How does ripgrep exclude binary files?

[d3lm]

I was wondering how ripgrep excludes binary files. Does it use something like from_utf8 and if it fails it won't further process the file? I'd imagine that's not the case because it's very optimized and performs search on raw bytes.

The reason I am asking this is because ripgrep is used by VSCode and it's not returning file matches within binary files. I was wondering if that's something ripgrep does or if VSCode maybe passes some exclude pattern.

[d3lm]

Oh I think I actually just found the answer here. It looks for a NUL byte.

[BurntSushi]

Yup that is indeed correct! I wrote that section because binary data detection is quite complicated in how it mingles with the various filtering strategies and search strategies.

And note that the -a/--text flag can be passed which will disable all binary detection whatsoever.

[d3lm]

Yep, thanks a lot! For my smaller ripgrep clone I am using from_utf8 actually even though that's also not 100% correct, but I figured it might be a bit better than just checking for a NUL byte.