When a download attempt fails with an HTTP error, the body of the failed
request will be written to the output file with a 0 exit status, causing the
execution of an error message as a script.
This issue is resolved by adding the --fail option to curl.
Co-authored-by: David Runge <dvzrv@archlinux.org>
1) wait for network-online.target before invoking curl
as there's no synchronization with network setup for this script
2) don't hide curl errors - it may be easier to debug the issues
3) add log and comments
The curl --retry-connrefused option is used with not instead of the --retry <num> option to add an extra type of failure to retry on, without --retry <num> it does not retry at all even on a connection refused.
https://man.archlinux.org/man/curl.1.en
configs/releng/airootfs/root/.automated_script.sh:
Add the `--location` curl parameter (see `man 1 curl`) to allow for curl to retrieve a remote script even if the source
is being redirected (e.g. moved permanently) when using the `script=` kernel commandline parameter.
Fixes#113
configs/releng/airootfs/root/.automated_script.sh:
Replacing wget with curl for the use-case of downloading a
pre-specified remote startup script (changing flags where needed).
configs/releng/packages.x86_64:
Removing wget from the list of installed packages, as
.automated_script.sh was the only reason for it being installed.
Closes#16
