Note by @fastfinge

If you use the eloquence64 addon for #eloquence with #NVDA, release v22 is available. This one has some interesting new things:
• speedups: we've lowered some buffers, and tightened several calls, making eloquence less laggy overall
• antivirus false positive fixes: we've changed the way the 32-bit eloquence host is built, in an effort to load faster, and hopefully reduce some of the false positives from Windows Defender
• experimental support for OpenEVV: we refactored the internals of how eloquence_64 works, so that when OpenEVV is ready, it will be easy for us to switch over. For now, OpenEVV is off by default (turn it on in settings and restart NVDA if you want it), and not recommended for average users. Only US English is supported, and no dictionary loading is available. However, if you only use US English and don't need dictionaries, OpenEVV is ready for you to try. As OpenEVV gains fixes, they can be adopted quickly into this addon.
github.com/fastfinge/eloquence_64/releases/tag/v22
#nvdasr #openevv

@fastfinge Congratulations! Yours now beats mine in the keypress delay war.

@nick That’s why it took a while lol. Release isn’t done until I’m faster than you. Jk

@fastfinge LOL. Which puts me in an interesting position. Admittedly, i still have the edge case when I occasionally use Windows on my macBook Neo, where every bite of RAM matters. I'll have to compare to see if your RAM usage matches mine, but if it does...Genuinely, the only reason to keep developing Eloquence RS becomes the stuff other people use, not me.

@nick It might be with openevv. I doubt it otherwise. But being entirely in Python made it easy for me to add. It’s the identical code, just running in 64 bit.

@fastfinge Indeed. OpenEVV still does a few things that drive me up the wall, which is why I'm not using it, but...Yeah.

@fastfinge Well, if it does end up that RAM usage of the hosts is still mine being smaller, I'll continue, but cherry picking your code.