James Tucker [Sun, 21 Apr 2013 20:16:20 +0000 (13:16 -0700)]
Prevent signals from being sent to pid 0
* Closes #544
James Tucker [Fri, 8 Feb 2013 03:01:06 +0000 (19:01 -0800)]
Bump version number
James Tucker [Fri, 8 Feb 2013 02:58:07 +0000 (18:58 -0800)]
Update README for todays releases
James Tucker [Thu, 7 Feb 2013 00:25:22 +0000 (16:25 -0800)]
Prevent symlink path traversals
* Closes CVE-2013-0262
James Tucker [Thu, 7 Feb 2013 22:45:06 +0000 (14:45 -0800)]
Use secure_compare for hmac comparison
* Closes CVE-2013-0263
James Tucker [Wed, 6 Feb 2013 23:31:53 +0000 (15:31 -0800)]
Add secure_compare to Rack::Utils
Conflicts:
test/spec_utils.rb
James Tucker [Mon, 21 Jan 2013 21:30:16 +0000 (13:30 -0800)]
Use Dir.tmpdir instead of hardcoded /tmp
Closes #492
James Tucker [Sun, 13 Jan 2013 22:01:29 +0000 (14:01 -0800)]
Bump version
James Tucker [Sun, 13 Jan 2013 21:55:50 +0000 (13:55 -0800)]
Update README for release. Add security section.
James Tucker [Sun, 13 Jan 2013 21:33:08 +0000 (13:33 -0800)]
Squash warnings in spec_auth
James Tucker [Sun, 13 Jan 2013 21:10:20 +0000 (13:10 -0800)]
Reimplement auth scheme fix
* Add Rack::Auth.add_scheme to enable folks to fix anything that breaks
* Add common auth schemes, MS ones, AWS ones, etc are missing, as unlikely
* Checked Rails - they don't use our authorization code
* Checked Warden - uses rails
* Checked Omniauth - uses rails
* Checked doorkeeper - users rails
* Checked rack-authentication - does it's own thing
* Checked warden-oauth - doesn't do headers
* Checked devise - uses rails
* Checked oauth2-rack - header creation only
* Checked rack-oauth2-server - does it's own thing
* Probably missed a bunch, but that'll have to do
Carlos Antonio da Silva [Mon, 7 Jan 2013 21:12:42 +0000 (19:12 -0200)]
Remove warnings: 'not used variable' and 'shadowing outer variable'
James Tucker [Mon, 7 Jan 2013 18:46:43 +0000 (10:46 -0800)]
Add release announcements to README
James Tucker [Mon, 7 Jan 2013 18:41:16 +0000 (10:41 -0800)]
Bump to 1.4.3
Eric Wong [Wed, 22 Aug 2012 22:48:23 +0000 (22:48 +0000)]
multipart/parser: avoid unbounded #gets method
Malicious clients may send excessively long lines
to trigger out-of-memory errors in a Rack web server.
James Tucker [Mon, 7 Jan 2013 01:30:59 +0000 (17:30 -0800)]
Bump to 1.4.2
James Tucker [Sun, 6 Jan 2013 23:50:56 +0000 (15:50 -0800)]
Update README based on master
James Tucker [Sun, 13 May 2012 22:02:17 +0000 (15:02 -0700)]
Fix parsing performance for unquoted filenames
Special thanks to Paul Rogers & Eric Wong
Conflicts:
test/spec_multipart.rb
Sokolov Yura [Fri, 9 Nov 2012 09:58:56 +0000 (13:58 +0400)]
Fix parsing multiple ranges
Fix parsing miltiple ranges in HTTP_RANGE header according to w3 rfc2616 (according to last example in sec14.35.1 http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.35.1 ) (according to BNF rules in http://www.w3.org/Protocols/rfc2616/rfc2616-sec2.html#sec2.1 )
James Tucker [Sat, 29 Dec 2012 20:47:59 +0000 (15:47 -0500)]
.woff now has an official mime type!
* Closes #405
James Tucker [Sat, 29 Dec 2012 20:28:25 +0000 (15:28 -0500)]
Do not fail on cookies that are not URI escaped
* Closes #360
Conflicts:
test/spec_request.rb
James Tucker [Sat, 29 Dec 2012 18:04:46 +0000 (13:04 -0500)]
Add a note in KNOWN-ISSUES regarding ECMA escape
James Tucker [Fri, 28 Dec 2012 22:38:51 +0000 (17:38 -0500)]
Refactor spec_cascade and spec_head
* StringIO is a better choice than a struct here.
James Tucker [Fri, 28 Dec 2012 22:36:36 +0000 (17:36 -0500)]
Rack::Response now conforms to body.close SPEC
* Previously 204, 205 and 304 bodies were not closed correctly.
James Tucker [Fri, 28 Dec 2012 22:28:26 +0000 (17:28 -0500)]
Rack::Head now conforms to body.close SPEC
James Tucker [Fri, 28 Dec 2012 22:19:04 +0000 (17:19 -0500)]
Cascade now conforms to the body.close SPEC
James Tucker [Fri, 28 Dec 2012 22:02:06 +0000 (17:02 -0500)]
Clarify the body.close spec section
* This item is frequently missed, including in core.
* This is not a change in semantic requirement, and does not update the SPEC
version.
Konstantin Haase [Wed, 12 Dec 2012 10:37:24 +0000 (11:37 +0100)]
fixes for 1.8
James Tucker [Sat, 3 Nov 2012 15:35:04 +0000 (08:35 -0700)]
Ensure that deflater always closes bodies.
Closes #349
Zachary Scott [Fri, 2 Nov 2012 23:24:53 +0000 (19:24 -0400)]
Rack::BodyProxy#each, fixes rack/rack#434
James Tucker [Fri, 2 Nov 2012 20:54:39 +0000 (13:54 -0700)]
Prevent infinite recursions from Response#to_ary
Closes #419
James Tucker [Fri, 2 Nov 2012 20:03:59 +0000 (13:03 -0700)]
Return a bad request for malformed basic auth
Closes #438
Thomas Klemm [Fri, 2 Nov 2012 17:05:28 +0000 (18:05 +0100)]
Rack::Static: Rename methods
Thomas Klemm [Fri, 2 Nov 2012 16:57:01 +0000 (17:57 +0100)]
Remove .rbenv-version from .gitignore
Zachary Scott [Thu, 25 Oct 2012 03:59:49 +0000 (23:59 -0400)]
rescue Errno::ESRCH for windows, fixes #391
Zachary Scott [Thu, 25 Oct 2012 03:25:33 +0000 (23:25 -0400)]
CommonLogger Documentation, fixes #412
Christian Neukirchen [Mon, 22 Oct 2012 14:58:43 +0000 (16:58 +0200)]
Add redrawn logos by Zachary Scott
Date: Mon, 22 Oct 2012 10:29:22 -0400
Message-ID: <CAH6G9XNkVkUWh9JPT9HuHzJ4KmRNoPSr8ov3q0rgzH3b=u3cGw@mail.gmail.com>
On Mon, Oct 22, 2012 at 10:17 AM, Christian Neukirchen
<chneukirchen@gmail.com> wrote:
> Is it ok to put the other files as MIT license into contrib/?
> (Perhaps add a copyright message to the .svg)
Whatever you want, they're all yours.
Thanks!
Santiago Pastorino [Sun, 16 Sep 2012 17:48:09 +0000 (10:48 -0700)]
Make multipart parser work when there is no Content-Length header
Closes #418
Thomas Klemm [Sat, 15 Sep 2012 20:55:07 +0000 (22:55 +0200)]
Fix docs for Rack::Static http headers
Thomas Klemm [Sat, 15 Sep 2012 20:49:42 +0000 (22:49 +0200)]
Fix docs for Rack::Static http headers
Thomas Klemm [Sat, 15 Sep 2012 20:42:41 +0000 (22:42 +0200)]
Rack::Static: Provide custom HTTP header rules as array only and more concise docs
Thomas Klemm [Sat, 15 Sep 2012 19:24:58 +0000 (21:24 +0200)]
Rack::Static: Refactor array of arrays implementation
Sam Saffron [Wed, 5 Sep 2012 02:59:52 +0000 (12:59 +1000)]
- correct existing raise test
- amend it so rack lock releases the mutex on throws as well and raises
- added raise test
Thomas Klemm [Sat, 1 Sep 2012 00:03:49 +0000 (02:03 +0200)]
Only run header rules tests that use a hash on ruby 1.9 and above
Thomas Klemm [Fri, 31 Aug 2012 23:42:54 +0000 (01:42 +0200)]
Allow HTTP header rules to be provided in array of arrays (useful on Ruby 1.8.7)
Thomas Klemm [Fri, 31 Aug 2012 20:46:12 +0000 (22:46 +0200)]
Try to get headers working on Ruby 1.8.7
Thomas Klemm [Fri, 31 Aug 2012 19:24:55 +0000 (21:24 +0200)]
Allow Rack::Static to assign HTTP Headers based on rules
James Tucker [Sun, 26 Aug 2012 19:15:05 +0000 (12:15 -0700)]
Fix script name escaping in Rack::Directory
Closes #415 and replaces it, which came with no tests and an insecure
implementation.
James Tucker [Sun, 26 Aug 2012 18:50:49 +0000 (11:50 -0700)]
Deprecation notice for cache_control parameter
Thomas Klemm [Fri, 17 Aug 2012 13:09:52 +0000 (15:09 +0200)]
Adjust Documentation
Thomas Klemm [Fri, 17 Aug 2012 10:31:11 +0000 (12:31 +0200)]
Allow Rack::File and Rack::Static to serve custom HTTP Headers
Timothy Elliott [Sat, 24 Mar 2012 00:22:22 +0000 (17:22 -0700)]
Use aref instead of merge!
Jamie Macey [Thu, 19 Jul 2012 19:46:50 +0000 (12:46 -0700)]
load session data for merge!
ID#prepare_session calls merge! on the newly-created SessionHash, but
this method is not overridden to parse existing data. As such, any
previous session data passed in from an earlier middleware is discarded.
For me, this was breaking Rack::Test while testing a Sinatra app.
Stephen Celis [Tue, 17 Jul 2012 19:48:21 +0000 (12:48 -0700)]
Don't set blank Cache-Control header in Rack::ETag
A Cache-Control header with an empty string is meaningless (confusing,
even, to those inspecting response headers) and slightly wasteful.
Signed-off-by: Stephen Celis <stephen@stephencelis.com>
Lars Gierth [Mon, 19 Mar 2012 17:04:04 +0000 (18:04 +0100)]
Explicitly declare Enumerator
Lars Gierth [Mon, 19 Mar 2012 15:10:24 +0000 (16:10 +0100)]
Update Rack::Deflater specs
- Always wrap apps in Rack::Lint
- Never assume response body is an Array (!)
Conflicts:
test/spec_deflater.rb
Lars Gierth [Mon, 9 Jan 2012 01:19:41 +0000 (02:19 +0100)]
Use const_defined? instead of defined? to check for Enumerator
Lars Gierth [Wed, 28 Dec 2011 22:49:28 +0000 (23:49 +0100)]
Don't use Object#tap, 1.8.6 doesn't have it
Lars Gierth [Wed, 28 Dec 2011 22:33:51 +0000 (23:33 +0100)]
Squash warnings caused by test code
Lars Gierth [Wed, 28 Dec 2011 22:24:16 +0000 (23:24 +0100)]
Fix declaration of Enumerator for 1.8.7
Lars Gierth [Wed, 28 Dec 2011 22:13:15 +0000 (23:13 +0100)]
Wrap test apps in Rack::Lint and fix uncovered errors (seventh pass)
Update tests for ShowExceptions, ShowStatus, Static and URLMap.
Lars Gierth [Wed, 28 Dec 2011 22:05:06 +0000 (23:05 +0100)]
Wrap test apps in Rack::Lint and fix uncovered errors (sixth pass)
Update tests for Sendfile, Session::Cookie, Session::Memcache and Session::Pool.
Conflicts:
test/spec_session_cookie.rb
Lars Gierth [Wed, 28 Dec 2011 21:32:46 +0000 (22:32 +0100)]
Wrap test apps in Rack::Lint and fix uncovered errors (fifth pass)
Update tests for MockRequest, MockResponse, NullLogger, Recursive and Runtime.
Lars Gierth [Wed, 28 Dec 2011 20:44:45 +0000 (21:44 +0100)]
Wrap test apps in Rack::Lint and fix uncovered errors (fourth pass)
Update tests for Lobster, Lock, Logger and MethodOverride.
Lars Gierth [Wed, 28 Dec 2011 00:59:40 +0000 (01:59 +0100)]
Wrap test apps in Rack::Lint and fix uncovered errors (third pass)
Update tests for Directory, ETag, File and Head.
Lars Gierth [Wed, 28 Dec 2011 00:59:13 +0000 (01:59 +0100)]
Wrap test apps in Rack::Lint and fix uncovered errors (second pass)
Update tests for Config, ContentLength, ContentType and Deflater.
Fabio Kreusch [Thu, 23 Feb 2012 14:42:37 +0000 (12:42 -0200)]
Skips Etag even if Cache-Control has a no-cache value together with other options
Bruno Aguirre [Tue, 31 Jan 2012 13:14:31 +0000 (11:14 -0200)]
Rack::Static :index can handle multiple folders
* Tries to serve the defined :index in every folder
* Useful for documentation like nanoc.
Konstantin Haase [Mon, 30 Jul 2012 13:36:54 +0000 (16:36 +0300)]
Update years in license
Jamie Macey [Thu, 19 Jul 2012 19:46:50 +0000 (12:46 -0700)]
load session data for merge!
ID#prepare_session calls merge! on the newly-created SessionHash, but
this method is not overridden to parse existing data. As such, any
previous session data passed in from an earlier middleware is discarded.
For me, this was breaking Rack::Test while testing a Sinatra app.
ITO Nobuaki [Fri, 13 Jul 2012 06:51:50 +0000 (15:51 +0900)]
Use backport of URI module even on ruby 1.9.2-p320.
ITO Nobuaki [Fri, 13 Jul 2012 06:03:17 +0000 (15:03 +0900)]
Added missing dependence on 'timeout' in test/spec_utils.
ITO Nobuaki [Thu, 12 Jul 2012 02:52:57 +0000 (11:52 +0900)]
Fix request loop on non-stale nonce with time_limit parameter.
Conrad Irwin [Mon, 4 Jun 2012 07:28:42 +0000 (00:28 -0700)]
Set __LINE__ correctly for rackup files.
Before this change the line numbers were off by one, which broke
debugging tools like Pry in addition to causing a smidgen of user
confusion.
Reported-At: https://github.com/pry/pry/issues/571
Hrvoje Šimić [Tue, 22 May 2012 15:00:02 +0000 (17:00 +0200)]
update the dead link
remove the dead link
use internet archive for a dead link
added working link
James Tucker [Sun, 13 May 2012 17:51:45 +0000 (10:51 -0700)]
Improve pidfile reporting and test coverage
* Output reduced to a single line
* Integration test added that also suppresses and checks output
Jean Boussier [Sat, 5 May 2012 16:05:32 +0000 (12:05 -0400)]
Check if the PID in pidfile is still running #371
Then abort or remove the pidfile
Jean Boussier [Wed, 21 Mar 2012 10:31:02 +0000 (11:31 +0100)]
Check that the pidfile is not already present before overriding it
James Tucker [Sun, 13 May 2012 00:56:44 +0000 (17:56 -0700)]
Squash argument warning on 1.8
Konstantin Haase [Sat, 12 May 2012 19:21:18 +0000 (21:21 +0200)]
point to github page instead of rubyforge
Anurag Priyam [Thu, 26 Apr 2012 09:29:59 +0000 (14:59 +0530)]
rackup: include the value of each -I command line option in $LOAD_PATH
`rackup -h` says '-I' can be used more than once. However, instead of adding
each value to $LOAD_PATH, rackup would discard the value of all but the last
'-I' option.
Signed-off-by: Anurag Priyam <anurag08priyam@gmail.com>
Josef Sin [Thu, 12 Apr 2012 13:28:45 +0000 (15:28 +0200)]
Skipping empty params inside query what lead to parsing error.
Example of these cookies would be: "foo=bar,;bar=foo" or ",foo=bar;,"
Konstantin Haase [Sun, 11 Mar 2012 11:56:56 +0000 (12:56 +0100)]
don't load broken backports on jruby, either
Konstantin Haase [Sun, 11 Mar 2012 10:40:36 +0000 (11:40 +0100)]
fix version comparison, again
Konstantin Haase [Sun, 11 Mar 2012 10:21:28 +0000 (11:21 +0100)]
fix version comparison
Konstantin Haase [Sun, 11 Mar 2012 10:03:34 +0000 (11:03 +0100)]
regexp issue has been fixed in Ruby 1.9.2-p381
Ben Hamill [Wed, 7 Mar 2012 03:20:32 +0000 (21:20 -0600)]
Use `const_defined?` so that it only checks withing the scope of URI.
Stefano Cobianchi [Mon, 5 Mar 2012 22:17:53 +0000 (23:17 +0100)]
Don't load the backport for uri/common.rb when running 1.9.3-p125 and beyond
Ben Hamill [Sun, 4 Mar 2012 03:36:32 +0000 (21:36 -0600)]
Uh. Put the conditional on the CORRECT line. :sweat:
Ben Hamill [Sun, 4 Mar 2012 03:25:34 +0000 (21:25 -0600)]
Only remove a constant if it exists.
David Lee [Wed, 22 Feb 2012 16:00:37 +0000 (08:00 -0800)]
Add #patch to MockRequest
Aaron Patterson [Wed, 25 Jan 2012 19:25:17 +0000 (11:25 -0800)]
Backported fixes for:
http://bugs.ruby-lang.org/issues/5925
Relevant commit:
https://github.com/ruby/ruby/commit/
edb7cdf1eabaff78dfa5ffedfbc2e91b29fa9ca1
Jonathan Tron [Tue, 24 Jan 2012 12:39:24 +0000 (13:39 +0100)]
Prevent error in Utils.parse_query/KeySpaceConstrainedParams when a key is nil
Before this when parsing a query string or a cookie string, an error were
raised because of the `key.size` in `KeySpaceConstrainedParams#[]=`.
This caused an error when params contained something parsed as a `nil` key.
Kyle Drake [Mon, 23 Jan 2012 08:19:33 +0000 (00:19 -0800)]
Trivial readme fixes to 1.4.1 release history
James Tucker [Mon, 19 Mar 2012 01:36:31 +0000 (18:36 -0700)]
Add warning to strongly recommend to people to have secrets protecting their cookies
Conflicts:
lib/rack/session/cookie.rb
test/spec_session_pool.rb
James Tucker [Mon, 19 Mar 2012 01:08:51 +0000 (18:08 -0700)]
Documentation and Rakefile updates! Yay!
Conflicts:
Rakefile
James Tucker [Mon, 23 Jan 2012 06:48:35 +0000 (22:48 -0800)]
Multipart percentage fail, round 3, the final character. Fixes strings terminated with %. See #323. Revisit for 1.5.
James Tucker [Mon, 23 Jan 2012 06:20:56 +0000 (22:20 -0800)]
Prep for 1.4.1
James Tucker [Mon, 23 Jan 2012 06:01:50 +0000 (22:01 -0800)]
Merge remote-tracking branch 'thedarkone/nested_params_key_space2'
* thedarkone/nested_params_key_space2:
Rack::Utils#normalize_params should be ignorant of the provided params class.
Correctly count the key space size for nested param queries.
James Tucker [Mon, 23 Jan 2012 01:42:34 +0000 (17:42 -0800)]
Correct multipart parser skips for cases where we have data that looks partially like it's percent-hex encoded