Uwe Schuster
b3b2903c75
Replace substring `html.replace(old_src, new_src)` with a regex anchored to <script src="…"> / <link href="…"> attribute values. Inert occurrences in comments, JSON literals, or unrelated attributes are left alone. Loud warning (stderr) when zero matches occur — previously the script silently skipped a typo'd old_src. Also rewrites <link href> in the same pass so adjacent CSS hashing doesn't need a follow-up edit. Tests: tests/test_inject_hashed_filenames.py covers happy path (both quote styles, extra attributes), inert-substring cases (comment, JSON literal, data-attr, anchor href), and link-href rewriting. Closes #3 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
102 lines
3.3 KiB
Python
Executable file
102 lines
3.3 KiB
Python
Executable file
#!/usr/bin/env python3
|
|
"""Rewrite HTML script tags to point at Vite's hashed production bundle.
|
|
|
|
Usage:
|
|
inject-hashed-filenames.py CONFIG_JSON [BUILD_DIR]
|
|
|
|
CONFIG_JSON is a path to a JSON file listing the entries to rewrite. The
|
|
schema mirrors the two hard-coded entries fewo-webapp had baked into an
|
|
earlier version of this script:
|
|
|
|
[
|
|
{
|
|
"manifest": "static/dist/.vite/manifest.json",
|
|
"html": "static/index.html",
|
|
"old_src": "/static/dist/app.js"
|
|
},
|
|
{
|
|
"manifest": "static/guest/dist/.vite/manifest.json",
|
|
"html": "static/guest/index.html",
|
|
"old_src": "/guest/dist/guest-app.js"
|
|
}
|
|
]
|
|
|
|
All paths are resolved relative to BUILD_DIR (defaults to the repo root
|
|
containing the config file's parent chain → `$PWD`).
|
|
"""
|
|
import json
|
|
import os
|
|
import re
|
|
import sys
|
|
|
|
|
|
def _build_pattern(old_src: str) -> "re.Pattern[str]":
|
|
# Match `src` on <script> and `href` on <link>, single- or double-quoted.
|
|
# We anchor to the tag name so an `old_src` substring sitting inside an
|
|
# HTML comment, a JSON literal, or a `data-…` attribute is not rewritten.
|
|
return re.compile(
|
|
r'(<(?:script|link)\b[^>]*?\b(?:src|href)\s*=\s*["\'])'
|
|
+ re.escape(old_src)
|
|
+ r'(["\'])',
|
|
re.IGNORECASE,
|
|
)
|
|
|
|
|
|
def rewrite(html: str, old_src: str, new_src: str) -> "tuple[str, int]":
|
|
"""Return (new_html, count). Tag-aware: only rewrites <script src> /
|
|
<link href> attributes, never substring matches in comments or JSON."""
|
|
pattern = _build_pattern(old_src)
|
|
return pattern.subn(lambda m: m.group(1) + new_src + m.group(2), html)
|
|
|
|
|
|
def inject(manifest_path: str, html_path: str, old_src: str) -> None:
|
|
if not os.path.exists(manifest_path):
|
|
print(f"skip: no manifest at {manifest_path}")
|
|
return
|
|
if not os.path.exists(html_path):
|
|
print(f"skip: no html at {html_path}")
|
|
return
|
|
with open(manifest_path) as f:
|
|
manifest = json.load(f)
|
|
for entry in manifest.values():
|
|
if not entry.get("isEntry"):
|
|
continue
|
|
hashed = entry["file"]
|
|
new_src = f"{os.path.dirname(old_src)}/{hashed}"
|
|
with open(html_path) as f:
|
|
html = f.read()
|
|
new_html, count = rewrite(html, old_src, new_src)
|
|
if count == 0:
|
|
# Loud warning — silent skip used to mask typos in `old_src`.
|
|
print(
|
|
f"WARN: no <script src> or <link href> matching {old_src!r} "
|
|
f"in {html_path} — leaving file unchanged",
|
|
file=sys.stderr,
|
|
)
|
|
return
|
|
with open(html_path, "w") as f:
|
|
f.write(new_html)
|
|
print(f"{old_src} -> {new_src} ({count} occurrence{'s' if count != 1 else ''})")
|
|
return
|
|
print(f"skip: no isEntry row in {manifest_path}")
|
|
|
|
|
|
def main() -> int:
|
|
if len(sys.argv) < 2:
|
|
print("usage: inject-hashed-filenames.py CONFIG_JSON [BUILD_DIR]", file=sys.stderr)
|
|
return 2
|
|
cfg_path = sys.argv[1]
|
|
build_dir = sys.argv[2] if len(sys.argv) > 2 else os.getcwd()
|
|
with open(cfg_path) as f:
|
|
entries = json.load(f)
|
|
for e in entries:
|
|
inject(
|
|
os.path.join(build_dir, e["manifest"]),
|
|
os.path.join(build_dir, e["html"]),
|
|
e["old_src"],
|
|
)
|
|
return 0
|
|
|
|
|
|
if __name__ == "__main__":
|
|
sys.exit(main())
|