Jim Cheetham

Right now it's used to provide authentication to a family tree app that doesn't have a robust auth layer of its own; I don't want to manage new user accounts...

I think the correct patch should not be to just remove special treatment for REMOTE_VERSION = "dropbear*", but to instead detect the presence of OpenWRT through some other way, and...

I set up an OpenWrt in a VM, made some changes and I've sent him a patch that might help. I was using the presence of `/etc/openwrt_release` but `/etc/os-release` might...

Unsurprisingly that looks almost exactly like the patch I sent upstream :-) but I left the "AUTH_KEY_DIR=$(dirname ..." call in the outer loop, because I wasn't sure it was available...

@MichaIng how do you propose to check for the presence of `/etc/openwrt_release`, or otherwise to differentiate between OpenWrt and other Dropbear systems, while in the outer REMOTE_VERSION check for `dropbear*`?...

For future ease of auditing downstream inclusion of the fix to this problem, can you please identify here on this PR, the commits or version numbering made that address this...

Peertube addresses that difference by including two Contact lines - one pointing to the project's documentation for reporting, and the other with the email of the server admin.

Using color alone isn't a wonderful differentiator, for accessibility reasons. Perhaps *every* glyph/character should have an indicator or some sort, identifying its type (upper/lower case, number, non-"ascii") ... not necessarily...

Absolutely not my intention to make you feel bad - color isn't a bad indicator! Just not on its own, without other elements :-) You could argue in this case...