Bug: Log File Collects PII

[kryzet]

Description

When uploading the log file for a bug report, personally identifiable information (PII) is included in the log, such as email addresses (and possibly full names). I did not thoroughly investigate the issue.

Steps To Reproduce

Open a log file from one of the GitHub Issues.

Requirements

A good solution might be to:

• Redact log lines that include PII
• Entirely avoid logging events that might involve PII

Files Version

3.7.11.0

Windows Version

10.0.22631.4317

User ID

No response

Log File

debug.log

[0x5bfa]

Thanks for the feedback.
We can't detect personal information but we have no problem with the log to be deducted to protect it.
Where was your email address recorded, like drive name?

[kryzet]

If you check the log file, you will find redacted lines that included my email address. You will also find lines that were redacted despite not including personal information, but that were related to the other lines that were redacted. Every line was either left unmodified or redacted with a comment explaining what information was stored on that line.

[yaira2]

In line with our privacy policy, only non-identifiable information is uploaded to Sentry. The local log file you're referring to generally doesn't contain private information, although it does log drive paths for fixed and cloud drives. This log file is never uploaded to Sentry, but it can be incredibly helpful for debugging issues when users choose to upload it to GitHub.

If you find differently, please let us know. However, it's most likely that the email address you found in the log file is the name for a Google Drive or iCloud account listed in the sidebar.

[kryzet]

If you find differently, please let us know. However, it's most likely that the email address you found in the log file is the name for a Google Drive or iCloud account listed in the sidebar.

That's actually the exact way my email address landed in the log. Why isn't that information excluded? Isn't it a matter of a simple regex to detect email addresses and remove them?

[yaira2]

The drive names are logged to help debug issues where they aren't detected properly.

[kryzet]

You could redact just the email addresses and leave the rest. If you need to distinguish between multiple cloud provider drives, you can assign every email address an anonymous identifier, such as <email #>.