fkoyer opened a new pull request, #49:
URL: https://github.com/apache/spamassassin/pull/49
**Target release: 4.0.4** (new plugin loaded via `rules/v404.pre`).
## Summary
Adds a new `main_body` rule type. A `main_body` rule works like a `body`
rule,
but matches only the "main" body: the rendered text of the inline body parts,
excluding attachments.
Part selection:
- Named parts (attachments) are skipped.
- For a `multipart/alternative` section, the one part named by
`multipart_alternative_preferred_part` is used; when that setting is unset,
`text/html` is preferred, falling back to `text/plain`.
- Other multipart containers (e.g. `multipart/mixed`) are descended into, and
every inline text part they contain is included, in display order.
As with `body` rules, the message Subject is prepended and the same
whitespace
normalization is applied, and the text is the full rendered text (including
any
text not visible in a mail client). The `multiple`, `maxhits`, and
`nosubject`
tflags are supported.
## Motivation
`body` rules match the concatenated rendered text of every body part of a
message, including text extracted from attachments (PDFs, images, etc.) and
every alternative. There is currently no way to write a rule that matches
only
the text a mail client would present as the body. `main_body` fills that gap.
## Implementation
- `Mail::SpamAssassin::Plugin::MainBody` — a self-contained plugin that
registers
the `main_body` config keyword and runs the rules, reusing
`run_generic_tests`
semantics so scoring, tflags, and meta-rule integration behave like `body`.
- `Mail::SpamAssassin::Message::get_main_body_parts()` — new shared accessor
that
performs the part selection and returns the inline body-part nodes in
display
order.
- `Mail::SpamAssassin::Message::get_main_body_text_array()` — returns the
concatenated rendered text for the `main_body` rules, reusing the existing
body
whitespace-normalization path.
- `Mail::SpamAssassin::Plugin::BodyEval::vertical_whitespace` is refactored
to use
the new shared `get_main_body_parts()` accessor instead of its own inline
walk.
- Loaded via `rules/v404.pre`.
## Testing
Adds `t/main_body.t` (16 subtests): attachment exclusion, the
`multipart/alternative` default (text/html over text/plain) and override, the
text/plain fallback, gathering both parts of a `multipart/mixed`, the
`multiple`
and `maxhits` tflags, Subject handling with and without `nosubject`, and a
meta
rule depending on a `main_body` rule. The full suite (`t/*.t`) passes.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
