Hello all,

A lot of work has been done over the past two weeks. An announcement would have been made sooner, but I've been busy responding to the aftermath.

• The item history page (/item/123456789) will now resolve current item IDs, making it easier to resolve item histories if you do not know an item's original ID.
  
• Chemistry sets[backpack.tf] now display variants on their stats pages.
  
• Stats pages for Festivized items have been added.
  
• The site now uses the Roboto font, as opposed to Arial and other fallback fonts, except on macOS where the font is still Helvetica.
  
• New navbars for account/classifieds areas of the site to simplify navigation.
  
• Inventory lookups should now be more reliable with less API failures.

• Fixed items appearing several times in name suggestions on the item filter modal.
  
• Fixed certain Normal-quality items appearing twice in the pricegrid.
  
• Fixed the particle effect filter on the suggestion browser.
  
• Fixed a dozen or so CSRF/XSS vulnerabilities (thanks "FREE SHKRELI").
  
• Fixed certain crates not showing the right contents on stats pages
  
• Removed the crate item chance values, because there have been so many redistributions and changes that a lot of these are wrong
  
• Fixed the Random Craft Hat appearing with a normal quality USD price in search
  
• Fixed search redirects for Strange Unusuals
  
• Fixed the "gifted by" link on the Something Special for Someone Special
  
• Fixed classifieds API names parameter not being functional
  
• Fixed the display of the negative trust icon on mini profiles
  
• Fixed profile report button behaviour on different pages which had the profile banner
  
• Premium search has been capped to 1,000 pages because people would click the last page button for some reason which would take them to page 34,567 or something and Elasticsearch would throw a fit
  
• Fixed a bunch of admin utilities you don't know about

The update that hit a few Fridays ago was actually rather huge underneath -- this was the result of about two months spent porting the site to a new templating engine. Was this necessary? Yes, as it alleviated concerns with security and modularity of the software.

backpack.tf is simply huge, and right now there's only one person working on the project. It is difficult to cover all function points when large changes are made, and often I need to rely on user feedback if something goes wrong. This was definitely the case, and I have spent the last couple weeks trying to alleviate these problems. Some issues are more time consuming than others, and over time the TF2 economy has become ridiculously complex with the amount of item states that need to be supported, between Festivizers, Decorated Weapons and so on. There are still some issues that remain unresolved, and I'm trying to fix these -- apologies for this. On the bright side, the code base is looking better than it has ever been.


Automatic is no longer being supported because of concerns for security and architecture of the software. We couldn't simply shut it down, however, as many people still depend on the software, but it is not receiving any further updates with the exception of possible security patches. I simply do not have the time to maintain this on top of the projects I'm already managing.

Hereforth, Automatic is no longer being treated as first class user agent. It didn't feel right to couple the software to the site as closely as it was, especially in its current state. My vision is for Automatic's successor to be one of several classifieds user agents that work with the site.

There's plenty of things we could do to improve Automatic, from how it's developed to how it's deployed. We will take what we have learned and use it to build something better.

• A full restructuring of the web API: A lot of people are criticising the web APIs for several valid reasons. I want to release a new set of APIs that have a sensible and fluent design. This will arrive prior to the classifieds agents successor, and I'll be discussing my plans with people who use the APIs currently to formulate a specification that can be agreed on.
  
• Open-source contributions: backpack.tf has a lot of code, and certain solutions just need a project of their own with which ample code coverage can be guaranteed. This has begun with the marketplace.tf API bindings[github.com] used by the site, which have been rewritten and published to Packagist. I hope to do the same with components and tools that I feel will find use by other people.
  
• Frontend improvements: Client-side JavaScript is still broken in Safari, and there are some weird problems here and there. This will take some time to resolve because some of these problems are fundamental to the JavaScript stack and will take some time to rework. Apologies for this.
  
• Calculator improvements: The current Calculator was written way before the economy became as diverse as it did, and hasn't seen much change since its introduction three years ago, and so it was written pretty optimistically. I want to rework this into a useful tool again, but this is on hold until the frontend is improved.
  
• Improved marketplace.tf integration: I'm not sure how to approach this just yet. This would mean a third inventory value that would be displayed alongside SCM and backpack.tf community value, and your marketplace.tf stock would be merged with your current inventory on the inventory viewer (because after all, you still technically own those items). However, I'm not 100% sure if this behaviour would be desired -- I feel that users might prefer a tab to switch between inventory sources, as if the 'My Marketplace Inventory' page were public and accessible to anyone. Thoughts on this are on hold until after the frontend improvements are complete.

There are still a few issues I need to resolve, and from there we will begin work on these projects. Again, apologies. It'll get better from here.