admin/brew
1
0
Fork 0
Code Issues Packages Projects Releases 6 Wiki Activity

Strategy: Use Curl output parsing methods

Browse Source
This commit is contained in:
Sam Ford 2021-05-04 16:35:21 -04:00
parent 13e3f704c9
commit 1c4faaa5d5
No known key found for this signature in database
GPG Key ID: 95209E46C7FFDEFE
1 changed files with 9 additions and 37 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

46
Library/Homebrew/livecheck/strategy.rb
View File

@ -72,11 +72,6 @@ module Homebrew
retries: 0, retries: 0,
}.freeze }.freeze
# HTTP response head(s) and body are typically separated by a double
# `CRLF` (whereas HTTP header lines are separated by a single `CRLF`).
# In rare cases, this can also be a double newline (`\n\n`).
HTTP_HEAD_BODY_SEPARATOR = "\r\n\r\n"
# A regex used to identify a tarball extension at the end of a string. # A regex used to identify a tarball extension at the end of a string.
TARBALL_EXTENSION_REGEX = / TARBALL_EXTENSION_REGEX = /
\.t \.t
@ -180,22 +175,17 @@ module Homebrew
headers = [] headers = []
[:default, :browser].each do |user_agent| [:default, :browser].each do |user_agent|
stdout, _, status = curl_with_workarounds( output, _, status = curl_with_workarounds(
*PAGE_HEADERS_CURL_ARGS, url, *PAGE_HEADERS_CURL_ARGS, url,
**DEFAULT_CURL_OPTIONS, **DEFAULT_CURL_OPTIONS,
use_homebrew_curl: homebrew_curl, use_homebrew_curl: homebrew_curl,
user_agent: user_agent user_agent: user_agent
) )
next unless status.success?
while stdout.match?(/\AHTTP.*\r$/) parsed_output = parse_curl_output(output)
h, stdout = stdout.split("\r\n\r\n", 2) parsed_output[:responses].each { |response| headers << response[:headers] }
break if headers.present?
headers << h.split("\r\n").drop(1)
.to_h { |header| header.split(/:\s*/, 2) }
.transform_keys(&:downcase)
end
return headers if status.success?
end end
headers headers
@ -211,8 +201,6 @@ module Homebrew
# @return [Hash] # @return [Hash]
sig { params(url: String, homebrew_curl: T::Boolean).returns(T::Hash[Symbol, T.untyped]) } sig { params(url: String, homebrew_curl: T::Boolean).returns(T::Hash[Symbol, T.untyped]) }
def self.page_content(url, homebrew_curl: false) def self.page_content(url, homebrew_curl: false)
original_url = url
stderr = nil stderr = nil
[:default, :browser].each do |user_agent| [:default, :browser].each do |user_agent|
stdout, stderr, status = curl_with_workarounds( stdout, stderr, status = curl_with_workarounds(
@ -229,27 +217,11 @@ module Homebrew
# Separate the head(s)/body and identify the final URL (after any # Separate the head(s)/body and identify the final URL (after any
# redirections) # redirections)
max_iterations = 5 parsed_output = parse_curl_output(output)
iterations = 0 final_url = curl_response_last_location(parsed_output[:responses], absolutize: true, base_url: url)
output = output.lstrip
while output.match?(%r{\AHTTP/[\d.]+ \d+}) && output.include?(HTTP_HEAD_BODY_SEPARATOR)
iterations += 1
raise "Too many redirects (max = #{max_iterations})" if iterations > max_iterations
head_text, _, output = output.partition(HTTP_HEAD_BODY_SEPARATOR) data = { content: parsed_output[:body] }
output = output.lstrip data[:final_url] = final_url if final_url.present? && final_url != url
location = head_text[/^Location:\s*(.*)$/i, 1]
next if location.blank?
location.chomp!
# Convert a relative redirect URL to an absolute URL
location = URI.join(url, location) unless location.match?(PageMatch::URL_MATCH_REGEX)
final_url = location
end
data = { content: output }
data[:final_url] = final_url if final_url.present? && final_url != original_url
return data return data
end end