CHANGELOG.md

@@ -29,10 +29,6 @@ Versionsnummern nach [Semantic Versioning](https://semver.org/lang/de/).
 
 ### Behoben
 
-- CSS-Auswertung: Es zählen nur noch Stylesheets, die ein Dokument per `<link>` (oder `@import`) einbindet.
-  Bisher wirkten alle CSS-Dateien des Manifests, auch mitgelieferte Layouts anderer Buchreihen. Im C-Buch
-  machte das aus fetten Buchstaben Inline-Code (`` `st`an`d`ard ``) und setzte falsche `<mark>`; eine
-  fremde `display:none`-Regel hätte Text gelöscht (QA3-01, Regressionstests in `tests/test_styles.py`).
 - `pandoc/callouts.lua`: Die Kästen blieben ohne Rahmen, wenn man Pandoc zusätzlich `-H`/`--include-in-header`
   übergab, und im EPUB hatten nur die Titelseite, nicht die Kapitel, das Kasten-CSS. Das CSS steht jetzt in
   `pandoc/callouts.css`; der Filter setzt es bei HTML an den Anfang des Dokuments und fügt es bei EPUB den

README.md

@@ -249,8 +249,6 @@ weil CommonMark keine `[^…]`-Syntax kennt.
 - **CSS-Formatierung:** Viele Verlage setzen Hervorhebungen nur über Stylesheets. epub2md wertet die
   CSS-Dateien des Buches aus: Klassen-, Element- und ID-Selektoren, Nachfahren- und Kind-Kombinatoren,
   Spezifität und `style`-Attribute. `@media`-Blöcke und Selektoren mit Pseudoklassen werden ignoriert.
-  Berücksichtigt werden nur Stylesheets, die ein Dokument per `<link>` (oder `@import`) einbindet;
-  mitgelieferte Layouts anderer Buchreihen bleiben wie im E-Book-Reader wirkungslos (Info `css-unlinked`).
   Direkt aneinanderstoßende gleiche Hervorhebungen werden zusammengefasst (`*x**y*` → `*xy*`).
 - **Farben und Schriften:** Markdown kennt keine Farben, Schriftarten oder -größen. Mit `--styles html`
   bleiben sie als HTML-Stil erhalten. Übernommen wird nur, was vom Grundstil abweicht (häufigste Werte aller

epub2md/html_converter.py

@@ -1233,8 +1233,8 @@ class Renderer:
             finally:
                 self.strong -= 1
         if name in ("del", "s", "strike"):
-            del_open, del_close = ("~~", "~~") if self.opt.format == "gfm" else ("<del>", "</del>")
+            marker = ("~~", "~~") if self.opt.format == "gfm" else ("<del>", "</del>")
-            return pre + self._wrap(self.inline_children(el), del_open, del_close)
+            return pre + self._wrap(self.inline_children(el), marker[0], marker[1])
         if name in ("u", "ins", "mark", "kbd", "sup", "sub"):
             if name in ("sup", "sub"):
                 elems = [c for c in el if is_element(c)]

epub2md/pipeline.py

@@ -13,7 +13,6 @@ from dataclasses import asdict, dataclass, field
 from datetime import datetime, timezone
 from pathlib import Path
 from typing import Any
-from urllib.parse import unquote
 
 from . import TOOL_NAME, __version__
 from .chapters import ChapterBuilder
@@ -28,7 +27,7 @@ from .media import MediaRegistry
 from .models import Chapter
 from .normalize import Doc, book_uses_title_tables, is_html_toc_doc, make_doc, normalize_doc, remove_keep_tail
 from .references import LinkResolver
-from .utils import is_element, is_inside, iter_elements, local, resolve_href, sha256_file, slugify, text_content
+from .utils import is_element, is_inside, iter_elements, local, sha256_file, slugify, text_content
 from . import writer
 
 MARKER_FILES = ("conversion-report.json", "manifest.json")
@@ -234,39 +233,18 @@ def _build_structure(docs: list[Doc], package: Any, opts: Options,
     return notes, chapters, resolver
 
 
-CSS_IMPORT = re.compile(r"""@import\s+(?:url\(\s*)?["']?([^"')\s;]+)""", re.I)
-
-
 def _load_stylesheet(reader: EpubReader, package: Any, docs: list[Doc], diag: Diagnostics) -> StyleSheet:
-    """The book's CSS: manifest stylesheets that a spine document links (directly or via @import), in manifest
+    """All CSS of the book: manifest stylesheets (manifest order) plus <style> elements of the documents."""
-    order, plus <style> elements of the documents. Unlinked stylesheets (publishers ship the layouts of other
+    texts: list[str] = []
-    series) are ignored – a reading system would not apply them either."""
+    for item in package.manifest.values():
-    pending: list[str] = []
+        if item.media_type == "text/css" or item.path.lower().endswith(".css"):
+            try:
+                texts.append(decode_markup(reader.read(item.path), item.path, diag))
+            except (KeyError, ContainerError):
+                diag.warn("css-unreadable", f"Stylesheet {item.path} nicht lesbar; seine Formatierung entfällt.", item.path)
     for doc in docs:
-        for e in iter_elements(doc.root):
+        texts.extend(text_content(e) for e in iter_elements(doc.root) if local(e.tag) == "style")
-            if local(e.tag) == "link" and "stylesheet" in (e.get("rel") or "").lower().split():
+    return StyleSheet(texts)
-                pending.append(resolve_href(doc.path, unquote((e.get("href") or "").split("#")[0].split("?")[0])) or "")
-    texts: dict[str, str] = {}
-    while pending:
-        path = pending.pop()
-        if not path or path in texts:
-            continue
-        try:
-            texts[path] = decode_markup(reader.read(path), path, diag)
-        except (KeyError, ContainerError):
-            diag.warn("css-unreadable", f"Stylesheet {path} nicht lesbar; seine Formatierung entfällt.", path)
-            texts[path] = ""
-            continue
-        pending.extend(resolve_href(path, unquote(m)) or "" for m in CSS_IMPORT.findall(texts[path]))
-    css_items = [i.path for i in package.manifest.values() if i.media_type == "text/css" or i.path.lower().endswith(".css")]
-    ordered = [texts[p] for p in css_items if p in texts] + [t for p, t in texts.items() if p not in css_items]
-    unlinked = len([p for p in css_items if p not in texts])
-    if unlinked:
-        diag.info("css-unlinked", f"{unlinked} Stylesheet(s) im Manifest werden von keinem Dokument eingebunden "
-                                  "und daher nicht ausgewertet.")
-    for doc in docs:
-        ordered.extend(text_content(e) for e in iter_elements(doc.root) if local(e.tag) == "style")
-    return StyleSheet(ordered)
 
 
 def _render(reader: EpubReader, package: Any, chapters: list[Chapter], resolver: LinkResolver, notes: Any,

tests/conftest.py

@@ -3,7 +3,6 @@
 from __future__ import annotations
 
 import json
-import posixpath
 import sys
 import zipfile
 from dataclasses import dataclass, field
@@ -26,7 +25,7 @@ SVG = b'<svg xmlns="http://www.w3.org/2000/svg" width="10" height="10"><rect wid
 XHTML = """<?xml version="1.0" encoding="utf-8"?>
 <!DOCTYPE html>
 <html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops" lang="de">
-<head><title>{title}</title>{head}</head>
+<head><title>{title}</title></head>
 <body>{body}</body>
 </html>"""
 
@@ -37,7 +36,6 @@ class Doc:
     body: str
     title: str = ""
     linear: bool = True
-    css: list[str] | None = None  # linked stylesheets (OEBPS-relative) – default: every .css in Book.media
 
 
 @dataclass
@@ -92,7 +90,7 @@ def build_epub(path: Path, book: Book) -> Path:
   <spine{' toc="ncx"' if book.ncx else ''}>{''.join(spine)}{book.spine_extra}</spine>
 </package>"""
     nav_xhtml = XHTML.format(
-        title="Inhalt", head="",
+        title="Inhalt",
         body='<nav epub:type="toc"><ol>' + "".join(f'<li><a href="{h}">{lbl}</a></li>' for lbl, h in nav_entries) + "</ol></nav>",
     )
     ncx = (
@@ -116,10 +114,7 @@ def build_epub(path: Path, book: Book) -> Path:
         if book.ncx:
             z.writestr("OEBPS/toc.ncx", ncx)
         for d in book.docs:
-            css = d.css if d.css is not None else [n for n in book.media if n.lower().endswith(".css")]
+            z.writestr(f"OEBPS/{d.name}", XHTML.format(title=d.title, body=d.body))
-            head = "".join(f'<link rel="stylesheet" type="text/css" href="{posixpath.relpath(c, posixpath.dirname(d.name) or ".")}"/>'
-                           for c in css)
-            z.writestr(f"OEBPS/{d.name}", XHTML.format(title=d.title, head=head, body=d.body))
         for name, data in book.media.items():
             z.writestr(f"OEBPS/{name}", data)
         for name, data in book.extra_entries.items():

tests/test_styles.py

@@ -211,22 +211,3 @@ def test_page_marker_in_box_title_goes_before_the_box(convert):
     text = convert(Book(docs=[Doc("c.xhtml", body)]), "--page-markers", "visible").chapter("001")
     assert "\\[S. 62\\]\n\n> [!NOTE]\n> **Wozu Deklarationen?**" in text
     assert "> Weil es \\[S. 63\\]hilft." in text  # markers in the box body stay in place
-
-
-def test_qa3_01_unlinked_stylesheets_are_ignored(convert):
-    """Publishers ship layout CSS of other series in the manifest; only linked stylesheets apply (QA3-01)."""
-    css = {"main.css": b".bold { font-weight: bold; }",
-           "fremd.css": b"p > span.bold { font-weight: normal; font-family: monospace; } .weg { display: none; }"}
-    body = ('<h1>K</h1><p>Die (<span class="bold">st</span>an<span class="bold">d</span>ard '
-            '<span class="bold">in</span>put) und <span class="weg">sichtbarer Text</span>.</p>')
-    text = convert(Book(docs=[Doc("c.xhtml", body, css=["main.css"])], media=css)).chapter("001")
-    assert "`" not in text  # monospace rule of fremd.css must not turn bold into code
-    assert "**st**an**d**ard **in**put" in text
-    assert "sichtbarer Text" in text  # display:none of fremd.css must not hide content
-
-
-def test_qa3_01_imported_stylesheets_still_apply(convert):
-    css = {"main.css": b'@import url("extra.css");', "extra.css": b".k { font-style: italic; }"}
-    body = '<h1>K</h1><p><span class="k">kursiv</span> normal.</p>'
-    text = convert(Book(docs=[Doc("c.xhtml", body, css=["main.css"])], media=css)).chapter("001")
-    assert "*kursiv* normal." in text