Scrapy 1.1.0rc3 release candidate is out!

[Paul Tremberth]

Hello, Scrapy users!

We are happy to announce that we just shipped Scrapy 1.1 Release Candidate 3!

It supersedes the shortlived (and wrongly deployed) 1.1.0rc2 from Feb 29th (so this announcement is nearly the same).

This release candidate fixes a handful of issues that were discovered with RC1 and Python 3.

Thanks to all contributors, and reporters of issues with 1.1rc1!

What's new in 1.1.0rc3 compared to 1.1.0rc1:

- new HTTP(S) download handler with updated TLS "context factory" (it does not force TLS 1.0 anymore)

this should solve most of the open issues users had with HTTPS connections

(works best with Twisted>=14 and OpenSSL>=0.14)

Let us know if you're still having SSL errors with this release.

And if you really need to force a TLS protocol version, use the new DOWNLOADER_CLIENT_TLS_METHOD setting.

- support for S3 in Python 3 with the help of botocore (https://github.com/boto/botocore)

- default ACL policy for S3 uploads is now "private" (was "public" previously):

This is backward incompatible, yes, but we felt it's safer not to expose uploaded files by default.

Please check if you require uploaded files to S3 to be public (there's a new setting to change the ACL policy -- FILES_STORE_S3_ACL)

- sending non-ASCII emails (but still no Python3 support for sending emails)

See the rest of the changes between 1.1.0rc1 and 1.1.0rc3 at this URL:

https://github.com/scrapy/scrapy/compare/1.1.0rc1...1.1.0rc3

It's available on PyPI, as universal wheel and source distribution.

Please give Scrapy 1.1.0rc3 a try, and report any problems at: https://github.com/scrapy/scrapy/issues

Install Scrapy 1.1.0rc3 via pip:

    $ pip install Scrapy==1.1.0rc3

Thank you, happy scraping!

Paul.