Replace Cyrillic c with c in rest of caCert variables #2492
Conversation
|
|
|
Too many Russians in the team, but source of this one is actually external contribution. |
|
Happily this is the only one case but I think we should add a linter/validator to check entire code base for anything but standard ASCII symbols. |
@alexshtin that sounds like a great idea, especially since there are a diverse set of contributors to the repo, this looks like it is an easy mistake to make if you are using a non-English keyboard. I've seen other GitHub issues on OSS repos punt on these kind of linters with concern that it pushes everyone to only use English. I think from the security/reliability prospective it can make sense though, given the complexity of unicode. Another idea is to restrict homoglyphs / similar looking unicode characters and bidirectional unicode characters. Rust has a good example of linters to do this (the Go ones I have seen aren't as comprehensive as these in coverage):
This would keep everything unicode friendly but protect from these kind of bugs. |
Summary
Following along #2397, there were still a couple more Cyrillic letters checked-in, likely from copy-pasting the config.
Motivation
To ensure caCert is set from the config.
Testing
Verified presence of the letters with
rg "[^\x00-\x7F]", changed them toc, and then verified again with ripgrep that the letters were no longer present.Risks
Should be low-risk, worst case perhaps someone will hit an error now that their caCert file is set correctly.
Hotfix?
Potentially not, as this is an sqlite development config which if I understand correctly is a feature still in development. Will defer to Temporal for this.