flathub-infra
Support rewriting scraped URLs
Anydesk is checked by scraping their human-readable download page. This links to a URL of the form https://download.anydesk.com/linux/anydesk-$VERSION-$ARCH.tar.gz.
When a new version is released, those files disappear, rendering the Flatpak non-installable. However, older (and current) versions remain available at https://download.anydesk.com/linux/generic-linux/anydesk-$VERSION-$ARCH.tar.gz.
We could teach htmlchecker to be able to mangle URLs. What I had in mind was something like:
"x-checker-data": {
"type": "html",
"url": "https://anydesk.com/en/downloads/linux",
"url-pattern": "(https://download.anydesk.com/linux/anydesk-([0-9.]+)-amd64.tar.gz)",
"version-pattern": "https://download.anydesk.com/linux/anydesk-([0-9.]+)-amd64.tar.gz",
"url-substitution": ["/linux/", "/linux/generic-linux/"]
}
and the checker would just use str.replace(url_substitution[0], url_substitution[1]).
An alternative that would work in this case would be to instead scrape the index at https://download.anydesk.com/linux/, find all matches of the url-pattern, and take the one with the larger version number.
Are there any updates on this issue regarding its implementation? This feature would be really useful even in a slimmed down variant that checks only a directory of an ftp server for a newer version.
