Filter format for URL list-based policies

• 05/01/2020
• 3 minutes to read

In this article

This article describes the filter format used for the Microsoft Edge URL list-based policies (For example, URLBlocklist, URLAllowList, and CertificateTransparencyEnforcementDisabledForUrls policies.

The filter format

The filter format is:

    [scheme://][.]host[:port][/path][@query]

The fields in the filter format are:

Comparing the filter format to the URL format

The filter format resembles the URL format, except for the following differences:

• If you include "user:pass", it will be ignored. For example, http://user:[email protected]/pub/example.iso.
• If you include a fragment identifier ("#"), it and everything that follows the identifier is ignored.
• You can use a wildcard ("*") as the host and you can prefix it with a dot (".").
• You can use a forward slash ("/") or a dot (".") as a suffix for the host. In this case, the suffix is ignored.

Filter selection criteria

The filter selected for a URL is the most specific match found after processing the following filter selection rules:

1. Filters with the longest host match are selected first.

2. From the selected filters, any filter with a non-matching scheme or port is discarded.

3. From the remaining filters, the filter with the longest matching path is selected.

4. From the remaining filters, the filter with the longest set of query tokens is selected. At this step, the allow list filter takes precedence over the block list filter if both filters have the same path length and number of query tokens.

5. If there's no valid filter remaining, then the left-most subdomain is removed from host and the selection process starts over at step 1. The special asterisk ("*") host is the last searched and it matches all hosts.

6. If a filter's available, it blocks or allows the URL request.

   The default behavior is to allow the URL request if no filter is matched.

Example filter selection criteria

In this example, when searching for a match to "https://sub.contoso.com/docs" the filter selection will:

1. Search for a filter for "sub.contoso.com". If it finds a filter, the search moves to step 2. If a filter isn't found, then it tries again with "contoso.com", "com", and finally "".
2. From the selected filters, any that don't have "http" in the scheme are removed.
3. From the remaining filters, any that have an exact port number that isn't "80" are removed.
4. From the remaining filters, any that don't have "/docs" as a prefix of the path are removed.
5. From the remaining filters, the filter with the longest path prefix is selected and applied. If a filter isn't found, the selection process starts over again at step 1. The process is repeated with the next subdomain.

Additional filter information

If a filter has a dot (".") prefixing the host then only exact host matches are filtered. For example:

• "contoso.com" (no dot) will match "contoso.com", "www.contoso.com", and "sub.www.contoso.com"
• ".www.contoso.com" (with a dot prefix) will only match "www.contoso.com"

You can use either a standard or customer schema. Supported standard schemas include:

• about, blob, content, edge, cid, data, file, filesystem, ftp, gopher, http, https, javascript, mailto, ws, and wss.

Any other schema is treated as a custom schema, but only the schema:* and schema://* patterns are allowed. For example:

• "custom:" or "custom://" will match "custom:app"
• "custom:app" or "custom://app" are invalid

schema and host aren't case-sensitive. For example:

• "http://contoso.com" filter matches "HTTP://contoso.com", "http://contoso.COM", and "http://contoso.com"

path and query are case-sensitive. For example:

• "http://contoso.com/path?query=A" filter doesn't match "http://contoso.com/Path?query=A" or "http://contoso.com/path?Query=A". It does match "http://contoso.COM/path?query=A".

Content license

This work is licensed under a Creative Commons Attribution 4.0 International License.

See also

• Microsoft Edge Enterprise landing page