Work-In-Progress Reports for September 2008

Analyser: Specific Detection of Geometry Inconsistency: Another item from our to-do list is now complete, specifically: "Script generator specific detection of geometry inconsistency; i.e. the tracks currently marked as “extra”, highlighting any suspect/modified track even if the write splice test passes. (more)

Analyser: Track Sizes and Geometry Changes: Tracks containing fake big sectors, and smaller real sectors as well, fully resolve to their real world size. Although technically the data content was previously correct, finding the real size helps some algorithms in the code. Found one image (out of a few hundred...) where the current method of detecting geometry changes is not enough to find the first written block. (more)

Analyser: Script Generator Data Checks: Changed the encoding check used by the script generator to the recording property check used by the main analysation process. This has significantly improved the reliability of detecting bad data, write splices, recorded areas and so on during the scripting process - although processing wise it takes slightly more time of course. (more)

Analyser: Gap Changes + Protection Support: Slightly changed the code used for detecting sudden changes in gap size; when there are only 2 gaps to compare, each one could qualify as a sudden change compared to the other. Instead, the gap size itself should be compared to get the expected result. The protection used on some US Gold games for the Atari ST is fully supported now. The RUBI protection v2 used on CPC Puffy's Saga fully resolves now. (more)

Analyser: "Anco" Variant: When checking the images using the recently supported Copylock protection variant it was noticed that Atari ST games by Anco (we have some Kick Off 2 dumps as samples) seemed to have similar signs of protection to the recently supported Copylock format. Investigating further revealed that they indeed use a very similar protection and these games now are fully supported.

Analyser: Complex Copylock Variant Part II: When checking the images using the recently supported Copylock protection variant it was noticed that the check retried and failed several times if the protection was written at any track other than track 0. As it turned out, there was an additional element: with each track and side, the distance of the overlapping area changed slightly. (more)

Analyser: Complex Copylock Variant: There was some missing functionality required to support a Copylock ST variant as described in the previous wip. This has now been implemented to fully support the geometry based Copylock variant, mainly: * The ability to remove any group(s) of segments from the pre-processed data if they match a pattern, and then restart the analysation process. (more)

Analyser: Copylock ST Additions: * Added a new functionality to pattern matching where the size of any logical area within a segment can be checked to be within predefined boundaries. This was required for the reliable detection of one specific Copylock ST variant. * Added one Copylock ST variant using short blocks and a key. In order to properly replicate the exact geometry though some new functionality needs to be added.

Analyser: Misc Work: Visualisation * New, 'O' flag displayed for data overlay. * Flags that signal special, auto-detected attributes (X, O, M, W) are now displayed with protection names. New Pattern-Matching Descriptors * Added KBI 1986 protection pattern. * New patterns for protections used on Infogrames ST titles ("starter"). * HLS ST variants patterns. * Protection used on at least the ST versions of Sargon III, Rugby Leage Boss and Crash Garrett. (more)

Analyser: Data Overlays: All possible data overlays are now supported automatically: * Example: A sector embedded within another sector, a longer sector continuing on another, etc. * All checksums and checksum areas are correctly scripted. * Multiple dependencies are supported so sectors involved in the overlay can themselves create another overlay. * Sectors involved in a valid overlay dependency are no longer marked as truncated. (more)