after forcing UTF-8 on all files before we read them in https://github.com/librariesio/bibliothecary/pull/565 , it was revealed that a pip-compile regexp we have is very generic and catching everything with "require" in the filename, which causes Bibliothecary.analyze() to blow up on binary files.
This rescues the encoding error and skips analysis, although a better fix might be to detect binary files and skip reading them altogether (and to just fix that regexp so it's less generic).
after forcing UTF-8 on all files before we read them in https://github.com/librariesio/bibliothecary/pull/565 , it was revealed that a pip-compile regexp we have is very generic and catching everything with
"require"
in the filename, which causes Bibliothecary.analyze() to blow up on binary files.This rescues the encoding error and skips analysis, although a better fix might be to detect binary files and skip reading them altogether (and to just fix that regexp so it's less generic).
closes https://github.com/librariesio/bibliothecary/issues/590