DietPi-Software | frp #7292
DietPi-Software | frp #7292
Conversation
|
Shall we do some kind of migration from ini to toml? |
|
thougt about it, but we would need to check whole ini file for custom changes. Isn't it? |
|
Yeah, and the number of config keys is much too large to do all of this:
Sadly there is no internal conversion feature. So it looks like this is something we need to ask users to do manually. Best we can probably do is, checking whether the ini files exist, and if so, prompt an information, and rename them to *.bak (to not repeat the prompt). Btw there is a way to generate bash completion files: frpc completion bash
frps completion bashProbably something for our docs? |
- DietPi-Software | frp: Fix parsing client input, allow any characters for the token (as we do not know any limitations), and make use of the native G_WHIP_INPUTBOX input validation. Also expose the server bind address in our default config, which might be more commonly changed to limit access to particular networks.
- DietPi-Software | frp: Inform about new toml config file and leave old ini as backup in place. Do not offer client inputs, if the config exists already (and is hence not generated anyway). Allow an empty token input, which is generally possible. Allow unattended installs, using both server and client as default, and applying default values to all inputs and menus. - DietPi-Globals | G_WHIP: Return default value for inputbox, checkbox and menu, if not interactive, but keep returning error code, so that it can be handled in parent script. Minor coding enhancements. - CI | DietPi-Software: Enable frp checks now that it can be installed unattended
There was a problem hiding this comment.
Works well now. An unrelated issue is that the frp server can sometimes take longer to start than the frp client, and hence the client then fails to connect. It restarts up to 3 times, but during GitHub Actions test, for ARMv7 emulation, even that sometimes was not sufficient, but it failed ultimately. But I have no idea how to fix this properly. The daemon would need to return some systemd notify call to lock the start sequence until a defined state, like when the TCP listener is up. But as they removed their systemd examples a while ago, I don't think this has any priority.
|
Yes have noticed this as well. At least until know we don't have a report this being an issue. Should be fine i guess |
|
Yeah, and if needed, we can raise |
initotomlconfiguration file becauseiniformat is deprecated and the support will be removed in the future