Arch Linux

Summary and Info: Pacman syncs gibberish data if behind a captive portal, overwriting your local repo cache.


Steps to Reproduce:
1) Be behind a captive portal (e.g. hotel WiFi without logging on)
2) pacman -Syyu:

(downloads of repo data seem to complete successfully), then:

[..]
error: could not open file /var/lib/pacman/sync/extra.db: Unrecognized archive format
error: could not open file /var/lib/pacman/sync/extra.db: Unrecognized archive format
error: could not open file /var/lib/pacman/sync/community-testing.db: Unrecognized archive format
error: could not open file /var/lib/pacman/sync/community-testing.db: Unrecognized archive format
error: could not open file /var/lib/pacman/sync/community.db: Unrecognized archive format
error: could not open file /var/lib/pacman/sync/community.db: Unrecognized archive format
[..]
there is nothing to do

Now I have botched my local repo data?

Expected results:

1) pacman verifies that the repo file is sane (not HTML as it turned out to be in my case when behind the portal, but an actual repo file (.tar.gz?))

2) If the file is not sane, do not replace the local copy with the downloaded garbage. In theory this could lead to some repos being updated locally and some not. That should be fine as the user could have any amount and kinds of repos enabled.