MyHeadlines Change Log

v5.0.0:

Release Date: TBD

• TODO: AJAX update with RESTfull interface (can you AJAX a div, and not traverse the a "a href" link? I'm gonna find out)
• TODO: Drag and drop site layout
• TODO: minimize / maximize / remove site elements.
• TODO: are cookies accessible within the AJAX fetch?
• TODO: remove cruft: link dump, promote, stock ticker, google search.
• TODO: Stand Alone support ONLY. No more PHPNukes, or it's clones.
• TODO: Proper XML parsing of RSS and ATOM. No more regex.

v4.3.7:

• Improved PHP5 support. (5.2.5 has done away with some legacy global server vars) See ticket 4816651.
• New Stock Ticker scraper definition integrated in the reset function, and the install script.

v4.3.6:

Not a public release

• Added some javascript to notify user that they can't comment unless they are signed in. (I'm seeing comment spam on my site, this should help reduce it, won't help everyone though...)
• Fixed Blog Search!
• Minor bug fixing

v4.3.5:

• Minor bug fixing

v4.3.4:

• Minor improvements in templates for link dump and blog
• More XSS scrubbing. Should have all of them now.

v4.3.3:

• More performance tweaks
• Comments are now scrubbed to remove undesirable tags and javascripts

v4.3.2:

• Many performance improvements
• Some PHP warnings, notices, and infos removed
• Large reduction in DB transactions: When I started a single page load could make over 400 DB queries! Now we are about 100. 75% reduction in DB traffic!
• SHM Module created. A fast cache object with assumed lossy data retention. Shared memory is available on most linux systems, and this should make it fast!
• XSS hole has been plugged. Thanks Thomas Pollet for the find, and the prompt notice.

v4.3.1:

• Link Dump: Admin capable of deleting an entry.
• Strip Tags removed from the backend RSS service: Comics Page as an example will now send image in RSS feed.
• Blog publish flag hides search, archive, URL access, backend, etc. if unchecked in DB
• Blog Admin enhancements for search (shows hidden and adds Edit option)
• Blog Admin enhancements for archive (shows hidden)
• Promote rejects posts without a valid Cache ID.
• Digg, del.icio.us, Furl Support promote system
• rewrite: Promote System is now more easily extensible
• BugFix: Promote Defualt now works
• Removed all PNG graphics
• Atom support (remix) - ATOM is a filthy swamp, but at least it is valid XML.
• OPML!!! Import and Export.
• SEO for Blog Archive. Drop the buttons and go REST to make things easier for the spiders.
• RSS output for Blog and Link Dump now detected by Firefox AND MS IE 7.0. It has a working XSL too.
• Blog Search works again
• backend URL construction fixed
• Fixed the FULLTEXT support for Cache tables bug on newer (>4.x) MySQL DBs. Thanks to Mark DeWar for the support and troubleshooting this issue with me.
• Links not using frames: you have the option to redirect outside the site.
• Backend error reporting not using static XML files any more. Now generating the RSS feed programatically.
• register_globals - I override this "security feature" MyHeadlines works on or off and of course this breaks the "security" feature.
• Entry Page Selector - (new type of Configuration type too).

v4.3.0: Service Pack B

• Code clean up of the SOAP class (same as the xml-rpc code clean up)
• Fixed up the PHP5.0.5 Issues. PHP5 IS SUPPORTED!
• Fixed up the GTU MySQL bug that was introduced in 4.3.0.

v4.3.0:

• blog preview now uses your blog templates. Preview actually shows you what you're gonna get on the blog page.
• tunable "show_comment_count" instead of "More" on link dump and blog.
• Skins Are Back!!! MyHeadlines has been looking kinda dated. So I have added some skins, and will play with CSS to see if I can make it cool again.
• A little fixup for a MySQL bug when using 4.1.x or newer. (Thanks Patrick VanVickle!)
• Many performance improvements on the DB side of things.
• Improved Stock Ticker: templatable, and faster page loading times.
• fixed a template bug so Google AdSense can be added to your page template.

v4.2.9:

(Experimental)

• Fixed up some minor bugs: Link dump promote..., removed stockticker
• Added {blog_rss} to the global template.
• Added global variable $myh_meta_title (useful for search engine placement optimizations)
• Hidden menu items now remain hidden (unless you edit the language setting)
• Comment system for both Link Dump and Blog entries
• Updated the Cache function, it should not re-cache old items anymore...
• Blog System
• Link Dump System
• Promote system updated to allow configuration and to include Blog and Link Dump Options
• Added Chunked support for HTTP 1.1
• Link system (framing a URL within MYH). Handy if you want to extend MyHeadlines with other content, but have it appear native to the MyHeadlines object. Thinking about adding the url for the Admin Login....

v4.2.8:

• Upgrades to the search mechanism. Now can sort by date or score, and there's an search entry form at the bottom of the results page if you want to adjust your query. (my very first radio button control, hard to believe)
• Backend Subscribers now trigger Cache Control activity
• Added a new template tag: {bullet} this replaces the statically hardcoded value for the news paper icon as the bullet for each item. This tag is tied to the new bullet_enabled parameter. If turned on the bullet switched from the paper icon to the "favicon.ico" file from the source! Kinda neat, but still experimental, don't know how this works in lesser browsers... Use at your own risk. Safest setting is 0. (also has no impact if you don't add the tag to your template definition)
• No More RCH... Back to the donation model.
• Other Bug fixing... (Top Ten, ATOM parser, etc.)

v4.2.7:

• ATOM parser added.
• Backend Logging and Administration has been upgraded. You can now Ban users, or rate limit their access to the backend services.
• Fixed rename in category tree. Now adjusts the weight for the sort.
• New import files. removed old import files.
• Fixed multiple menu entries due to hide/show conflict with change a config parameter or install (Bug Track:866924)
• More bugs fixed with the Clean Source Database. Now you can recover lost sources or just delete them...
• Untouch (expire) a source after the admin edits the source definition.
• Removed "RSS_SHOW_SUMMARY" option - this is confusing since the template system now offers that functionality.

v4.2.6:

• Fixed the installation issue with v4.2.5. If anything goes wrong please verify the Administration->Configure->Parameters settings. Pay particular attention to "webroot" if it is wrong then others may be too... (autodetect is a bit flaky)

v4.2.5:

• Database maintenance functions have been added. This should help with the garbabge collection. All sorts of junk accumulates over time, this will clean the DB. If that sort of thing matters to you.
• I have hacked a prototype Menu System with bare bones functionality. You can now show/hide the user lever menu items. this will allow you to remove features from plain view and will serve as a test for how to handle adding new menu items. This may be completely ripped out in future releases, but for now it is a "good enough" solution.
• Track ID: 833928 This bug Sqaushed with GUSTO!!! Now a configurable parameter "find_a_source_limit" (Administration->Configure->Parameters) to set the MAX number of search results. Default to 50, but has been working with 250... whatever your preference. NOTE: I also IMPROVED this feature by having the result come back sorted by the last update timestamp. This means that the results are far more relevant and are more likely to be for good sources, since they have recently been updated.
• Track ID: 833935 New Template definition {image_url}!!! If you want to see (not see) a transparent 1x1.gif for sites that dont have an image, turn off show_shameless_plug, and add either {image} or {image_url} to dispBlockNoImage.
• Bug fix for sites using older versions of PHP: http.php line 109
• Enclosures RSS 2.0 - Now supported by the scraper template system Use {enclosure_0}, {enclosure_1}, etc.
• Major Backend overhaul!!!! The following subsytems now have RSS output:
  • RSS Sections
  • RSS Top Ten.
  • RSS Search.
  • RSS Latest News.

v4.2.4:

• RCH Error messages are a little more explicit as to the actual failure. Not much, but it may help.
• Removed date helper functions from xmlrpc.php. These were not being used and were poluting the RCH namespace.
• More work on the HTTP1.1 machine:
  • Check out the register_shutdown_function() in PHP! Learn something new every day.
  • Added a TimedOut state to the statemachine. Same as Disconnected...
  • Trashed a couple For Loops that made very little sense, this code was just too damn clever to be useful. I am making it much more explicit! No more fucking around in someone's ideal world... The Net is a dirty environment and it never works according to spec, so now this code is somewhat more hardened.
  • Added "Connection: close" to outbound HTTP1.1 header
  • Added "User-Agent:" to outbound HTTP1.1 header. It was a configurable parameter that I broke with the HTTP1.1 implementation.
  • Max download size of 200KB is hard coded into the app
  • New defaults: RSS_TIMEOUT = 50 (seconds)
  • Bad Sources (410 and 403) are now insta-blocked... Other 4?? results go into the (3?) strikes queue.
• Added a critical flush() to the backend cache control... (should have done this long time ago, I saw this bug many moons past)
• New Help Links now point back to my site and the Administrators guide.

v4.2.3:

• Skins for those who don't want to pay for template access, but insist that my code is too "Garbaged Up" for their site.
• Possible bug squashed: Sections code.
• Code Clean-Up: I have learned a valuable lesson - Just because everyone uses a particular library (xml-rpc) doesn't mean that the code is clean. I have fixed numerous bugs in the xml-rpc library, and renamed the class to avoid it conflicting with other modules.
• Administration Category tree now shows the number of subscribers to any given source. Thanks Remco de Lang
• HTTP1.1 error codes are now handled gracefully, redirects are followed, sources are updated.
• Stock Ticker scraper updated.

v4.2.2:

• SECURITY FIX: All 4.x.x versions prior to this release are vulnerable to a cross site scripting attack!!! This hole is now plugged.

v4.2.1:

• Bug Fixing, Bug Fixing, Bug Fixing. Just goes to show you that you should never trust a ".0" release.

v4.2.0:

• New import files, and removal of dead ones... (magportal, rssengine). The old NewsIsFree import file is now deprecated. Remove the newsisfree.imp and then re-import the new file: newsisfree_2.imp
• Template Layout System, and Sections are documented.
• Added new Parameter: rss_User-Agent_string - This is the string that identifies what software is making the request to the far end server. Polite usage is to leave this as "MyHeadlines" so that remote admins may monitor traffic patterns and that they have the opportunity to block usage on their servers. Hiding behind fake strings is not cool, but there could be a technical reason for changing this so it is there.
• Fixed the backend style sheet. (removed it)
• Upgraded Maintenance Page (delete without preview)
• RSD - Really Simple Discovery: Now auto detecting your blogger API!!! Thanks to Dennis Slagers for testing, and bringing this feature to my attention.
• Fixed a small bug in the eport function. Now only good sources are exported, this will not export known to be bad sources.
• The pruneTree function now checks for source name, if it is null the source is pruned from the tree.

v4.1.4:

NOT A PUBLIC RELEASE

• Sections code added to RCH.
• New DEFAULT Layout!!! (Stock Ticker under the menu, and more...)
• XML_RPC to SYNDIC8: "Find A Feed"
• XML_RPC for Blogger API (and email) is now working. Thanks to Mike Krus' HPE code, pretty much a complete ripoff ;-) Slight architectural changes to fit MyHeadlines, but otherwise it's HPE code. Mr. Krus' code is quite clean, and raises the bar. Something for me to aim for.
• NEW TECHNOLOGY: RCH SubSystem - This provides code security and access to potential revenue stream. I rejected embedding ads in the MyHeadlines module as a source of funding. Instead I have protected some features via RCH and implemented a micro-payment system on a per feature basis. You don't have to buy all of the new features, just the ones you want.
• NEW FEATURE: Template based layout system. End of the much maligned Table borders!
• NEW FEATURE: Sections - just like my Static News module, now this functionality is contained within the MyHeadlines Module. Static separate news pages with default subscription now allow the site admin to construct "Sections" just like you would in a print news paper.
• DOCUMENT LOCATION - Everything points to my site now. The docs were making the download very heavy. Now I host all the docs, keep them current, and bring the user to my site :-) This should shave 2 MB off of the download.
• GTU UPGRADES:
  • GTU rendering has been dramatically improved!!! Standard plus/minus metaphore has been implemented.
  • code refactoring, optimizations via tracking root ids
  • gravity functionality added - Tree SORTING!!!

v4.1.3 Service Pack:

• template based scraper extension -->> Story Images!!! {pic_0}, {pic_1}, etc.
• changed a directory name "core" becomes "apps". A bright user noted that some lazy admins have bad scripts that delete anything named core.
• Anti-Anti-Scraper functions...
• memory conservation for import/export functions: Looks like on low memory servers the import files end up in the root of the tree instead of the intended category I have tried a little work around to ease the recursive nature of the import function. Works reasonably well, I will double the margin and it should work nearly every where...
• BUG FIX: Google search had a commented line over 200 chars long. Looks like a "//" comment has a fixed limit and this line exceeded it. PHP4.3.0? Anyway, I have simply removed the line. All is fine.
• BRAIN CRAMP: I had two "setTimeout()" functions within the agar_stck class, I can't believe it has worked this long. PHP4.3.0 finally bitched appropriately. I've fixed it.

v4.1.3:

• Fixed the "allow_anonymous" behaviour (Thanks: R.M. de Lang for this one!)
• Fixed the backend RSS for scraped feeds, now it will service the request BEFORE attempting to update the cache. External news readers will not see a delay in the feed.
• MyHeadlines Administrators Guide!!!

v4.1.2:

• Vladan came through again with a big one: Usability for failed fetches while user is subscribing. It now posts a message stating that the feed is offline, and marks the source as bad. The source will appear on the Maintenance page for admin to review.
• Export now has a scraped feeds only option. (used to create the new import file...)
• Fixed the import function for non-scraped feeds (thanks Vladan for this bug report)
• Now fully RSS 0.91 Valid as the default RSS output. (also passes rss1.0, and rss0.9)
• goofed on the install script for 4.1.1

v4.1.1:

• RSS output for the scraped feeds is now active (but does not pass RSS Validator) This feature needs more work but it should be good enough to get started. MyHeadlines successfully reads this output, but other news readers may not.
• BUG FIX: rss::fixURL now handles paths from the original URL
• New Feature: The Stock Ticker component output now defaults to a javascript fader. This will be used to stop theme break-up. Read the API.txt doc for more info on getDashBoardComponent_StockTicker($javascript=1)
• Improved bad feed detection algorithm (scraper requires more stringent error checking)
• BUG FIX: include/src.php: $cmd .= " scraped bool NOT NULL DEFUALT 0,"; THANKS: Vladan Zirojevic

v4.1.0:

• Brand Spankin' New Scraper Engine! Template Based Headline Scraping!!! Don't have an RSS/RDF/XML feed for a site? No Problem, build a template instead!!! >>> Read: headline_scraper.htm file for complete instructions.
• Fixed the Source table: No More BLOBs!
• Expanded the Source table: Now there's a flag for scraped feeds, and a scraper template column.
• Fixed the Top Ten page, ' and other now converted back to proper characters.
• Apperance textarea now uses stripslashes() to clean up \\\" type garbage
• Un-Import feature added. This will undo an import action. Useful for times where you no longer qualify to receive feeds from an aggregator: Moreover recently changed their terms of use, and as such we needed a way to remove the Moreover imported feeds.
• RSS Backend for serving up the scrapped feed to others. This will actually serve all feeds in the DB, but hopefully people will be nice enough to go direct to the source...
• Garbage Collection for anonymous user data: There's a new button on the maintenance screen. This will delete all anonymous user data from the DB which grows considerably over time and offers little value to the system.
• It's Back!!! Allow Anonymous configuration parameter is back. I had a couple requests for this feature. I still think that the CMS should handle this functionality but in some cases it is simply not offered at the CMS layer. So as a user driven kind of guy, I have bent to the will of the people. It is set to allow access by default but you can change it easily through the administration menu.
• BUG FIX: If you chnaged the description of a source, your changes were not saved... now, it works as designed.
• ENHANCEMENT: The Maintenance screen now works better with submissions that have null value for the title of the feed.
• New Configuration Variable: "max_user_time" this limits the time spent trying to update expired items if this is a user interface. cron and backend functions will not be affected, but if there is a user on the end of the line then this will clamp the time spent looking for new data, and better your site's load time.

v4.0.8:

• minor hacks, nothing worth noting here...

v4.0.7:

• PHPNuke Version ONLY -> I now honour the $prefix variable for DB prefix. Sorry about that one, it slipped through and has been around for a while. You will lose everything in this upgrade, all subscriptions, and sources will need to be re-configured.
• PHPNuke Version ONLY -> To compensate the PHPNuke community for my bad behaviour, I have included in this release 2 special surprises:
  Stock Ticker Block - WooHoo!!!
  Subscriptions Block - This is meant to be placed as a centre block to make it easy for you to place the user subscriptions on the front page of your site.
• Redo of the "scrub/unscrub" functions that caused the "\'" proliferation.
• Repaired the "Cache Feature Improved" from 4.0.6 that wasn't...
• Implicit "Move" feature added to the edit portion of the Admin interface for headline sources: if you change the Category of a feed in the admin interface the headline now moves to the new location.
• New import files (all of them have been updated). There is no harm in re-importing these files, so if you already have MyHeadlines do it again. Now we have over 3000 feeds!!!
• Fixed the SOAP client pass by reference error message.

v4.0.6:

• Documentation Menu Item added: THANK YOU Hilary E. House Jr.
• Minor changes to the language file: Now Menu Items have different entries so that you may edit the menu appearance with out changing other common language file elements. Hint: Look for the "Menu" prefix in the lang file.
• New Feature: Added new functionality to the Subscribe->Edit Subscriptions page. The "Move" option will allow you to easily reposition a source and quickly adjust the layout of your subscriptions.
• New Feature: Added a "New Search" button to the search results page.
• New Feature: Flush Cache on the Administration->Maintenance page you have the option to delete the entire cache table. This has been added to allow admins to quickly reduce the size of the DB in extreme situations.
• New Feature: code named "untouch" - On the Administration->Maintenance page there is a new button that allows the admin to force all sources to the expired state. This is handy if you want to refresh the cache in a hurry.
• Install: if you changed the "Front Page" I will not reset the value to the default credit screen. (trying to play nice...)
• Cache Feature Improved: MyHeadlines will now attempt to detect the previous existence of a story, and NOT re-cache and expired item. This is apparent on slow sources that don't produce a lot of new content. As the content expires from the MyHeadlines cache the expired stories used to re-appear as new items. I have tried to limit this behaviour.

v4.0.5:

• Bug Fix: For Post Nuke not running v7.1.0 - Database prefix is now honoured.
• Bug Fix: Latest News and Search results now show the Source URL in Post Nuke v.7.1.x This one only seemed to happen with Post Nuke, but it was due to core PHP array index issues (my fault).
• Cookie Code has been adjusted: header already sent warning messages are squashed.
• New Nag Screen: Shown only to administrators, before they can access the MyHeadlines configuration parameters.
• New Import File: meerkat.imp contains and additional 413 sources!
• New Import File: newsisfree.imp contains and additional 854 sources!
• CMS API Change: the getCache() method now returns more data. See the API.txt file.
• New Feature: The "Show All" screen now sports an XML icon and link to the source. -->> Thanks Dennis Slagers
• New Feature: The "Show All" screen now shows the time left before the cache for the source is updated. -->> Thanks Dennis Slagers
• Typo Fix: vlaign to valign -->> Thanks mike -at- typosquad.com
• Bug Fix: Fixed the RSS v1.0 parsing code. (Makes a lot of sources work again...)
• Bug NOT Fixed: Type casting should resolve the bad type errors on Google searches. But it has not helped. I am still working on this. If you encounter problems with google SOAP API just disable the search feature.
• CMS API Additions: getDashBoard(), Please refer to the API.txt file for more. getDashBoardComponent_Menu(), getDashBoardComponent_MainScreen(), getDashBoardComponent_StockTicker()
• New Parameter: "front_page" Now you may change the frontpage text to anything you want. You have acces to the HTML of the "MyHeadlines v4.0.x by Mike Agar..." and can change it to match up with your site. If you haven't donated to support the development of MyHeadlines, please leave a link or credit indicating how grateful you are to have free access to my code. ;-)

v4.0.4:

• Bug Fix: Admin interface - Category Pulldown box "snaps" to the proper category for the headline source definition.
• New Feature: Google Soap API is now part of the Search feature. You will search the cache always, but have an option to search Google via the new SOAP API.
• New Parameter: You must enter your registration key to enable the Google API. You can register with Google at: http://www.google.com/apis/
• New API function: registerMenuURL() - see API.txt.
• New API function: doGoogleSearch() - see API.txt.

v4.0.3:

• Bug Fix: Multiple entries in the cache for the same story. This is now fixed.

v4.0.2:

• Modified RSS_ITEM class.
• Bug Fix: Admin Category Tree "New" and "Bad" states for sources repaired
• New Feature: "Graduated Error Checking" There is now a threshold variable for the number of consecutive failures before a source is marked "bad". The old method marked it as bad the first time a failure is detected, which caused a number of good sources to be marked bad due to network failures and not source failures. Now you can set the threshold before a source is removed from the User Visible category tree, and placed on the Maintenance list. The default is 3 consecutive failures, which is working for me right now, and you shouldn't have to change it, but it's there if you need to play with it.
• New Feature: "Info" option in admin interface. This shows the Source ID (sid) from the DB and will allow the creation of static pages similar to the very old "Nuke News" of days gone by.
• New Feature: "Date Format" This config variable allows you to use whatever international date format you want. See the "Help" link on the Config screen for more info.

v4.0.1:

• Enriched the cache table object: (click thrus per day)

v4.0.0:

• nearly a complete re-write...
• Improved "look and feel" the HTML rendering functions are much cleaner and special chars like "&" are converted back to the normal values allowing for links and formating to be preserved.
• New Architecture: CMS Abstraction Layer! MyHeadlines is now a "Stand Alone Product". It no longer requires PostNuke, or PHPNuke! But it can easily be converted to modules for both CMS. Please refer to the included document "API.txt" for more information.
• New Feature: Click Thru's are now measured and reported in the Top Ten Lists
• New Feature: Top Ten List (and cache table statistics)
• New Feature: (restored old feature from v2.3) "Rename" function for Categories Admin.
• New Feature: Mandatory Site Default Subscriptions through the admin interface
• New Feature: Mandatory Site Default Stock Symbols through the admin interface
• Retired Feature: "allow_anonymous" configuration parameter has been removed. This activity should be the responsibility of the CMS. If you don't want to allow access to anonymous users, then your CMS should stop their entry. MyHeadines will work with registered and anonymous users.
• Retired Feature: "num_collumns" configuration parameter has been removed. This activity should be the responsibility of the CMS. There is a default user subscription function in the CMS API, but if you want to do something different, the CMS API is very flexible and you're free to render the subscriptions anyway you see fit.

v3.1.1:

• New Config Variables: Proxy Server support now available.
• New Feature: Direct link to all of the source's stories for search and latest news.
• Bug Fix: Show Shameless Plug image works again...

v3.1.0:

• New Feature: "Stock Ticker"
• New Block: "Stock Ticker Block"
• New feature: "show_shamless_plug" switch. You can show/hide the Powered by MyHeadlines banners now.
• New feature: "More..." this will show all the stories from this source currently in the cache.
• New Feature: "More" button to browse deeper into searches, and latest news
• New Feature: "Allow Anonymous" configurable parameter.
• Bug Fix: Cleaned the acyion of some of the "Done" buttons
• Adjustment: No longer display a message to the user if a source timed out. Only the admin will know if a source is bad (Admin->Maintenance)
• New Admin Feature: Show User Subscriptions.

v3.0.1:

• Updated to work with both v.6.x and v.7 PostNuke versions
• Implemented transparent incremental upgrade script

v3.0 Final:

• Dramatically improved the search feature
• auto refresh for cron impaired users
• code cleanup

v3.0 Beta 3b:

• Bug Fix: addCategory() error line 195 core.php
• Bug Fix: reset() error in rss.php

v3.0 Beta 3:

• Bug Fix: Some Headlines Never Get Cached
• New Feature: Search
• New Feature: Latest News
• New Feature: cron
• New setting: override_logo_width is now turned on by default

v3.0 Beta 2:

• Improvement: Bulleted Headline Lists
• Configuration Variable Added: max_logo_width This variable helps reduce theme break-up. Refer to the docs.
• Bug Fix: Left Bottom & Right Bottom Contention Thanks to David Bailey, http://www.patriotfiles.com
• Bug Fix: Can't Create MyHeadlines Blocks Thanks to David Bailey, http://www.patriotfiles.com

v3.0 Beta 1:

• Complete Re-write.

v2.3:

• PHPNuke Addon that needs to be converted to a PostNuke Module