To save space in the profile, we can deduplicate strings in marker fields by using the unique-string field type. However, how do I use deduplicated URLs? Or deduplicated sanitized strings? I have to pick either url or unique-string, there's no unique-url or unique-sanitized-string:
|
// ---------------------------------------------------- |
|
// String types. |
|
|
|
// Show the URL, and handle PII sanitization |
|
| 'url' |
|
// Show the file path, and handle PII sanitization. |
|
| 'file-path' |
|
// Show regular string, and handle PII sanitization. |
|
| 'sanitized-string' |
|
// Important, do not put URL or file path information here, as it will not be |
|
// sanitized. Please be careful with including other types of PII here as well. |
|
// e.g. "Label: Some String" |
|
| 'string' |
|
/// An index into a (currently) thread-local string table, aka UniqueStringArray |
|
/// This is effectively an integer, so wherever we need to display this value, we |
|
/// must first perform a lookup into the appropriate string table. |
|
| 'unique-string' |
|
// ---------------------------------------------------- |
I think it would be better to remove the unique-string type and turn all string fields into string indexes.