One thing I noticed. We have a resource in our bundle with the name 'Leads/µcomputer.fxp`
I had to strip this from my bundle because the unzip stage after install gave me this error
[4.843 info src/plugin.cpp:193 extractPackages] Extracting package /Users/paul/Documents/Rack2/plugins/SurgeRack-2.1.7.0-mac.vcvplugin
[5.015 warn src/plugin.cpp:198 extractPackages] Plugin package /Users/paul/Documents/Rack2/plugins/SurgeRack-2.1.7.0-mac.vcvplugin failed to extract: Unarchiver could not read entry from archive: Pathname can't be converted from UTF-8 to current locale.
and then I got an incomplete plugin install and insta-crash.
Seems somewhere in your plugin creation / unpack chain you have lost uff-8 encoding.
I got around it by just deleting that asset from our bundle when we make the dist stage. It’s the only one with a non-ascii filename. But wanted to flag the bug to you.
Yeah, it seems wrong that the unarchiver is trying to convert a UTF-8 path into an ascii one using the current locale. I assume this is on windows? On windows of course you need to convert UTF-8 to a wide character string and pass that to the OS…
yeah i don’t know which part of the toolchain is doing it; but some part is and is doing it newly since we had this patch included in our 1.0 bundle. So definitely a rack 2 regression.
like i said i can work around it. The patch isn’t that great except inasmuch as it forces us to make sure we don’t get our utf-8 filename handling wrong in the surge build and scan processes
yeah, it’s no joke keeping you stuff working with non-ascii paths. I don’t know how one logs a bug against rack these days, now that the git issues database seems closed for VCV. I guess you post a query here and hope andrew notices?
For community-backed support and other resources, see
Community forum
Rack manual
other VCV communities
from https://vcvrack.com/support
“Community backed support” is not going to be easy without the collective “memory” of github issues.
I am not going to try helping out without an issue tracker to point to.
Yeah, I think it’s a mistake. Bad bug reports don’t get better by being sent by email, and now everyone looses the opportunity to see already reported bugs, so tons of duplicate reporting.
