From 5be966c2ed6d1f1692dbb308557373d67216a7e7 Mon Sep 17 00:00:00 2001
From: RibatTRW <aydinhrrs@gmail.com>
Date: Sun, 27 Sep 2026 14:37:41 +0300
Subject: [PATCH 1/2] fix(lightnovelwp): parse chapter text when bottomnav is
 absent
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

parseChapter anchored the chapter text on a `bottomnav` div to close the
`epcontent` block. Current versions of the theme no longer emit that wrapper,
so the match returned undefined and every chapter parsed as an empty string —
no error, just blank text on affected sources.

Close the block at whichever structural landmark follows it (bottomnav, a
script tag, or </body>), and fall back to the whole document when no epcontent
block is present at all. This also drops the regex's hard-coded `epcontent `
class spelling, which would not match `class="epcontent entry-content"` when
epcontent is the only class.

Output is byte-identical to the old expression on pages that still have
bottomnav, verified against captured markup, so the other lightnovelwp
sources are unaffected.

Kol Novel was the source that surfaced this.

Co-Authored-By: Claude Code <noreply@anthropic.com>
---
 plugins/multisrc/lightnovelwp/template.ts | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/plugins/multisrc/lightnovelwp/template.ts b/plugins/multisrc/lightnovelwp/template.ts
index 8234b5cf4..cb1ef7676 100644
--- a/plugins/multisrc/lightnovelwp/template.ts
+++ b/plugins/multisrc/lightnovelwp/template.ts
@@ -438,11 +438,16 @@ export class LightNovelWPPlugin implements Plugin.PluginBase {
         throw error;
       }
     }
+    // The chapter text lives in the `epcontent` block. Older child themes
+    // closed it with a `bottomnav` div, but that wrapper is gone on current
+    // ones, so anchoring on it silently returned an empty chapter. Cut the
+    // block off at the next structural landmark instead, whichever is
+    // present, and fall back to the whole document when none of them is.
+    const content = data.match(
+      /<div[^>]*class="[^"]*\bepcontent\b[^"]*"[^>]*>([^]*?)(?=<div[^>]*class="?bottomnav|<script|<\/body)/,
+    );
     return (
-      data
-        .match(/<div.*?class="epcontent ([^]*?)<div.*?class="?bottomnav/g)?.[0]
-        .match(/<p[^>]*>([^]*?)<\/p>/g)
-        ?.join('\n') || ''
+      (content?.[1] ?? data).match(/<p[^>]*>([^]*?)<\/p>/g)?.join('\n') || ''
     );
   }
 

From d89b9e2a2a17d934b502b47b613d0dde4297ce77 Mon Sep 17 00:00:00 2001
From: RibatTRW <aydinhrrs@gmail.com>
Date: Sun, 27 Sep 2026 18:36:24 +0300
Subject: [PATCH 2/2] fix(lightnovelwp): bound chapter extraction to the
 epcontent element
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Follow-up to the bottomnav fix. Cutting the document at the next `<script>`
or `</body>` guessed the boundary from surrounding text, which is wrong in
both directions on a real page:

- an inline `<script>` inside the block ends the match early, so paragraphs
  after it are dropped — measured on Kol Novel, the first script after the
  block cuts 6 of 96 paragraphs
- when no script intervenes the match runs past the block's real close, so
  paragraphs from recommendations and comments are appended to the chapter —
  the same page yields 145 paragraphs that way against 96 inside the block

Walk the markup with the parser instead and close on the element that actually
ends the block, tracking div depth. This drops the first and last false
landmark in one change rather than picking a better guess.

Also return nothing when there is no epcontent block at all. A 200 response
from a login wall or an error page has paragraphs but no chapter, and
falling back to the whole document showed that page's text as the chapter.

Verified against captured Kol Novel markup: 96 paragraphs inside the block, and
an error page now yields an empty chapter rather than its notice text.

Co-Authored-By: Claude Code <noreply@anthropic.com>
---
 plugins/multisrc/lightnovelwp/template.ts | 66 +++++++++++++++++++----
 1 file changed, 56 insertions(+), 10 deletions(-)

diff --git a/plugins/multisrc/lightnovelwp/template.ts b/plugins/multisrc/lightnovelwp/template.ts
index cb1ef7676..080b4a452 100644
--- a/plugins/multisrc/lightnovelwp/template.ts
+++ b/plugins/multisrc/lightnovelwp/template.ts
@@ -440,17 +440,63 @@ export class LightNovelWPPlugin implements Plugin.PluginBase {
     }
     // The chapter text lives in the `epcontent` block. Older child themes
     // closed it with a `bottomnav` div, but that wrapper is gone on current
-    // ones, so anchoring on it silently returned an empty chapter. Cut the
-    // block off at the next structural landmark instead, whichever is
-    // present, and fall back to the whole document when none of them is.
-    const content = data.match(
-      /<div[^>]*class="[^"]*\bepcontent\b[^"]*"[^>]*>([^]*?)(?=<div[^>]*class="?bottomnav|<script|<\/body)/,
-    );
-    return (
-      (content?.[1] ?? data).match(/<p[^>]*>([^]*?)<\/p>/g)?.join('\n') || ''
-    );
-  }
+    // ones, so anchoring on it silently returned an empty chapter.
+    //
+    // Walk the markup with the parser rather than cutting the document with a
+    // regex: the block has to end at the element that closes it, and any
+    // landmark guessed from the surrounding text is wrong on some site. A
+    // <script> inside the block is prose-adjacent, not the end of it, so
+    // treating the first one as a boundary truncates the chapter.
+    let depth = 0;
+    let inside = false;
+    let found = false;
+    let isReadingParagraph = false;
+    let paragraph = '';
+    const paragraphs: string[] = [];
 
+    const parser = new Parser({
+      onopentag(name, attribs) {
+        if (name === 'div' && !inside) {
+          const className = attribs['class'] || '';
+          if (/(^|\s)epcontent(\s|$)/.test(className)) {
+            inside = true;
+            found = true;
+            depth = 1;
+          }
+          return;
+        }
+        if (!inside) return;
+        if (name === 'div') depth++;
+        else if (name === 'p') isReadingParagraph = true;
+      },
+      ontext(text) {
+        if (inside && isReadingParagraph && text.trim()) {
+          paragraph += text;
+        }
+      },
+      onclosetag(name) {
+        if (!inside) return;
+        if (name === 'p') {
+          if (paragraph.trim()) paragraphs.push(paragraph.trim());
+          paragraph = '';
+          isReadingParagraph = false;
+        } else if (name === 'div') {
+          depth--;
+          if (depth === 0) inside = false;
+        }
+      },
+    });
+
+    parser.write(data);
+    parser.end();
+
+    // No epcontent block means this was not a chapter page (a login wall or
+    // an error page served with 200). Returning the whole document here would
+    // show that page's text as the chapter, so report nothing instead.
+    if (!found) return '';
+
+    return paragraphs.join('\n');
+  }
   async searchNovels(
     searchTerm: string,
     page: number,