Module: Utils::Curl Private

Included in:
CurlDownloadStrategy, GitHubReleases, SPDX, SPDX, SharedAudits, SharedAudits
Defined in:
utils/curl.rb

Overview

This module is part of a private API. This module may only be used in the Homebrew/brew repository. Third parties should avoid using this module if possible, as it may be removed or changed without warning.

Helper function for interacting with curl.

Class Method Summary collapse

Class Method Details

.clear_path_cacheObject

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.



45
46
47
# File 'utils/curl.rb', line 45

def clear_path_cache
  @curl_path = nil
end

.curl(*args, print_stdout: true, **options) ⇒ Object

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.



162
163
164
165
166
# File 'utils/curl.rb', line 162

def curl(*args, print_stdout: true, **options)
  result = curl_with_workarounds(*args, print_stdout: print_stdout, **options)
  result.assert_success!
  result
end

.curl_args(*extra_args, connect_timeout: nil, max_time: nil, retries: Homebrew::EnvConfig.curl_retries.to_i, retry_max_time: nil, show_output: false, user_agent: nil) ⇒ Array<T.untyped>

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.

Parameters:

  • extra_args (T.untyped)
  • connect_timeout (Integer, Float, nil) (defaults to: nil)
  • max_time (Integer, Float, nil) (defaults to: nil)
  • retries (Integer, nil) (defaults to: Homebrew::EnvConfig.curl_retries.to_i)
  • retry_max_time (Integer, Float, nil) (defaults to: nil)
  • show_output (Boolean, nil) (defaults to: false)
  • user_agent (String, Symbol, nil) (defaults to: nil)

Returns:

  • (Array<T.untyped>)


60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
# File 'utils/curl.rb', line 60

def curl_args(
  *extra_args,
  connect_timeout: nil,
  max_time: nil,
  retries: Homebrew::EnvConfig.curl_retries.to_i,
  retry_max_time: nil,
  show_output: false,
  user_agent: nil
)
  args = []

  # do not load .curlrc unless requested (must be the first argument)
  args << "--disable" unless Homebrew::EnvConfig.curlrc?

  # echo any cookies received on a redirect
  args << "--cookie" << "/dev/null"

  args << "--globoff"

  args << "--show-error"

  args << "--user-agent" << case user_agent
  when :browser, :fake
    HOMEBREW_USER_AGENT_FAKE_SAFARI
  when :default, nil
    HOMEBREW_USER_AGENT_CURL
  when String
    user_agent
  else
    raise TypeError, ":user_agent must be :browser/:fake, :default, or a String"
  end

  args << "--header" << "Accept-Language: en"

  unless show_output == true
    args << "--fail"
    args << "--progress-bar" unless Context.current.verbose?
    args << "--verbose" if Homebrew::EnvConfig.curl_verbose?
    args << "--silent" unless $stdout.tty?
  end

  args << "--connect-timeout" << connect_timeout.round(3) if connect_timeout.present?
  args << "--max-time" << max_time.round(3) if max_time.present?

  # A non-positive integer (e.g., 0) or `nil` will omit this argument
  args << "--retry" << retries if retries&.positive?

  args << "--retry-max-time" << retry_max_time.round if retry_max_time.present?

  args + extra_args
end

.curl_check_http_content(url, url_type, specs: {}, user_agents: [:default], check_content: false, strict: false, use_homebrew_curl: false) ⇒ Object

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.



239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
# File 'utils/curl.rb', line 239

def curl_check_http_content(url, url_type, specs: {}, user_agents: [:default],
                            check_content: false, strict: false, use_homebrew_curl: false)
  return unless url.start_with? "http"

  secure_url = url.sub(/\Ahttp:/, "https:")
  secure_details = nil
  hash_needed = false
  if url != secure_url
    user_agents.each do |user_agent|
      secure_details = begin
        curl_http_content_headers_and_checksum(
          secure_url,
          specs:             specs,
          hash_needed:       true,
          use_homebrew_curl: use_homebrew_curl,
          user_agent:        user_agent,
        )
      rescue Timeout::Error
        next
      end

      next unless http_status_ok?(secure_details[:status_code])

      hash_needed = true
      user_agents = [user_agent]
      break
    end
  end

  details = nil
  user_agents.each do |user_agent|
    details =
      curl_http_content_headers_and_checksum(
        url,
        specs:             specs,
        hash_needed:       hash_needed,
        use_homebrew_curl: use_homebrew_curl,
        user_agent:        user_agent,
      )
    break if http_status_ok?(details[:status_code])
  end

  unless details[:status_code]
    # Hack around https://github.com/Homebrew/brew/issues/3199
    return if MacOS.version == :el_capitan

    return "The #{url_type} #{url} is not reachable"
  end

  unless http_status_ok?(details[:status_code])
    return if details[:responses].any? do |response|
      url_protected_by_cloudflare?(response) || url_protected_by_incapsula?(response)
    end

    # https://github.com/Homebrew/brew/issues/13789
    # If the `:homepage` of a formula is private, it will fail an `audit`
    # since there's no way to specify a `strategy` with `using:` and
    # GitHub does not authorize access to the web UI using token
    #
    # Strategy:
    # If the `:homepage` 404s, it's a GitHub link, and we have a token then
    # check the API (which does use tokens) for the repository
    repo_details = url.match(%r{https?://github\.com/(?<user>[^/]+)/(?<repo>[^/]+)/?.*})
    check_github_api = url_type == SharedAudits::URL_TYPE_HOMEPAGE &&
                       details[:status_code] == "404" &&
                       repo_details &&
                       Homebrew::EnvConfig.github_api_token

    unless check_github_api
      return "The #{url_type} #{url} is not reachable (HTTP status code #{details[:status_code]})"
    end

    "Unable to find homepage" if SharedAudits.github_repo_data(repo_details[:user], repo_details[:repo]).nil?
  end

  if url.start_with?("https://") && Homebrew::EnvConfig.no_insecure_redirect? &&
     (details[:final_url].present? && !details[:final_url].start_with?("https://"))
    return "The #{url_type} #{url} redirects back to HTTP"
  end

  return unless secure_details

  return if !http_status_ok?(details[:status_code]) || !http_status_ok?(secure_details[:status_code])

  etag_match = details[:etag] &&
               details[:etag] == secure_details[:etag]
  content_length_match =
    details[:content_length] &&
    details[:content_length] == secure_details[:content_length]
  file_match = details[:file_hash] == secure_details[:file_hash]

  http_with_https_available =
    url.start_with?("http://") &&
    (secure_details[:final_url].present? && secure_details[:final_url].start_with?("https://"))

  if (etag_match || content_length_match || file_match) && http_with_https_available
    return "The #{url_type} #{url} should use HTTPS rather than HTTP"
  end

  return unless check_content

  no_protocol_file_contents = %r{https?:\\?/\\?/}
  http_content = details[:file]&.scrub&.gsub(no_protocol_file_contents, "/")
  https_content = secure_details[:file]&.scrub&.gsub(no_protocol_file_contents, "/")

  # Check for the same content after removing all protocols
  if (http_content && https_content) && (http_content == https_content) && http_with_https_available
    return "The #{url_type} #{url} should use HTTPS rather than HTTP"
  end

  return unless strict

  # Same size, different content after normalization
  # (typical causes: Generated ID, Timestamp, Unix time)
  if http_content.length == https_content.length
    return "The #{url_type} #{url} may be able to use HTTPS rather than HTTP. Please verify it in a browser."
  end

  lenratio = (100 * https_content.length / http_content.length).to_i
  return unless (90..110).cover?(lenratio)

  "The #{url_type} #{url} may be able to use HTTPS rather than HTTP. Please verify it in a browser."
end

.curl_download(*args, to: nil, try_partial: false, **options) ⇒ Object

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.



168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
# File 'utils/curl.rb', line 168

def curl_download(*args, to: nil, try_partial: false, **options)
  destination = Pathname(to)
  destination.dirname.mkpath

  if try_partial
    range_stdout = curl_output("--location", "--head", *args, **options).stdout
    parsed_output = parse_curl_output(range_stdout)

    headers = if parsed_output[:responses].present?
      parsed_output[:responses].last[:headers]
    else
      {}
    end

    # Any value for `accept-ranges` other than none indicates that the server supports partial requests.
    # Its absence indicates no support.
    supports_partial = headers.key?("accept-ranges") && headers["accept-ranges"] != "none"

    if supports_partial &&
       destination.exist? &&
       destination.size == headers["content-length"].to_i
      return # We've already downloaded all the bytes
    end
  end

  args = ["--location", "--remote-time", "--output", destination, *args]
  # continue-at shouldn't be used with servers that don't support partial requests.
  args = ["--continue-at", "-", *args] if destination.exist? && supports_partial

  curl(*args, **options)
end

.curl_executable(use_homebrew_curl: false) ⇒ Object

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.



35
36
37
38
39
# File 'utils/curl.rb', line 35

def curl_executable(use_homebrew_curl: false)
  return HOMEBREW_BREWED_CURL_PATH if use_homebrew_curl

  @curl_executable ||= HOMEBREW_SHIMS_PATH/"shared/curl"
end

.curl_http_content_headers_and_checksum(url, specs: {}, hash_needed: false, use_homebrew_curl: false, user_agent: :default) ⇒ Object

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.



363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
# File 'utils/curl.rb', line 363

def curl_http_content_headers_and_checksum(
  url, specs: {}, hash_needed: false,
  use_homebrew_curl: false, user_agent: :default
)
  file = Tempfile.new.tap(&:close)

  # Convert specs to options. This is mostly key-value options,
  # unless the value is a boolean in which case treat as as flag.
  specs = specs.flat_map do |option, argument|
    next [] if argument == false # No flag.

    args = ["--#{option.to_s.tr("_", "-")}"]
    args << argument unless argument == true # It's a flag.
    args
  end

  max_time = hash_needed ? 600 : 25
  output, _, status = curl_output(
    *specs, "--dump-header", "-", "--output", file.path, "--location", url,
    use_homebrew_curl: use_homebrew_curl,
    connect_timeout:   15,
    max_time:          max_time,
    retry_max_time:    max_time,
    user_agent:        user_agent
  )

  parsed_output = parse_curl_output(output)
  responses = parsed_output[:responses]

  final_url = curl_response_last_location(responses)
  headers = if responses.last.present?
    status_code = responses.last[:status_code]
    responses.last[:headers]
  else
    {}
  end
  etag = headers["etag"][ETAG_VALUE_REGEX, 1] if headers["etag"].present?
  content_length = headers["content-length"]

  if status.success?
    open_args = {}
    # Try to get encoding from Content-Type header
    # TODO: add guessing encoding by <meta http-equiv="Content-Type" ...> tag
    if (content_type = headers["content-type"]) &&
       (match = content_type.match(/;\s*charset\s*=\s*([^\s]+)/)) &&
       (charset = match[1])
      begin
        open_args[:encoding] = Encoding.find(charset)
      rescue ArgumentError
        # Unknown charset in Content-Type header
      end
    end
    file_contents = File.read(file.path, open_args)
    file_hash = Digest::SHA2.hexdigest(file_contents) if hash_needed
  end

  {
    url:            url,
    final_url:      final_url,
    status_code:    status_code,
    headers:        headers,
    etag:           etag,
    content_length: content_length,
    file:           file_contents,
    file_hash:      file_hash,
    responses:      responses,
  }
ensure
  file.unlink
end

.curl_output(*args, **options) ⇒ Object

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.



200
201
202
# File 'utils/curl.rb', line 200

def curl_output(*args, **options)
  curl_with_workarounds(*args, print_stderr: false, show_output: true, **options)
end

.curl_pathObject

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.



41
42
43
# File 'utils/curl.rb', line 41

def curl_path
  @curl_path ||= Utils.popen_read(curl_executable, "--homebrew=print-path").chomp.presence
end

.curl_response_last_location(responses, absolutize: false, base_url: nil) ⇒ String?

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.

Returns the URL from the last location header found in cURL responses, if any.

Parameters:

  • responses (Array<Hash{Symbol => T.untyped}>)

    An array of hashes containing response status information and headers from #parse_curl_response.

  • absolutize (Boolean) (defaults to: false)

    Whether to make the location URL absolute.

  • base_url (String, nil) (defaults to: nil)

    The URL to use as a base for making the location URL absolute.

Returns:

  • (String, nil)

    The URL from the last-occurring location header in the responses or nil (if no location headers found).



495
496
497
498
499
500
501
502
503
504
505
506
507
# File 'utils/curl.rb', line 495

def curl_response_last_location(responses, absolutize: false, base_url: nil)
  responses.reverse_each do |response|
    next if response[:headers].blank?

    location = response[:headers]["location"]
    next if location.blank?

    absolute_url = URI.join(base_url, location).to_s if absolutize && base_url.present?
    return absolute_url || location
  end

  nil
end

.curl_supports_tls13?Boolean

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.

Returns:

  • (Boolean)


434
435
436
437
438
439
# File 'utils/curl.rb', line 434

def curl_supports_tls13?
  @curl_supports_tls13 ||= Hash.new do |h, key|
    h[key] = quiet_system(curl_executable, "--tlsv1.3", "--head", "https://brew.sh/")
  end
  @curl_supports_tls13[curl_path]
end

.curl_with_workarounds(*args, secrets: nil, print_stdout: nil, print_stderr: nil, debug: nil, verbose: nil, env: {}, timeout: nil, use_homebrew_curl: false, **options) ⇒ Object

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.

Raises:

  • (Timeout::Error)


112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
# File 'utils/curl.rb', line 112

def curl_with_workarounds(
  *args,
  secrets: nil, print_stdout: nil, print_stderr: nil, debug: nil,
  verbose: nil, env: {}, timeout: nil, use_homebrew_curl: false, **options
)
  end_time = Time.now + timeout if timeout

  command_options = {
    secrets:      secrets,
    print_stdout: print_stdout,
    print_stderr: print_stderr,
    debug:        debug,
    verbose:      verbose,
  }.compact

  result = system_command curl_executable(use_homebrew_curl: use_homebrew_curl),
                          args:    curl_args(*args, **options),
                          env:     env,
                          timeout: end_time&.remaining,
                          **command_options

  return result if result.success? || !args.exclude?("--http1.1")

  raise Timeout::Error, result.stderr.lines.last.chomp if timeout && result.status.exitstatus == 28

  # Error in the HTTP2 framing layer
  if result.status.exitstatus == 16
    return curl_with_workarounds(
      *args, "--http1.1",
      timeout: end_time&.remaining, **command_options, **options
    )
  end

  # This is a workaround for https://github.com/curl/curl/issues/1618.
  if result.status.exitstatus == 56 # Unexpected EOF
    out = curl_output("-V").stdout

    # If `curl` doesn't support HTTP2, the exception is unrelated to this bug.
    return result unless out.include?("HTTP2")

    # The bug is fixed in `curl` >= 7.60.0.
    curl_version = out[/curl (\d+(\.\d+)+)/, 1]
    return result if Gem::Version.new(curl_version) >= Gem::Version.new("7.60.0")

    return curl_with_workarounds(*args, "--http1.1", **command_options, **options)
  end

  result
end

.http_status_ok?(status) ⇒ Boolean

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.

Returns:

  • (Boolean)


441
442
443
# File 'utils/curl.rb', line 441

def http_status_ok?(status)
  (100..299).cover?(status.to_i)
end

.parse_curl_output(output, max_iterations: 25) ⇒ Hash{Symbol => T.untyped}

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.

Separates the output text from curl into an array of HTTP responses and the final response body (i.e. content). Response hashes contain the :status_code, :status_text, and :headers.

Parameters:

  • output (String)

    The output text from curl containing HTTP responses, body content, or both.

  • max_iterations (Integer) (defaults to: 25)

    The maximum number of iterations for the while loop that parses HTTP response text. This should correspond to the maximum number of requests in the output. If curl’s --max-redirs option is used, max_iterations should be max-redirs + 1, to account for any final response after the redirections.

Returns:

  • (Hash{Symbol => T.untyped})

    A hash containing an array of response hashes and the body content, if found.



458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
# File 'utils/curl.rb', line 458

def parse_curl_output(output, max_iterations: 25)
  responses = []

  iterations = 0
  output = output.lstrip
  while output.match?(%r{\AHTTP/[\d.]+ \d+}) && output.include?(HTTP_RESPONSE_BODY_SEPARATOR)
    iterations += 1
    raise "Too many redirects (max = #{max_iterations})" if iterations > max_iterations

    response_text, _, output = output.partition(HTTP_RESPONSE_BODY_SEPARATOR)
    output = output.lstrip
    next if response_text.blank?

    response_text.chomp!
    response = parse_curl_response(response_text)
    responses << response if response.present?
  end

  { responses: responses, body: output }
end

.url_protected_by_cloudflare?(response) ⇒ Boolean

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.

Check if a URL is protected by CloudFlare (e.g. badlion.net and jaxx.io).

Parameters:

  • response (Hash{Symbol => T.untyped})

    A response hash from #parse_curl_response.

Returns:

  • (Boolean)

    Whether a response contains headers indicating that the URL is protected by Cloudflare.



209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
# File 'utils/curl.rb', line 209

def url_protected_by_cloudflare?(response)
  return false if response[:headers].blank?
  return false unless [403, 503].include?(response[:status_code].to_i)

  set_cookie_header = Array(response[:headers]["set-cookie"])
  has_cloudflare_cookie_header = set_cookie_header.compact.any? do |cookie|
    cookie.match?(/^(__cfduid|__cf_bm)=/i)
  end

  server_header = Array(response[:headers]["server"])
  has_cloudflare_server = server_header.compact.any? do |server|
    server.match?(/^cloudflare/i)
  end

  has_cloudflare_cookie_header && has_cloudflare_server
end

.url_protected_by_incapsula?(response) ⇒ Boolean

This method is part of a private API. This method may only be used in the Homebrew/brew repository. Third parties should avoid using this method if possible, as it may be removed or changed without warning.

Check if a URL is protected by Incapsula (e.g. corsair.com).

Parameters:

  • response (Hash{Symbol => T.untyped})

    A response hash from #parse_curl_response.

Returns:

  • (Boolean)

    Whether a response contains headers indicating that the URL is protected by Incapsula.



231
232
233
234
235
236
237
# File 'utils/curl.rb', line 231

def url_protected_by_incapsula?(response)
  return false if response[:headers].blank?
  return false if response[:status_code].to_i != 403

  set_cookie_header = Array(response[:headers]["set-cookie"])
  set_cookie_header.compact.any? { |cookie| cookie.match?(/^(visid_incap|incap_ses)_/i) }
end