From fed28c3ecadffdfc5be38895872590b6b0cfdb20 Mon Sep 17 00:00:00 2001
From: rustmailer <pollybase@zohomail.com>
Date: Wed, 28 Jan 2026 13:19:33 +0800
Subject: [PATCH] fix: add tolerant HTML-to-text extraction (#141)

---
 src/modules/envelope/extractor.rs | 11 +++--------
 src/modules/utils/html.rs         | 28 ++++++++++++++++++++++++++++
 src/modules/utils/mod.rs          |  1 +
 3 files changed, 32 insertions(+), 8 deletions(-)
 create mode 100644 src/modules/utils/html.rs

diff --git a/src/modules/envelope/extractor.rs b/src/modules/envelope/extractor.rs
index d428bdb..9a9d58e 100644
--- a/src/modules/envelope/extractor.rs
+++ b/src/modules/envelope/extractor.rs
@@ -21,6 +21,7 @@ use crate::modules::envelope::utils::normalize_subject;
 use crate::modules::error::code::ErrorCode;
 use crate::modules::error::BichonResult;
 use crate::modules::utils::create_hash;
+use crate::modules::utils::html::extract_text;
 use crate::{calculate_hash, raise_error, utc_now};
 use crate::{id, modules::indexer::envelope::Envelope};
 use async_imap::types::Fetch;
@@ -48,10 +49,7 @@ pub fn extract_envelope(fetch: &Fetch, account_id: u64, mailbox_id: u64) -> Bich
     let text = if let Some(text) = message.body_text(0).map(|cow| cow.into_owned()) {
         text
     } else if let Some(html) = message.body_html(0).map(|cow| cow.into_owned()) {
-        html2text::config::plain()
-            .allow_width_overflow()
-            .string_from_read(html.as_bytes(), 100)
-            .map_err(|e| raise_error!(format!("{:#?}", e), ErrorCode::InternalError))?
+        extract_text(html)
     } else {
         String::new()
     };
@@ -150,10 +148,7 @@ pub fn extract_envelope_from_eml(
     let text = if let Some(text) = message.body_text(0).map(|cow| cow.into_owned()) {
         text
     } else if let Some(html) = message.body_html(0).map(|cow| cow.into_owned()) {
-        html2text::config::plain()
-            .allow_width_overflow()
-            .string_from_read(html.as_bytes(), 100)
-            .map_err(|e| raise_error!(format!("{:#?}", e), ErrorCode::InternalError))?
+        extract_text(html)
     } else {
         String::new()
     };
diff --git a/src/modules/utils/html.rs b/src/modules/utils/html.rs
new file mode 100644
index 0000000..b933fbd
--- /dev/null
+++ b/src/modules/utils/html.rs
@@ -0,0 +1,28 @@
+use std::panic;
+use tracing::error;
+
+pub fn extract_text(html: String) -> String {
+    let result = panic::catch_unwind(|| {
+        html2text::config::plain()
+            .allow_width_overflow()
+            .string_from_read(html.as_bytes(), 100)
+    });
+
+    match result {
+        Ok(Ok(text)) => text,
+        Ok(Err(err)) => {
+            error!("html2text error: {}", err);
+            html
+        }
+        Err(err) => {
+            if let Some(s) = err.downcast_ref::<&str>() {
+                error!("html2text panic: {}", s);
+            } else if let Some(s) = err.downcast_ref::<String>() {
+                error!("html2text panic: {}", s);
+            } else {
+                error!("html2text panic: unknown error");
+            }
+            html
+        }
+    }
+}
diff --git a/src/modules/utils/mod.rs b/src/modules/utils/mod.rs
index 68b3017..57610b7 100644
--- a/src/modules/utils/mod.rs
+++ b/src/modules/utils/mod.rs
@@ -26,6 +26,7 @@ use rand::{rng, Rng};
 use super::error::code::ErrorCode;
 
 pub mod encrypt;
+pub mod html;
 pub mod net;
 pub mod rate_limit;
 pub mod shutdown;