Twenty-nine commits since v0.12.0. See CHANGELOG.md for the full list and the upgrade notes. Ten additions — the SEO package (site URL, robots.txt, sitemap.xml, JSON-LD), slideshow mode, opt-in original downloads, copy-link, previous/next album navigation, alt text from the Immich description, per-album location precision, the update notice, a Help tab, and a subpage cover grid separate from its photo grid. Two fixes: CJK album slugs, and the lightbox control bar. Photo of the Day (#476) was built on `dev` and reverted as not planned; it never appears in a release. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01DmsHVRL9yQvckUB1ZpVWre --------- Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: ralksta <ralksta@users.noreply.github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
70 lines
2.5 KiB
TypeScript
70 lines
2.5 KiB
TypeScript
/**
|
|
* Filenames for downloaded originals (#475).
|
|
*
|
|
* The name comes from Immich, which took it from the camera or from whatever
|
|
* the photographer typed - so it is untrusted input on its way into a
|
|
* `Content-Disposition` header. A newline in it would end the header and let
|
|
* the rest be read as headers of its own; a path separator would suggest a
|
|
* directory to the browser.
|
|
*/
|
|
|
|
/**
|
|
* Drop control characters, including the CR and LF that would end a header.
|
|
*
|
|
* Written as a code-point filter rather than a regular expression: a character
|
|
* class of escaped control codes is the kind of source line that gets mangled
|
|
* by an editor or a copy-paste and then silently stops matching.
|
|
*/
|
|
function stripControlChars(value: string): string {
|
|
return Array.from(value)
|
|
.filter((character) => {
|
|
const code = character.charCodeAt(0);
|
|
return code > 31 && code !== 127;
|
|
})
|
|
.join('');
|
|
}
|
|
|
|
/** True for printable ASCII, the range the pre-RFC-5987 filename is limited to. */
|
|
function isPrintableAscii(character: string): boolean {
|
|
const code = character.charCodeAt(0);
|
|
return code >= 32 && code <= 126;
|
|
}
|
|
|
|
/**
|
|
* A filename safe to put in a `Content-Disposition` header.
|
|
*
|
|
* Strips anything that could break out of the header or out of the download
|
|
* directory, then falls back to a generic name if nothing usable is left - an
|
|
* empty filename would make the browser invent one from the URL, which is the
|
|
* opaque token.
|
|
*/
|
|
export function safeDownloadName(raw: string | undefined, fallback = 'photo'): string {
|
|
const cleaned = stripControlChars(raw ?? '')
|
|
// Directory separators and traversal.
|
|
.replace(/[/\\]/g, '_')
|
|
// Quotes and semicolons delimit the header's own parameters.
|
|
.replace(/["';]/g, '')
|
|
.replace(/^\.+/, '')
|
|
.trim();
|
|
|
|
if (!cleaned) return fallback;
|
|
// Long names are the browser's problem, but an unbounded one is ours.
|
|
return cleaned.slice(0, 200);
|
|
}
|
|
|
|
/**
|
|
* A full `Content-Disposition` value.
|
|
*
|
|
* Emits both `filename` and `filename*`: the plain parameter is stripped to
|
|
* ASCII for clients that predate RFC 5987, and the starred one carries the
|
|
* real name percent-encoded for everything since.
|
|
*/
|
|
export function contentDisposition(name: string | undefined): string {
|
|
const safe = safeDownloadName(name);
|
|
const ascii =
|
|
Array.from(safe)
|
|
.map((character) => (isPrintableAscii(character) ? character : '_'))
|
|
.join('') || 'photo';
|
|
return `attachment; filename="${ascii}"; filename*=UTF-8''${encodeURIComponent(safe)}`;
|
|
}
|