From 6e2c9856b98faa2fc79ae12883ddf5acd1098f8e Mon Sep 17 00:00:00 2001 From: Christian Cleberg Date: Mon, 3 Aug 2026 15:41:22 -0500 Subject: fix malformed image URL and harden img rewrite Correct a https:/// typo in the tumblr post org source that produced a 404ing cross-origin image URL. Because the extra slash meant the string no longer matched, rewrite_img_urls() skipped it and shipped the broken URL. Make the rewrite pattern tolerate stray slashes so a source typo can't silently pass through un-rewritten again. --- build.py | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'build.py') diff --git a/build.py b/build.py index 28af6ec..771c406 100755 --- a/build.py +++ b/build.py @@ -326,13 +326,18 @@ def rewrite_img_urls(build_dir=".build"): Requires the server to serve /var/www/img/ at /img/ on the cleberg.net vhost (e.g. `ln -s /var/www/img /var/www/cleberg.net/img`). + + The pattern tolerates a stray number of slashes after the scheme + (https:/img, https:///img, ...) so a typo in the org source can't silently + slip through un-rewritten and ship a broken cross-origin URL. """ + pattern = re.compile(r"https:/+img\.cleberg\.net/") count = 0 for html in Path(build_dir).rglob("*.html"): text = html.read_text(encoding="utf-8") - new_text = text.replace("https://img.cleberg.net/", "/img/") - if new_text != text: - count += text.count("https://img.cleberg.net/") + new_text, n = pattern.subn("/img/", text) + if n: + count += n html.write_text(new_text, encoding="utf-8") print(f"Rewrote {count} img.cleberg.net references to /img/") -- cgit v1.2.3