Status Update - September 2020
Added 2020-10-02 22:20:28 +0000 UTCSeptember was again a month full of emulation and experimentation with console games modding!
DuckStation (and PlayStation in general)
This month, I submitted too many little pull requests to DuckStation to mention them all (the link takes to a list of all my PRs from September), so I'll only mention the most important ones:
- Fixed issues with a -resume command line parameter - originally intended to load a resume save state of the game started from the command line, it effectively did not work
- Fix Discord RPC not reporting the game title - while DuckStation has Discord integration, it never displayed the name of the game the user was playing. This was a bug.
- libchdr - fix a memory leak - a bug in libchdr, a library DuckStation uses to read CHD files resulted in a tiny memory leak every time a CHD file was loaded. DuckStation had a ton of leaks like this which were fixed in an updated version of the library it did not use - however, even after updating the library to the newest version, one leak persisted. I identified and fixed it.
- CPU overclocking - not made by me, but I helped kickstart the development of that feature by "hacking around" an overclock implementation by modifying two lines of code, and this seemed to have given Stenzek some ideas on how to implement overclocking properly without any performance hit.
Finally, as I mentioned in a post from a few days ago, I started experimenting with mods/cheat codes for console games, starting with PS1 games running on DuckStation. I improved several existing codes for Gran Turismo and Gran Turismo 2, ported them to other game versions, and created one cheat of my own, which you can check HERE.
RPCS3
In September, I only fixed a single bug in RPCS3. I noticed that running games with a --no-gui parameter reset their playtime and fixed it.
Other
Although not related to anything I've done last month, it's worth mentioning that my reported Visual Studio compiler bug got reviewed and was confirmed to be an optimizer bug dating to the early 1990s! Even though we still don't know when will it be fixed, it's very rewarding to know that I (accidentally, to be honest) found a bug this old.
Additionally, just recently I found out that LLVM has picked up and fixed the bug I wrote a blog post about back in February! I never got to report it due to the high barrier of entry for their bug tracker, but turns out they have acknowledged the bug on their own.