiri
+  name: ffi
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
@@ -52,7 +52,7 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-description: 
+description:
 email: nakilon@gmail.com
 executables: []
 extensions: []
@@ -61,12 +61,12 @@ files:
 - LICENSE
 - ferrum_common.gemspec
 - lib/ferrum_common.rb
-homepage: 
+homepage:
 licenses:
 - MIT
 metadata:
   source_code_uri: https://github.com/nakilon/ferrum_common
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -81,8 +81,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.3.25
-signing_key: 
+rubygems_version: 3.1.6
+signing_key:
 specification_version: 4
 summary: "[WIP] common useful extensions for ferrum or cuprite"
 test_files: []