has there been a regression in 1.44.3 in deno ffi?
the smallest repro i could make:
https://gist.github.com/teidesu/199abcd47a3fb494564a2c41965be5ad
on linux amd64 host, put these 2 files in the same directory, and run
docker run --rm $(docker build -q .)
then, change deno version to 1.44.2 in dockerfile and run the same command again
on 1.44.3 the code fails with a random error from the library being ffi-ed (e.g. pcre2_substitute_8 error: -8. sometimes it's different error code, but always PCRE2_ERROR_NULL or PCRE2_ERROR_UTF8_ERR family), and on 1.44.2 everything works fine and outputs goodbye world correctly
on the other hand, everything seems to work fine when running outside docker or on arm64 linux, so i'm not really sure what could be the issue...Gist
deno ffi issue repro
deno ffi issue repro. GitHub Gist: instantly share code, notes, and snippets.
4 Replies
not opening a github issue because im not entirely sure if this is actually an issue with deno
We did a lot of FFI improvements around that time. You might want to give the latest version a try to see if the issue is resolved there
the issue is present in every version after 1.44.3, 1.44.2 is the newest version that works fine :c
I see, that definitely sounds like a bug. Best to open an issue for that here https://github.com/denoland/deno/issues/new/choose