wmelton / knproxy Goto Github PK

-- New Release (Ver 4.50) --
-Rewritten code in parser. Fixes problems in HTML parsing especially in dealing with javascript.
	New code should break less JSes and *hopefully* make more ajax requests actually work.
	Scores 95 on Acid3 now.
-Fixed bug in HTTP Request Module with websockets mode. [Note: Websockets mode still does not support POST forms!]
-Fixed bug in tag parsing (does not eat delimiters)

-- Feature Upgrade (Ver 4.40) --
-Added support for incremental downloads nd HTTP Range: Header
-Processes ETags and other cache headers to let client more effeciently handle cache
-i18n now works for the cookie manager
-Supports a special header X-KnProxy-Location: [Base64 Encoded absolute URL]
 Allows proxied server to serve up different pages if viewed by KnProxy and possibly to prevent looping
-Fixed small error if URL field is empty (frames will embed)
-Changed the Debugging page to provide more information.

-- Feature Upgrade (Ver 4.40-beta) --
-HTTP Request Module rewritten for better parsing of headers
-Implemented HTTP sockets (very buggy) for compatability
 HTTP Sockets is only a fallback for cURL if it is not available
 HTTP Sockets has buggy support for chunked and normal encodings and get queries but limited url escaping
 *IT DOES NOT SUPPORT POST REQUESTS OR HTTPS YET*
-Added a pseudo 'STREAM MODE' which allows incremental download of data in stream-like manner for those extra-large files
-Renamed many conf param names
-Implemented progressive download if remote server supports it and forced progressive download if it doesnt
-Fixed some errors in cookie manager

-- Feature Upgrade (Ver 4.36) --
-Added 'Cookie Manager' accessable thru the url 'about:cookies'
-Added 'System Internals' accessable thru url 'about:sysinternals' for debug
 These can be closed by defining DISABLE_ABOUT_PAGES as true.
-Fixed incorrect parsing of http header/body delimiters
-Fixed problem with multiple headers on servers that support them (solves a few login bugs)
-Now respects cookie expire times, cookies expire as defined or else are treated as session cookies.
-Removed some unused/redundant code
-Changed to pseudo-referer mode as default

-- Bugfix (Ver 4.34) --
-Fixes typo in includes/module_http.php:143 heaer
-Fixes undefined index error index.php:28 for HTTPS detection
-Fixes bug #3419799 in SF tracker, escapes ampersands in url possibly with them (still testing)
-"about:" pages introduced

-- Bugfix (Ver 4.33) --
-Fixed serious bug on version 4.32 encryption module accidentally sending unencrypted url in GET
 All users on 4.32, if highly security conscious, are STRONGLY RECOMENDED to upgrade or disable the navbar!
-Changed the random number key generator to use a higher region
-Fixed bug in Url Parser processing empty path strings (link to file on root of host)

-- New Release (Ver 4.32) --
- Added Navigation Bar Feature (default on) to show addresses (Requires support for frames)
- Improved basic encryption
- Fixed small problem in Url Parser with url fragments
- Fixed problems JS url parsing, now only parses url inside of strings.
- Allows users to set 'Blank Referer'
- Users can opt to shut off GZIP retreiving (many free servers do not support gzip decoding) and force remote servers to return unzipped
- Tested with Acid3 obtaining score of 92/100
- Internationalization Supported!

-- New Release (Ver 4.30) --
- Rewritten the URL parser, solving parsing problems for many URLs
- Fixed POST/GET determination method. Solves problem of mistaking post for GET forms
- Fixed problems with addresses embedded in JS
- Fixed problems with REGEX exceeding pcre recursive/backtrack limits and returning blank pages

-- New Release (Ver 4.28) --
-Fixed BUG #3374265, #3379261
 Problem with file encoding and Problem of jumping out of ssl
-Updated the main interface to prevent re-submitting the form causing address to be double-encrypted
-Fixed bug in error message to filter HTML injections
-Fixed problem parsing some GET forms
-Respects 'refresh:' http headers now

-- Bugfix --
-Fixed newline encoding in includes/plugins/youtube.php to UNIX because it caused problems

-- New Release (Ver 4.26) --
-Fixed bug with parser that changed almost all POST forms into GET forms
-Added native support for KnProxy V5

-- New Release (Ver 4.25) --
-Updated URL Parser removing problems generated in last bugfix
-Updated parser to ignore <! beginning tags (namely <!-- and <![CDATA[)
-Can now parse @import with "" '' wrapping strings
-New default salt

-- Bugfix (Ver 4.24 BUGFIX) --
-Fixed serious problem with last release caused by debugging code not removed
 Sorry about that.

-- New Release (Ver 4.24) --
-Now officially supports GET forms
 Which allows 'Google Search' To WORK :D
-Pushed new default SECRET into effect. Please omit config.php if you still wish to keep your urls.
-Support for You*Tube videos is in place though you cannot choose def rates [PACK:SSE-2011-07-10-PV2] Sites: You*tube
-Rolled back url parser as problems occurred in new version.

-- New Release (Ver 4.22) --
-Resolved Bug Artifacts #3358326 #3333079 on SourceForge Tracker
-Updated Parser to only process urls inside of tags
-Updated URL Parsing module resolving errors while processing urls like
 ?somewhere=http://someaddress/somepath/?something or /somepath/http://somesite/somepath2/

-- New Release (Ver 4.19) --
-Solved Warning Messages when php is in strict mode. 
 (Thanks to bug report from zhongxiaolee AT gmail DOT com)
-KNEN_SECRET allows you to define a salt for address encryption
 ( Graduated from -dev branch!)
-Supports GZIP on HTML/CSS/JS and other MIME:text/* files! Can be disabled in conf.php
-Autodetect page encoding (to be used in ver 4.2x for Page Encryption)
-Changed default interface to be slimmer (smaller download size)

-- Bugfix (Ver 4.16)--
-Fixed bug when parsing url( in css tags when there is a space after the bracket.
 Solves problem with some @import url( somewhere ); situations

-- New Release (Ver 4.15) --
-Added support for HTTP basic authentication (username and password stored in cookies)
-Enhanced parsing engine for JS urls (better support for face*book and google ajax calls!)
 Tested on Facebook, can now post status/share and all kinds of messages!
 This works for general websites as well!
-Site Specific Enhancements [PACK:SSE-2011-06-11-PV1] Sites: Facebook
-Lets 404 error pages pass through instead of getting handled! (because some error pages are creative ^_^)

-- Development Release (Ver 4.10-dev) --
-Note: The DEV branch is for small features that are seasonal and used to defeat GFW
 They do not provide a big improvement if you are using outside of China or networks
 that have intensive censorship.
 DEV branches will have a chinese versionlog file to guide users. It will not be included 
 in the main builds.
 Some features may eventually graduate from the dev into the main builds, others will only
 remain in dev.
-Added support for cloud proxying (still testing)
-Added config file. You can set your own proxy key to avoid url being caught by GFW.
-Added Chinese Versionlog (note: only dev versions have the chinese Versionlog)

-- New Release (Ver 4.10) --
-Edited the parser to support parsing of style= inside an HTML tag
 Note: This fixes the problem of the Google logo not showing up
-Added a small filter for urls captured in JS to remove \" and \/ stuff.
-Fixed small problem in page encryption, but still doesn't decode GB2312 correctly


-- New Release (Ver 4.08) --
-Added page encryption support(disabled by default)
 The GFW experienced an upgrade in blocking techniques.
 This new version solves problem with page encryption
 Small enhancement for users in China

 
-- New Release (Ver 4.07) --
-Fixed bug in knHTTP module. Only one request is sent to get headers and body.
-Fixed bug in parser for mime types
-Fixed regression problems in SSL enable/disable checks
-Added a JS for future use of enhanced client-side parsing

-- New Release (Ver 4.05) --
-Limited SSL Support included with a warning dialog
-Solved problem parsing cookies with Expire dates
-Solved problem of wrongly parsing Location: tags when X-Something-Location: was present
-Minor Enhancement for parser to find the mime type more accurately

-- Bugfix (Ver 4.01) --
-Last upload was not final v4 version!
-Fixed bug in parsing engine that didnt change the 'action=' field in html forms resulting in unproxified forms
 Now can work with f*cebook logins and all sorts of logins!
-Fixed bug in URL parser. Now ignores javascript tags, data url scheme and anchors (eg 'javascript:','data:','#anchor').
-Added debug mode, if you get a blank white screen on loading a page, try adding &debug=true to the end of the url
 you will enter debug mode. You can send the debug data to me at [email protected] or (if you can) solve it by yourself.

-- Initial release (Ver 4.00) --
Knproxy Theta (v4) is out with loads of new improvements!!
=======
1. Slimmer Engine
The method used for fetching and parsing has been greatly slimmed down. Knproxy now runs on half the processing resources!

2. Improved Parsing
New specialized parsing for css/js and html with checks on MIME type to give the best optimized effect.
 The regexes have also been fixed meaning a compatibility for more websites. Tags with href=" href=' and href= are all supported in this version.
 CSS parsing improvements prevent the page suffering from bad css.
 JS scripts are basically ignored to speed up the loading.
 Also, images and other attachments need not go through the parsing engine to prevent breaking (if a good content type is sent).

3. Better Support for POST forms and COOKIEs
Support for POST and COOKIES is now built into the knHTTP request engine. (And needs no special plugin as in v3)
COOKIES can now be stored locally instead of on server and are loaded faster.
GET requests are on the way!

4. Improved Error Checking
Error checks are built in natively. Tons of debugging info given on error to help pinpoint problem.

5. Native Encoding
Now, encoding for URL is embedded natively and not plugged in. However, we currently only support KNEN.
(You can implement your own obfuscation method!)

6. Speedup
Lots of redundant code removed. Proxy modualized and functions separately. Greatly reduced the size of master file!

7. Supports Shifting Code
You can give a shift value for the encoding to avoid blacklists. Though currently addresses are static.

Enjoy~
Author Jabbany