mawei/StarPilot
1
0
Fork 0
mirror of https://github.com/firestar5683/StarPilot.git synced 2026-06-29 10:32:10 +08:00
Code Issues Packages Projects Releases Wiki Activity

raylib: precompile regex patterns for faster HTML parsing (#36417)

Browse Source 
precompiled regex
This commit is contained in:
Dean Lee
2025-10-23 07:18:11 +08:00
committed by GitHub
parent c33c9ff22a
commit 215acefbb4
1 changed files with 8 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
system/ui/widgets/html_render.py
+8 -4
View File
@@ -31,6 +31,10 @@ class ElementType(Enum):
TAG_NAMES = '|'.join([t.value for t in ElementType])
START_TAG_RE = re.compile(f'<({TAG_NAMES})>')
END_TAG_RE = re.compile(f'</({TAG_NAMES})>')
COMMENT_RE = re.compile(r'<!--.*?-->', flags=re.DOTALL)
DOCTYPE_RE = re.compile(r'<!DOCTYPE[^>]*>')
HTML_BODY_TAGS_RE = re.compile(r'</?(?:html|head|body)[^>]*>')
TOKEN_RE = re.compile(r'</[^>]+>|<[^>]+>|[^<\s]+')
def is_tag(token: str) -> tuple[bool, bool, ElementType | None]:
@@ -104,14 +108,14 @@ class HtmlRenderer(Widget):
self._cached_width = -1
# Remove HTML comments
html_content = re.sub(r'<!--.*?-->', '', html_content, flags=re.DOTALL)
html_content = COMMENT_RE.sub('', html_content)
# Remove DOCTYPE, html, head, body tags but keep their content
html_content = re.sub(r'<!DOCTYPE[^>]*>', '', html_content)
html_content = re.sub(r'</?(?:html|head|body)[^>]*>', '', html_content)
html_content = DOCTYPE_RE.sub('', html_content)
html_content = HTML_BODY_TAGS_RE.sub('', html_content)
# Parse HTML
tokens = re.findall(r'</[^>]+>|<[^>]+>|[^<\s]+', html_content)
tokens = TOKEN_RE.findall(html_content)
def close_tag():
nonlocal current_content