fix(updater): download via curl with resume instead of Chromium fetch

Chromium's network stack restarts requests mid-stream on flaky Wi-Fi,
yielding a right-sized but corrupt gzip (progress hits 100% three times,
then 'archive corrupt'). curl resumes with HTTP Range so a dropped
connection picks up where it stopped, retries internally, and the whole
file is still byte-count + gunzip verified before handoff. Falls back to
the fetch pipeline when curl is absent.
This commit is contained in:
avi 2026-10-05 00:25:03 -05:00
commit 9afe49c378
2 changed files with 83 additions and 30 deletions

4
package-lock.json generated
View file

@ -1,12 +1,12 @@
{
"name": "folio",
"version": "0.1.8",
"version": "0.1.10",
"lockfileVersion": 3,
"requires": true,
"packages": {
"": {
"name": "folio",
"version": "0.1.8",
"version": "0.1.10",
"license": "MIT",
"dependencies": {
"@tiptap/core": "^2.27.2",