New Features

• New items can now be added to discmaster! Expect a steady stream of new items to be added every day going forward!
• 3D poly objects are now supported and can viewed (textured or wireframe), rotated and zoomed
• Shockwave Flash files can be played directly in the browser by clicking the icon
• Audio, music and video files are now transcribed and spoken words within them can now be found with search
• You can now download entire folders in addition to individual files
• When viewing a text file, you can now change what encoding you want to view the file as
• ANSI art files can now be viewed as animations with different baud rates and fonts
• You can now search for items on the category pages
• Symlinks that point to other files within the same item now properly link to that file when clicked
• A details pagecan now be seen for any file by clicking the icon or adding ?details to the end of the URL. On this page you can also:
  • View a hex dump of the file
  • See a list of all formats that the file was detected as
  • Open the file in a web based Mac OS 8.1 emulator
  • Open the file in a web based Windows 98 emulator
  • View the file as text (can also see this by adding ?text to the end of a URL)
  • View the raw JSON data of how it was processed by dexvert (can also see this by adding ?json to the end of a URL)

Improvements

• All files have been re-converted with the latest dexvert with over a year of new improvements:
  • Full support for 1,500 new file formats (~3,000 total)
  • Improved detection AND conversion of previously supported formats
  • Able to identify an additional 3,700 unsupported formats (~6,174 total)
  • The above improvements yielded over 100 million more files (~20%) being extracted from the old discmaster1 items
• Image filenames on browse/search pages are now visible even if you don't have CSS support. Additionally the filenames can be highlighted/selected and are findable via Ctrl-F. (Previously these names were generated via CSS which had a negative side effect of making them un-selectable and only visible on browsers with CSS support)
• Image filenames now show up on more image sizes. They also should be smaller on smaller images so they don't cover up as much of the image
• HTML files from items now render better in regards to differing charsets and links to other files
• Macintosh files show their file type and creator code
• Various ANSI-art formats now show title, author and group meta information when available
• File sizes are now shown to 1 decimal point such as 1.4MB (before you'd just see 1MB)
• You can tile images as a repeating backgound image
• Downloading an HTML file now downloads the original un-modified file (before there was no way to access the un-modified version, you always got the discmaster modified HTML)
• Browsing pages and search results now have sections for each family of file (before many were combined into 'Other')
• Changed video playback speed dropdown to be a regular SELECT element for better browser compatability (before it was a hacky CSS/label monstrosity)

Search Improvements

• Search is now substantially faster!
• Can search for folder names
• Extensions can now be searched separately from the main query. This allows you to restrict content queries to only include files with a given extension or exclude specific extensions (-.ext). The filename still has the extension in it, so just searching 'png' in the main 'search for' query will still return *.png files
• New advanced search operators
  • You can now use ? in the search to match any character, such as t?st matches test
  • You can now use astericks * at the *start or end* of a word to find words ending or starting with a word
  • Words can now be grouped with (parenthesies AND (use | combinators))
  • Results now properly honor multiple -word uses instead of just the first one
  • You can use $ to match the end of a field or ^ to match the start of a field
  • You can use << to ensure one word is matched before another
• New option available to 'Omit Unsupported' which excludes any files that are not currently supported by dexvert/discmaster
• When choosing to 'Omit Dups', the number of duplicates per file is shown. You can also now sort by the # of dups
• New detections field that you can search in:
  • Can enter any text from the 'Detection' column on the hex page of a file. This allows you to find files that were detected by dexvert's detector but were necessarily identified or handled by dexvert
  • Can search for mac file type/creator codes
  • Can search for the first 16 bytes of any file (HEX and ASCII (non-ascii bytes as •)). For example, to find any file that starts with IMPM you can search as ASCII [IMPM*] or as hex [494d504d*]
• You can now sort by: family, format and file path (file path groups files together in their respective folders/archives according to their path in the item)
• Search results now show original letter case of filenames (discmaster1 search results were allways lowercased)
• Search results now correctly sort by relevance when chosen (before it was always filename, even when relevance was chosen)
• Improved support for non-english searches, especially Japanese
• Can search the full contents of indexed HTML pages (discmaster1 it only indexed the text content of the page, which meant you couldn't find script content, specific HTML tags, etc.)
• Newly added items are available to search immediately (discmaster1 was on a 1-2 day delay)
• Can search for files based on the file hash (blake3) by adding &b3sum=<hash> to the search query in the address bar
• The JSON output of search now returns a TON more information about each result (including file hash (blake3/b3sum))
• When including the item name in search results, the item category is now also shown
• Maximum number of results per page is now 500 (up from 250)
• The search URL now uses human readable words for Genre and Category searches instead of id numbers

Other Changes

• Search counter added to index page
• Hit counter on index page now increments for every page served, rather than just the index landing page
• To save on disk space, hybrid Mac/PC CDs now replace duplicate files on the mac side with symlinks to the duplica PC file
• Images are no longer visually classified by AI and tagged with labels, thus searching for images by label is no longer supported. Freely available AI models to do this hallucinated and produced wildly inacurate labels and this feature also consumed a lot of GPU time when processing