On 16/03/17 05:07 AM, Yamaban wrote:
Think of bigger environs, a office with more than say 4 desktops, or a classroom, esp. in conjunction with "hostname-via-dhcp", and service done mostly remote, or even via puppet or similar tools, or even full-blown pxe / nfs setups.
I've done similar setups during my time in university / student-admin during the late 1990ies, most of the time we used xdm, which made that task easy.
Well, yes. I even remember SUN workstations back in the 3meg Ethernet days that were 'roaming login' or whatever the term that Microsoft has made popular. The workstation was minimal and /usr (as the $HOME was termed back then) was loaded via NFS. I forget what else was loaded, but IIR the RootFS had lots of symlinks that became live when the login completed. I've also worked at places that supplied me with a laptop that had enough for me to do basic work on local files but also had that 'roaming login' to any part of the (multi-site) network. [As a sidebar, I'll mention that they ran SAMBA on a pair of HP-500 HP/UX series machines.] What is significant about that is that when I logged in the remote host not only handed out local shares and my network workspace and shared spaces, but also rammed some executables down. [That was fine so long as I was running the "Corporate XP". It was when I tried logging in with my own Linux laptop running a slightly alter model version of SAMBA that the problems arose :-(] I'm aware that the DHCP protocol and deal with hostnames in a number of ways. Either my "personalized" workstation can tell the DHCP server my hostname and it perhaps it responds in a custom fashion. Alternatively the server gives me a hostname for the duration of the session. In either case the hostname should be available before the login screen appears. Which gets back to the OP question of why don't the other login screen mechanisms, themes allow for the text to be easily modified. Rather than being simply declarative (as was the case with kdmrc) is now seems to be procedural, using QtQucik2. Yes, this means that it can be compiled with language packages, so we get lines like errorMessage.text = textConstants.loginFailed What was that about adding a level of indirection? And so too in, for example, /usr/share/sddm/themes/maldives/ Main.qml we find text: textConstants.welcomeText.arg(sddm.hostName) There is similar in the code for "breeze", but it is a procedurally more complex and obscure instance. All this is, I suppose, logical, a logical progression. An emergent property of internationalization. But it is also written in QtQuick2, so it means another language to learn, with its specific quirks. Why not something like Ruby, which I enjoyed learning and which we already have as the basis for Yast? Its already an interpreter. The QtQuick script reminds me of Ruby in many ways. I suppose the answer is that Ruby isn't a particularly fast interpreter because its a OO language. I think that's a weak answer. Whenever you are dealing with user input then the user typing speed is the limiting factor. There does seem to have been a trend among developers recently to move away from systems and details that are for the convenience of end users and administrators. The absence of the arrow icons at the end of scroll bars is one example of that which has already come up in other threads. Sorry to be so rambling. I'll go put a pot of coffee on. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org