X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/848723ea972c09f28787db91d8c06e98a274ab89..83c8b40add31fb501d587b64fb17a936ffd5dbc5:/youtube_dl/downloader/http.py diff --git a/youtube_dl/downloader/http.py b/youtube_dl/downloader/http.py index 748f9f3..af405b9 100644 --- a/youtube_dl/downloader/http.py +++ b/youtube_dl/downloader/http.py @@ -1,15 +1,21 @@ +from __future__ import unicode_literals + +import errno import os +import socket import time +import re from .common import FileDownloader +from ..compat import compat_urllib_error from ..utils import ( - compat_urllib_request, - compat_urllib_error, ContentTooShortError, - encodeFilename, sanitize_open, - format_bytes, + sanitized_Request, + write_xattr, + XAttrMetadataError, + XAttrUnavailableError, ) @@ -21,13 +27,16 @@ class HttpFD(FileDownloader): # Do not include the Accept-Encoding header headers = {'Youtubedl-no-compression': 'True'} - if 'user_agent' in info_dict: - headers['Youtubedl-user-agent'] = info_dict['user_agent'] - basic_request = compat_urllib_request.Request(url, None, headers) - request = compat_urllib_request.Request(url, None, headers) + add_headers = info_dict.get('http_headers') + if add_headers: + headers.update(add_headers) + basic_request = sanitized_Request(url, None, headers) + request = sanitized_Request(url, None, headers) - if self.params.get('test', False): - request.add_header('Range', 'bytes=0-10240') + is_test = self.params.get('test', False) + + if is_test: + request.add_header('Range', 'bytes=0-%s' % str(self._TEST_FILE_SIZE - 1)) # Establish possible resume length if os.path.isfile(encodeFilename(tmpfilename)): @@ -37,7 +46,7 @@ class HttpFD(FileDownloader): open_mode = 'wb' if resume_len != 0: - if self.params.get('continuedl', False): + if self.params.get('continuedl', True): self.report_resuming_byte(resume_len) request.add_header('Range', 'bytes=%d-' % resume_len) open_mode = 'ab' @@ -49,7 +58,25 @@ class HttpFD(FileDownloader): while count <= retries: # Establish connection try: - data = compat_urllib_request.urlopen(request) + data = self.ydl.urlopen(request) + # When trying to resume, Content-Range HTTP header of response has to be checked + # to match the value of requested Range HTTP header. This is due to a webservers + # that don't support resuming and serve a whole file with no Content-Range + # set in response despite of requested Range (see + # https://github.com/rg3/youtube-dl/issues/6057#issuecomment-126129799) + if resume_len > 0: + content_range = data.headers.get('Content-Range') + if content_range: + content_range_m = re.search(r'bytes (\d+)-', content_range) + # Content-Range is present and matches requested Range, resume is possible + if content_range_m and resume_len == int(content_range_m.group(1)): + break + # Content-Range is either not present or invalid. Assuming remote webserver is + # trying to send the whole file, resume is not possible, so wiping the local file + # and performing entire redownload + self.report_unable_to_resume() + resume_len = 0 + open_mode = 'wb' break except (compat_urllib_error.HTTPError, ) as err: if (err.code < 500 or err.code >= 600) and err.code != 416: @@ -59,7 +86,7 @@ class HttpFD(FileDownloader): # Unable to resume (requested range not satisfiable) try: # Open the connection again without the range header - data = compat_urllib_request.urlopen(basic_request) + data = self.ydl.urlopen(basic_request) content_length = data.info()['Content-Length'] except (compat_urllib_error.HTTPError, ) as err: if err.code < 500 or err.code >= 600: @@ -80,48 +107,69 @@ class HttpFD(FileDownloader): self._hook_progress({ 'filename': filename, 'status': 'finished', + 'downloaded_bytes': resume_len, + 'total_bytes': resume_len, }) return True else: # The length does not match, we start the download over self.report_unable_to_resume() + resume_len = 0 open_mode = 'wb' break + except socket.error as e: + if e.errno != errno.ECONNRESET: + # Connection reset is no problem, just retry + raise + # Retry count += 1 if count <= retries: self.report_retry(count, retries) if count > retries: - self.report_error(u'giving up after %s retries' % retries) + self.report_error('giving up after %s retries' % retries) return False data_len = data.info().get('Content-length', None) + + # Range HTTP header may be ignored/unsupported by a webserver + # (e.g. extractor/scivee.py, extractor/bambuser.py). + # However, for a test we still would like to download just a piece of a file. + # To achieve this we limit data_len to _TEST_FILE_SIZE and manually control + # block size when downloading a file. + if is_test and (data_len is None or int(data_len) > self._TEST_FILE_SIZE): + data_len = self._TEST_FILE_SIZE + if data_len is not None: data_len = int(data_len) + resume_len - min_data_len = self.params.get("min_filesize", None) - max_data_len = self.params.get("max_filesize", None) + min_data_len = self.params.get('min_filesize') + max_data_len = self.params.get('max_filesize') if min_data_len is not None and data_len < min_data_len: - self.to_screen(u'\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len)) + self.to_screen('\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len)) return False if max_data_len is not None and data_len > max_data_len: - self.to_screen(u'\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len)) + self.to_screen('\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len)) return False - data_len_str = format_bytes(data_len) byte_counter = 0 + resume_len block_size = self.params.get('buffersize', 1024) start = time.time() + + # measure time over whole while-loop, so slow_down() and best_block_size() work together properly + now = None # needed for slow_down() in the first loop run + before = start # start measuring while True: + # Download and write - before = time.time() - data_block = data.read(block_size) - after = time.time() + data_block = data.read(block_size if not is_test else min(block_size, data_len - byte_counter)) + byte_counter += len(data_block) + + # exit loop when download is finished if len(data_block) == 0: break - byte_counter += len(data_block) - # Open file just in time + # Open destination file just in time if stream is None: try: (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode) @@ -129,45 +177,63 @@ class HttpFD(FileDownloader): filename = self.undo_temp_name(tmpfilename) self.report_destination(filename) except (OSError, IOError) as err: - self.report_error(u'unable to open for writing: %s' % str(err)) + self.report_error('unable to open for writing: %s' % str(err)) return False + + if self.params.get('xattr_set_filesize', False) and data_len is not None: + try: + write_xattr(tmpfilename, 'user.ytdl.filesize', str(data_len).encode('utf-8')) + except (XAttrUnavailableError, XAttrMetadataError) as err: + self.report_error('unable to set filesize xattr: %s' % str(err)) + try: stream.write(data_block) except (IOError, OSError) as err: - self.to_stderr(u"\n") - self.report_error(u'unable to write data: %s' % str(err)) + self.to_stderr('\n') + self.report_error('unable to write data: %s' % str(err)) return False + + # Apply rate limit + self.slow_down(start, now, byte_counter - resume_len) + + # end measuring of one loop run + now = time.time() + after = now + + # Adjust block size if not self.params.get('noresizebuffer', False): block_size = self.best_block_size(after - before, len(data_block)) + before = after + # Progress message - speed = self.calc_speed(start, time.time(), byte_counter - resume_len) + speed = self.calc_speed(start, now, byte_counter - resume_len) if data_len is None: - eta = percent = None + eta = None else: - percent = self.calc_percent(byte_counter, data_len) eta = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len) - self.report_progress(percent, data_len_str, speed, eta) self._hook_progress({ + 'status': 'downloading', 'downloaded_bytes': byte_counter, 'total_bytes': data_len, 'tmpfilename': tmpfilename, 'filename': filename, - 'status': 'downloading', 'eta': eta, 'speed': speed, + 'elapsed': now - start, }) - # Apply rate limit - self.slow_down(start, byte_counter - resume_len) + if is_test and byte_counter == data_len: + break if stream is None: - self.to_stderr(u"\n") - self.report_error(u'Did not get any data blocks') + self.to_stderr('\n') + self.report_error('Did not get any data blocks') return False - stream.close() - self.report_finish(data_len_str, (time.time() - start)) + if tmpfilename != '-': + stream.close() + if data_len is not None and byte_counter != data_len: raise ContentTooShortError(byte_counter, int(data_len)) self.try_rename(tmpfilename, filename) @@ -181,6 +247,7 @@ class HttpFD(FileDownloader): 'total_bytes': byte_counter, 'filename': filename, 'status': 'finished', + 'elapsed': time.time() - start, }) return True