Fix issues with echobot invite link in cmdeploy

The entire stack is setup to support 'localhost'  as a value for
ssh_host, returning LocalExec when that is the case, but this last step
made a new explicit SSHExec connection to config.mail_domain.

This changes that to simply use the subprocess module when ssh_host is
local.

This also fixes the issue where the connection was made to
'config.mail_domain' instead of the supplied ssh_host value, ensuring
that remains consistent.

Additionally, the entire process will be skipped if --dry-run is used
with cmdeploy, allowing a dry-run to complete without error.
This commit is contained in:
Alexandre Gauthier
2025-11-21 21:59:04 -05:00
committed by missytake
parent 166bf68915
commit 22c303ed7c

View File

@@ -109,15 +109,24 @@ def run_cmd(args, out):
try:
retcode = out.check_call(cmd, env=env)
if retcode == 0:
if not args.disable_mail:
if not args.disable_mail and not args.dry_run:
print("\nYou can try out the relay by talking to this echo bot: ")
sshexec = SSHExec(args.config.mail_domain, verbose=args.verbose)
print(
sshexec(
call=remote.rshell.shell,
kwargs=dict(command="cat /var/lib/echobot/invite-link.txt"),
echobot_cmd = "cat /var/lib/echobot/invite-link.txt"
if ssh_host in ["localhost", "@local", "@docker"]:
result = (
subprocess.check_output(echobot_cmd, shell=True)
.decode()
.strip()
)
print(result)
else:
echo_sshexec = get_sshexec(ssh_host, verbose=args.verbose)
print(
echo_sshexec(
call=remote.rshell.shell,
kwargs=dict(command=echobot_cmd),
)
)
)
out.green("Deploy completed, call `cmdeploy dns` next.")
elif not remote_data["acme_account_url"]:
out.red("Deploy completed but letsencrypt not configured")