优化不可改写区判定逻辑

Author: GTJasonMK

src-tauri/src/adapters/docx/simple_extract_runs.rs

@@ -409,162 +409,9 @@ pub(super) fn push_writeback_editable_text_regions(
 
 pub(super) fn split_editable_text_segments<'a>(
     text: &'a str,
-    presentation: &Option<TextPresentation>,
+    _presentation: &Option<TextPresentation>,
 ) -> Vec<(&'a str, bool)> {
-    let mut segments = Vec::new();
-    for (segment_text, allow_rewrite) in split_structured_text_segments(text, presentation) {
-        if !allow_rewrite {
-            segments.push((segment_text, false));
-            continue;
-        }
-        extend_url_locked_segments(&mut segments, segment_text);
-    }
-    segments
-}
-
-pub(super) fn split_structured_text_segments<'a>(
-    text: &'a str,
-    presentation: &Option<TextPresentation>,
-) -> Vec<(&'a str, bool)> {
-    if !presentation.as_ref().is_some_and(|item| item.underline) || !text_has_visible_content(text)
-    {
-        return vec![(text, text_has_visible_content(text))];
-    }
-    let Some((content_start, content_end)) = text_content_bounds(text) else {
-        return vec![(text, false)];
-    };
-    if content_start == 0 && content_end == text.len() {
-        return vec![(text, true)];
-    }
-    let mut segments = Vec::with_capacity(3);
-    if content_start > 0 {
-        segments.push((&text[..content_start], false));
-    }
-    segments.push((&text[content_start..content_end], true));
-    if content_end < text.len() {
-        segments.push((&text[content_end..], false));
-    }
-    segments
-}
-
-pub(super) fn text_content_bounds(text: &str) -> Option<(usize, usize)> {
-    let start = text.char_indices().find(|(_, ch)| !ch.is_whitespace())?.0;
-    let (end_start, end_ch) = text
-        .char_indices()
-        .rev()
-        .find(|(_, ch)| !ch.is_whitespace())?;
-    Some((start, end_start + end_ch.len_utf8()))
-}
-
-pub(super) fn extend_url_locked_segments<'a>(segments: &mut Vec<(&'a str, bool)>, text: &'a str) {
-    let spans = bare_url_spans(text);
-    if spans.is_empty() {
-        segments.push((text, true));
-        return;
-    }
-
-    let mut cursor = 0usize;
-    for (start, end) in spans {
-        if cursor < start {
-            let prefix = &text[cursor..start];
-            segments.push((prefix, text_has_visible_content(prefix)));
-        }
-        segments.push((&text[start..end], false));
-        cursor = end;
-    }
-    if cursor < text.len() {
-        let suffix = &text[cursor..];
-        segments.push((suffix, text_has_visible_content(suffix)));
-    }
-}
-
-pub(super) fn bare_url_spans(text: &str) -> Vec<(usize, usize)> {
-    let mut spans = Vec::new();
-    let mut index = 0usize;
-    while index < text.len() {
-        let slice = &text[index..];
-        let prefix_len = if slice.starts_with("https://") {
-            Some("https://".len())
-        } else if slice.starts_with("http://") {
-            Some("http://".len())
-        } else if slice.starts_with("www.") {
-            Some("www.".len())
-        } else {
-            None
-        };
-        let Some(prefix_len) = prefix_len else {
-            index += text[index..]
-                .chars()
-                .next()
-                .map(|ch| ch.len_utf8())
-                .unwrap_or(1);
-            continue;
-        };
-        if !url_start_allowed(text, index) {
-            index += prefix_len;
-            continue;
-        }
-        let end = find_bare_url_end(text, index, prefix_len);
-        if end > index + prefix_len {
-            spans.push((index, end));
-            index = end;
-        } else {
-            index += prefix_len;
-        }
-    }
-    spans
-}
-
-pub(super) fn url_start_allowed(text: &str, start: usize) -> bool {
-    let Some(prev) = text[..start].chars().next_back() else {
-        return true;
-    };
-    !(prev.is_ascii_alphanumeric() || matches!(prev, '/' | '.' | '_' | '-' | '@'))
-}
-
-pub(super) fn find_bare_url_end(text: &str, start: usize, prefix_len: usize) -> usize {
-    let bytes = text.as_bytes();
-    let mut end = start;
-    while end < bytes.len() && !bytes[end].is_ascii_whitespace() {
-        end += 1;
-    }
-    while end > start + prefix_len && url_trailing_punctuation(text[..end].chars().next_back()) {
-        end -= text[..end]
-            .chars()
-            .next_back()
-            .map(|ch| ch.len_utf8())
-            .unwrap_or(1);
-    }
-    end
-}
-
-pub(super) fn url_trailing_punctuation(ch: Option<char>) -> bool {
-    matches!(
-        ch,
-        Some(
-            '.' | ','
-                | ';'
-                | ':'
-                | '!'
-                | '?'
-                | ')'
-                | ']'
-                | '}'
-                | '"'
-                | '\''
-                | '。'
-                | '，'
-                | '；'
-                | '：'
-                | '！'
-                | '？'
-                | '）'
-                | '】'
-                | '」'
-                | '』'
-                | '、'
-        )
-    )
+    vec![(text, true)]
 }
 
 pub(super) fn parse_writeback_formula_region(

src-tauri/src/adapters/docx/slots.rs

@@ -101,7 +101,7 @@ fn build_paragraph_slots(
 
         let split_count = fragments.len();
         for (fragment_index, fragment) in fragments.into_iter().enumerate() {
-            let slot_editable = editable && text_has_visible_content(&fragment.text);
+            let slot_editable = editable && !fragment.text.is_empty();
             slots.push(WritebackSlot {
                 id: slot_id(&anchor, fragment_index, split_count),
                 order: start_order + slots.len(),

src-tauri/src/adapters/docx/tests/import_basic.rs

@@ -411,7 +411,7 @@ fn report_template_keeps_first_heading_numbered_as_chapter_one() {
 }
 
 #[test]
-fn imports_underlined_blank_runs_as_locked_underlined_text() {
+fn imports_underlined_blank_runs_as_editable_underlined_text() {
     let xml = r#"<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
 <w:document xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main">
   <w:body>
@@ -438,13 +438,13 @@ fn imports_underlined_blank_runs_as_locked_underlined_text() {
         .expect("underlined blank presentation");
 
     assert_eq!(rebuilt, "填写日期：　　　　");
-    assert!(blank_region.skip_rewrite);
+    assert!(!blank_region.skip_rewrite);
     assert!(presentation.underline);
     assert_eq!(presentation.protect_kind.as_deref(), None);
 }
 
 #[test]
-fn splits_underlined_run_edge_whitespace_into_locked_regions() {
+fn keeps_underlined_run_edge_whitespace_editable() {
     let xml = r#"<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
 <w:document xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main">
   <w:body>
@@ -473,14 +473,11 @@ fn splits_underlined_run_edge_whitespace_into_locked_regions() {
         .collect::<Vec<_>>();
 
     assert_eq!(rebuilt, "作品编号：　　ABC123　　　");
-    assert_eq!(
-        underlined_regions,
-        vec![("　　", true), ("ABC123", false), ("　　　", true)]
-    );
+    assert_eq!(underlined_regions, vec![("　　ABC123　　　", false)]);
 }
 
 #[test]
-fn writes_back_underlined_run_with_locked_edge_whitespace() {
+fn writes_back_underlined_run_with_editable_edge_whitespace() {
     let xml = r#"<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
 <w:document xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main">
   <w:body>
@@ -498,9 +495,9 @@ fn writes_back_underlined_run_with_locked_edge_whitespace() {
     let mut regions = DocxAdapter::extract_writeback_regions(&bytes).expect("extract regions");
     let editable_region = regions
         .iter_mut()
-        .find(|region| !region.skip_rewrite && region.body == "ABC123")
+        .find(|region| !region.skip_rewrite && region.body == "　　ABC123　　　")
         .expect("editable fill content");
-    editable_region.body = "ZX-9".to_string();
+    editable_region.body = "　　ZX-9　　　".to_string();
 
     let rewritten = DocxAdapter::write_updated_regions(&bytes, &source, &regions)
         .expect("write updated regions");
@@ -516,9 +513,5 @@ fn writes_back_underlined_run_with_locked_edge_whitespace() {
         .map(|region| (region.body.as_str(), region.skip_rewrite))
         .collect::<Vec<_>>();
 
-    assert_eq!(
-        underlined_regions,
-        vec![("　　", true), ("ZX-9", false), ("　　　", true)]
-    );
+    assert_eq!(underlined_regions, vec![("　　ZX-9　　　", false)]);
 }
-

src-tauri/src/adapters/docx/tests/regression_misc.rs

@@ -129,7 +129,7 @@ fn extracts_hyperlink_display_text_with_target_presentation() {
 }
 
 #[test]
-fn locks_bare_urls_inside_plain_docx_runs() {
+fn keeps_bare_urls_inside_plain_docx_runs_editable() {
     let xml = r#"<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
 <w:document xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main">
   <w:body>
@@ -150,16 +150,15 @@ fn locks_bare_urls_inside_plain_docx_runs() {
 
     assert_eq!(
         parts,
-        vec![
-            ("访问 ", false),
-            ("https://chat.deepseek.com/share/lzlvnjcj3o5uees841", true),
-            (" 查看答案", false),
-        ]
+        vec![(
+            "访问 https://chat.deepseek.com/share/lzlvnjcj3o5uees841 查看答案",
+            false
+        )]
     );
 }
 
 #[test]
-fn keeps_url_with_trailing_space_as_one_locked_region() {
+fn keeps_url_with_trailing_space_as_one_editable_region() {
     let xml = r#"<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
 <w:document xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main">
   <w:body>
@@ -180,7 +179,7 @@ fn keeps_url_with_trailing_space_as_one_locked_region() {
 
     assert_eq!(
         parts,
-        vec![("https://chat.deepseek.com/share/lzlvnjcj3o5uees841 ", true)]
+        vec![("https://chat.deepseek.com/share/lzlvnjcj3o5uees841 ", false)]
     );
 }
 

src-tauri/src/adapters/markdown/inline_lines.rs

@@ -188,9 +188,6 @@ fn protected_markdown_region(text: &str) -> TextRegion {
         || text.starts_with('$')
         || text.starts_with("![")
         || text.starts_with('[')
-        || text.starts_with("http://")
-        || text.starts_with("https://")
-        || text.starts_with("www.")
     {
         return TextRegion::inline_object(text);
     }