IOError, 'Not a gzipped file'

[DharmeshPandav]

while trying to access sitemap from robots.txt , Scrapy fails with IOError, 'Not a gzipped file' error

not sure if this issue is related to following issue(s)
#193 -> closed issue
#660 -> merged pull request to address issue 193
#951 -> open issue

line where code fails in gzip.py at line # 197

def _read_gzip_header(self):
        magic = self.fileobj.read(2)
        if magic != '\037\213':
            raise IOError, 'Not a gzipped file'

Response Header

Content-Encoding: gzip
Accept-Ranges: bytes
X-Amz-Request-Id: BFFF010DDE6268DA
Vary: Accept-Encoding
Server: AmazonS3
Last-Modified: Wed, 15 Jun 2016 19:02:20 GMT
Etag: "300bb71d6897cb2a22bba0bd07978c84"
Cache-Control: no-transform
Date: Sun, 19 Jun 2016 10:54:53 GMT
Content-Type: binary/octet-stream

Error Log:

 Traceback (most recent call last):
  File "c:\venv\scrapy1.0\lib\site-packages\scrapy\utils\defer.py", line 102, in iter_errback
    yield next(it)
  File "c:\venv\scrapy1.0\lib\site-packages\scrapy\spidermiddlewares\offsite.py", line 29, in process_spider_output
    for x in result:
  File "c:\venv\scrapy1.0\lib\site-packages\scrapy\spidermiddlewares\referer.py", line 22, in <genexpr>
    return (_set_referer(r) for r in result or ())
  File "c:\venv\scrapy1.0\lib\site-packages\scrapy\spidermiddlewares\urllength.py", line 37, in <genexpr>
    return (r for r in result or () if _filter(r))
  File "c:\venv\scrapy1.0\lib\site-packages\scrapy\spidermiddlewares\depth.py", line 58, in <genexpr>
    return (r for r in result or () if _filter(r))
  File "D:\projects\sitemap_spider\sitemap_spider\spiders\mainspider.py", line 31, in _parse_sitemap
    body = self._get_sitemap_body(response)
  File "c:\venv\scrapy1.0\lib\site-packages\scrapy\spiders\sitemap.py", line 67, in _get_sitemap_body
    return gunzip(response.body)
  File "c:\venv\scrapy1.0\lib\site-packages\scrapy\utils\gz.py", line 37, in gunzip
    chunk = read1(f, 8196)
  File "c:\venv\scrapy1.0\lib\site-packages\scrapy\utils\gz.py", line 21, in read1
    return gzf.read(size)
  File "c:\python27\Lib\gzip.py", line 268, in read
    self._read(readsize)
  File "c:\python27\Lib\gzip.py", line 303, in _read
    self._read_gzip_header()
  File "c:\python27\Lib\gzip.py", line 197, in _read_gzip_header
    raise IOError, 'Not a gzipped file'

i did download file manually and was able to extract the content so it is not like file is corrupted

as an example sitemap url : you can follow amazon robots.txt

[DharmeshPandav]

as a work around to this problem
if one is to set COMPRESSION_ENABLED = False in settings.py file, it will work as expected

[redapple]

Thanks @DharmeshPandav for reminding us of this issue.
It's the same problem as @juraseg commented on in #951 (comment)
For http://www.amazon.fr/sitemaps.f3053414d236e84.SitemapIndex_0.xml.gz for example you get

{b'Accept-Ranges': [b'bytes'],
 b'Content-Encoding': [b'gzip'],
 b'Content-Type': [b'binary/octet-stream'],
 b'Date': [b'Mon, 20 Jun 2016 10:50:16 GMT'],
 b'Etag': [b'"48f5a0d2cfff8d96700c49b5799c6d35"'],
 b'Last-Modified': [b'Wed, 15 Jun 2016 19:29:30 GMT'],
 b'Server': [b'AmazonS3'],
 b'Vary': [b'Accept-Encoding'],
 b'X-Amz-Request-Id': [b'674AEA9C05B9A833']}

from the web server, with HttpCompressionMiddleware decompressing the response based on Content-Encoding being "gzip".

When reading IANA's page on application/octet-stream

The recommended action for an implementation that receives an
"application/octet-stream" entity is to simply offer to put the data
in a file, with any Content-Transfer-Encoding undone, or perhaps to
use it as input to a user-specified process.

I'd go for considering Content-Type: binary/octet-stream the same way, taking precedence over Content-Encoding (for the same reasons as in #193 (comment)), i.e. not try to decompress it at HttpCompressionMiddleware level as a gzipped HTML/XML/JSON... response, but like a (compressed) file for another layer to interpret (here, SitemapSpider)

[redapple]

With #2389, I'm now reconsidering what I wrote earlier :)