ompchamber 2.0.0 → 2.0.1

package/src/shared/lib/code/line-endings.ts

@@ -0,0 +1,56 @@
+/**
+ * @license
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+/**
+ * Line-ending preservation for the file-editing surfaces.
+ *
+ * Two independent transports destroy the distinction, so the file's own ending
+ * has to be carried out of band rather than inferred from the buffer:
+ *
+ * - A `<textarea>` cannot hold a CRLF. The HTML spec defines its *API value* as
+ *   the raw value with every carriage return removed, so the instant a CRLF file
+ *   is placed in the editing surface the buffer is LF-only — verified in
+ *   Chromium: `textarea.value = 'a\r\nb'` reads back `'a\nb'`.
+ * - `multipart/form-data` normalizes the other way: every bare LF in a field
+ *   value becomes CRLF in transit (the HTML serializer on the way out, Bun's
+ *   parser agreeing on the way in). A save payload therefore cannot say "LF" no
+ *   matter what the buffer held.
+ *
+ * Saving the buffer as-is rewrote the whole file's line endings on a
+ * one-character edit — every line showed as changed in `git diff`. The ending is
+ * detected at read and travels as its own field; the server builds the bytes.
+ * An unedited save is then byte-identical and an edited save a one-line diff.
+ *
+ * Detection is by the FIRST terminator rather than a vote over all of them: a
+ * file is overwhelmingly uniform, and the first ending is the one the author's
+ * editor writes. A file with no terminator, or one using lone CR (classic Mac
+ * OS), reads as `lf`.
+ */
+
+export type LineEnding = 'lf' | 'crlf';
+
+/** The ending the file uses, judged by its first terminator. */
+export function detectLineEnding(text: string): LineEnding {
+  const index = text.indexOf('\n');
+  return index > 0 && text[index - 1] === '\r' ? 'crlf' : 'lf';
+}
+
+/**
+ * The buffer form: CRLF and lone CR collapsed to LF, matching what a textarea
+ * can actually hold. A string already free of CR is returned unchanged.
+ */
+export function toLf(text: string): string {
+  return text.includes('\r') ? text.replace(/\r\n?/g, '\n') : text;
+}
+
+/**
+ * The disk form: `text` normalized — whatever mix of endings it carries, and
+ * whether or not it came back through a transport that rewrote them — and then
+ * written with `ending`.
+ */
+export function toDiskText(text: string, ending: LineEnding): string {
+  const lf = toLf(text);
+  return ending === 'crlf' ? lf.replace(/\n/g, '\r\n') : lf;
+}