External/wayback-machine-downloader
1
0
Fork 0
mirror of https://github.com/StrawberryMaster/wayback-machine-downloader.git synced 2025-12-19 02:36:38 +00:00
Code Issues Packages Projects Releases Wiki Activity

Setting file modified time to value reported by Wayback Machine

Browse Source 
Implements 937306712c564e5757d898feacc14fbabd10722d, fixes Maintain original creation/modified dates of files while downloading #174
This commit is contained in:
Felipe 2024-06-26 19:52:12 +00:00
parent 0a7752eb41
commit 509d7034a1
1 changed files with 7 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
lib/wayback_machine_downloader.rb
View File

@ -6,6 +6,7 @@ require 'open-uri'
require 'fileutils' require 'fileutils'
require 'cgi' require 'cgi'
require 'json' require 'json'
requite 'time'
require_relative 'wayback_machine_downloader/tidy_bytes' require_relative 'wayback_machine_downloader/tidy_bytes'
require_relative 'wayback_machine_downloader/to_regex' require_relative 'wayback_machine_downloader/to_regex'
require_relative 'wayback_machine_downloader/archive_api' require_relative 'wayback_machine_downloader/archive_api'
@ -264,6 +265,7 @@ class WaybackMachineDownloader
file_id = file_remote_info[:file_id] file_id = file_remote_info[:file_id]
file_timestamp = file_remote_info[:timestamp] file_timestamp = file_remote_info[:timestamp]
file_path_elements = file_id.split('/') file_path_elements = file_id.split('/')
original_file_mtime = nil
if file_id == "" if file_id == ""
dir_path = backup_path dir_path = backup_path
file_path = backup_path + 'index.html' file_path = backup_path + 'index.html'
@ -285,6 +287,11 @@ class WaybackMachineDownloader
begin begin
http.get(URI("https://web.archive.org/web/#{file_timestamp}id_/#{file_url}")) do |body| http.get(URI("https://web.archive.org/web/#{file_timestamp}id_/#{file_url}")) do |body|
file.write(body) file.write(body)
if uri.meta.has_key?("x-archive-orig-last-modified")
original_file_mtime = Time.parse(uri.meta["x-archive-orig-last-modified"])
end
end end
rescue OpenURI::HTTPError => e rescue OpenURI::HTTPError => e
puts "#{file_url} # #{e}" puts "#{file_url} # #{e}"