- Fix logging configuration not propagating to child parser processes (#646).
- Update `mailsuite` dependency to `?=1.11.1` to solve issues with iCloud IMAP (#493).
Fixes
- Fix saving reports to OpenSearch ([#637](https://github.com/domainaware/parsedmarc/issues/637))
- Fix parsing certain DMARC failure/forensic reports
- Some fixes to type hints (incomplete, but published as-is due to the above bugs)
## Improvements
- Type hinting is now used properly across the entire library. (#445)
## Fixes
- Decompress report files as needed when passed via the CLI.
- Fixed incomplete removal of the ability for `parsedmarc.utils.extract_report` to accept a file path directly in `8.15.0`.
## Breaking changes
This version of the library requires consumers to pass certain arguments as keyword-only. Internally, the API uses a bare `*` in the function signature. This is standard per [PEP 3102](https://peps.python.org/pep-3102/) and as documented in the Python Language Reference.
.
8.19.0
- Add multi-tenant support via an index-prefix domain mapping file
- PSL overrides so that services like AWS are correctly identified
- Additional improvements to report type detection
- Fix webhook timeout parsing (PR #623)
- Output to STDOUT when the new general config boolean `silent` is set to `False` (Close#614)
- Additional services added to `base_reverse_dns_map.csv`
---------
Co-authored-by: Sean Whalen <seanthegeek@users.noreply.github.com>
Co-authored-by: Félix <felix.debloisbeaucage@gmail.com>
- Fix since option to correctly work with weeks (PR #604)
- Add 183 entries to `base_reverse_dns_map.csv`
- Add 57 entries to `known_unknown_base_reverse_dns.txt`
- Check for invalid UTF-8 bytes in `base_reverse_dns_map.csv` at build
- Remove unneeded items from the `parsedmarc.resources` module at build
- Move `__version__` to `parsedmarc.constants`
- Create a constant `USER_AGENT`
- Use the HTTP `User-Agent` header value `parsedmarc/version` for all HTTP requests
- Merged PR #603
- Fixes issue #595 - CI test fails for Elasticsearch
- Moved Elasticsearch to a separate Docker service container for CI testing
- Dropped Python 3.8 from CI testing
- Fixes lookup and saving of DMARC forensic reports in Elasticsearch and OpenSearch
- Updated fallback `base_reverse_dns_map.csv`, which now includes over 1,400 lines
- Updated included `dbip-country-lite.mmdb` to the June 2025 release
- Automatically fall back to the internal `base_reverse_dns_map.csv` if the received file is not valid (Fixes#602)
- Print the received data to the debug log
- Add support for Microsoft national clouds via Graph API base URL (PR #590)
- Avoid stopping processing when an invalid DMARC report is encountered (PR #587)
- Increase `http.client._MAXHEADERS` from `100` to `200` to avoid errors connecting to Elasticsearch/OpenSearch (PR #589)
- Ignore duplicate aggregate DMARC reports with the same `org_name` and `report_id` seen within the same hour ([#539](https://github.com/domainaware/parsedmarc/issues/539))
- Fix saving SMTP TLS reports to OpenSearch (PR #585 closed issue #576)
- Add 303 entries to `base_reverse_dns_map.csv`
- Fix processing of SMTP-TLS reports ([#549](https://github.com/domainaware/parsedmarc/issues/549)), which broke in commit [410663d ](410663dbca)(PR [#530](https://github.com/domainaware/parsedmarc/pull/530))
- This PR enforced a stricter check for base64-encoded strings, which SMTP TLS reports from Google did not pass
- Removing the check introduced its own issue, because some file paths were treated as base64-encoded strings
- Create a separate `extract_report_from_file_path()` function for processioning reports based on a file path
- Remove report extraction based on a file path from `extract_report()`
- Fix processing of SMTP-TLS reports (#549)
- Skip invalid aggregate report rows without calling the whole report invalid
- Some providers such as GoDaddy will send reports with some rows missing a source IP address, while other rows are fine
- Fix Dovecot support by using the seperator provided by the IPMAP namespace when possible (PR #552closes#551)
- Only download `base_reverse_dns_map.csv` once (fixes#542)
- Update included `base_reverse_dns_map.csv`
- Replace University category with Education to be more inclusive
- Update included `dbip-country-lite.mmdb`
- Skip invalid aggregate report rows without calling the whole report invalid
- Some providers such as GoDaddy will send reports with some rows missing a source IP address, while other rows are fine
- Fix Dovecot support by using the seperator provided by the IPMAP namespace when possible (PR #552closes#551)
- Only download `base_reverse_dns_map.csv` once (fixes#542)
- Update included `base_reverse_dns_map.csv`
- Replace University category with Education to be more inclusive
- Update included `dbip-country-lite.mmdb`
- Fix for deadlock with large report (#508)
- Build: move to kafka-python-ng (#510)
- Fix new config variables previously not propagated in the code (#524)
- Fixes for kafka integration (#522)
- Fix if base_domain is None before get_service_from_reverse_dns_base_domain (#514)
- Update base_reverse_dns_map.csv
- Actually save `source_type` and `source_name` to Elasticsearch and OpenSearch
- Reverse-lookup cache improvements (PR #501 closes issue #498)
- Update the included `dbip-country-lite.mmdb` to the 2024-03 version
- Update `base_reverse_dns_map.csv`
- Add new general config options (closes issue #500)
- `always_use_local_files` - Disables the download of the reverse DNS map
- `local_reverse_dns_map_path` - Overrides the default local file path to use for the reverse DNS map
- `reverse_dns_map_url` - Overrides the default download URL for the reverse DNS map
