Re the clock: The clock updates to be correct instantly on connecting to wifi. So it’s not a slow “tactful” correction. I suspect (but have not tested, this is me “spitballing”) that the clock simply doesn’t increment at all when the phone is off. So if that’s true, I turn it on where there’s no wifi, it would start up at whatever time it had shown when I turned it off. (I can test this a multitude of ways, obviously–I just had the thought a couple of hours ago.)
Really right now my major concern is the cell modem; if that’s working the clock will be updated at almost all times (though ideally I’d like that fixed, too). Plus, I still wear a watch (and half of my personal library is on clay tablets, too).
…annnnndddddd, it appears my clock issues were due to a nearly drained battery.
It appears to be fine now. I shut off automatic updates, was able to change the time without any trouble; the time was retained when I rebooted the phone, and reverted to the correct time when I turned on automatic updates again.
So that leaves the modem issue. ANd I am going to experiment with keyboard layouts (I hate the one that came with the phone).
Makes sense. In your typical desktop, there is a button cell battery that provides power to the real-time-clock (RTC) chip, which therefore runs even when the desktop is completely powered off (switched off at the wall) and maintains the time from shutdown to boot.
In devices that run off battery, it would be inefficient to include two batteries but then you get the downside that if the laptop or phone battery goes completely flat, you have lost the time out of the RTC chip.
(For regular phones, without a killswitch, that’s not such a big deal because as soon as you are on the mobile network, you can use NTP to get the time from the internet and/or as soon as you manage to connect to WiFi presuming that the WiFi gets you to the internet and again, without a killswitch.)
OK, the modem seems to be responsive…somewhat. It can’t find T-mobile (USA) usually, and when it does, calls hang up before I even lift my finger off the call button (which tells me something is wrong with phone calls). I don’t know yet about data; unfortunately mobile data with T-mobile here where I am at right now barely exists.
Under privacy you can supposedly set how long it takes for the screen to go dark, and how long afterwards it locks (or disable that, entirely). But no matter what I set it to, the phone locks the screen immediately after the screen goes dark. That’s rather annoying when I’m sitting right next to it (no security issue) but doing something else. I can either leave the screen on for half an hour or longer and drain the battery, or have to enter the dang passcode every time I pick the thing up.
I’ve just had a look at this - unfortunately the parts of the settings app that are needed to set this up on the L5 are not adaptive, and I found them impossible to use. I know that it is possible to do the setting up from the cli instead, but that needs knowledge, and is fiddly on the built-in terminal anyway. (By comparison, the corresponding functionality under Sailfish once you have installed the developer utilities is plug-and-play, I think that Sailfish starts up a DHCP server on connection so the host computer’s USB network address is set up automatically too.)
My advice for a wired connection on the L5 would still be to connect to an ethernet network if possible. Virtually all consumer ISP-supplied routers in their default configuration will have a DHCP server running, and the L5 will get an address on the network without any issues in most cases.
What kind of an oscillator runs the CPU? In this day and age, an accurate clock, to within mili-seconds of drift per day, is pretty much standard on every low-end device. How could the Librem 5 be so impaired when it comes to keeping time? A fifty year old mechanical wrist watch does better. Any real-time clock circuit with even a reasonably good (lower resolution) RC oscillator should perform much better. These performance of these time-keeping components are all specified in the manufacturer’s datasheets. It’s kind of hard to get wrong unless you make easily-fixable mistakes in the code and then choose not to fix those mistakes. Does anyone know why the device is having difficulty keeping time?
The phone’s time algorithms should keep all external network connections away from affecting the phone’s time, except perhaps (maybe) when changing time zones or adjusting for daylight savings time changes twice per year. In a phone that has network kill switches as a prominent feature, I wouldn’t want any network to have access or the phone itself, or to have any reliance on any network, especially by default, to always display an accurate time. It just doesn’t make sense to let any network be a part of that equation any more than is absolutely needed, which should leave “never” as a default and valid option in user defined time settings.
But doesn’t this require the L5 to be able to find a DHCP server? In my experience, a typical desktop or laptop Linux distribution doesn’t have one enabled by default (although maybe a Librem laptop or Librem mini does?). Anyway, when I connect my L5 to my Linux desktop or Linux laptop by USB, it only sets up a IPv6 address on the USB interface, not an IPv4 one, which limits its usefulness, and certainly doesn’t allow the L5 to check for updates via the USB-connected computer (which was the question that started this thread).
I know I discussed the problem in two different places, so maybe this is NOT the thread where I pointed out that the battery was most likely very low at the time. No oscillator can run worth beans without power. The reason desktop computers seem to be able to keep time even when off and even unplugged, is that that oscillator has its own battery (good for many years). No cellphone I am aware of has a separate battery for the clock.
Apparently, the mainstream smart phones all sync the time via some network, probably the cellular provider, every time they lose power and then boot-up again when power has been restored. But in a phone that prides itself on kill switches to prevent spying, maybe this isn’t such a useful method. I want to leave networking turned off most of the time. I would guess that time is likely derived also via GPS signals, which are one-way from space to the earth. You don’t even get pinged. Maybe another option might be to use cell-site paging services to communicate time to the phone. Personal pagers don’t transmit either (if they still exist). I think the idea for time should be to derive time on the phone from a source that doesn’t expect any kind af communications acknowledgement from the Librem 5 but that tells the Librem 5 what time and date it is.
A GPS receiver would do the trick. The difficulty would be in verifiably making sure the GPS location doesn’t leave the phone once established (as far as I know, it doesn’t go back out through GPS). Which, presumably, the kill switches would accomplish…though someone will no doubt (please do!) correct me if I am wrong.
If the phone has a charged battery (and it needs, probably, at least 5 percent), it should keep time fairly well without needing a sync. Mine has since the initial issues.
There is a part of Arizona and Nevada where even my Samsung Note 9 struggles to keep accurate time. Bullhead City (Arizona) and Laughlin (Nevada) operate in many respects, as one city, although they are each separate cities, each in in their own respective adjoining States. This dual-city is divided only by the Colorado River with several bridges connecting the two cities. The Laughlin Nevada half of this geography has daylight savings time for half of the year. The Bullhead City Arizona half never has daylight savings time. So you can gain or lose an hour (depending on the time of year), simply by crossing a city bridge. Scattered along the river that divides these two halves, are several cell phone towers on both sides of the river. The cell phone connection to the towers doesn’t seem to respect the city limits. One minute it’s 3:00 PM. A minute later, it’s 4:01 PM. Ten minutes after that, it’s 3:12 PM. You never really know what time it is because most of us don’t look at the time often enough to know for sure what time it really is in this situation. So before you leave home for either city, you set your phone to ignore the towers and just keep its own time. Then you make the one-hour time adjustment in your head, depending on which side of the river you’re on when you check your phone to see what time it is.
Well, that SHOULD work for the Librem 5 as long as its battery remains charged. For that matter, any other cell phone probably has that limitation, too.
Yes, I pointed this out to show that no matter how the time is maintained, there can always be challenges. It might be nice to display the time zone name, right next to the time. I didn’t realize initially, how complex keeping time on a phone can be. But I think there always needs to be a mode upon boot-up that says: 1.) Find GPS location and time, 2.) Set the phone time and date based upon this information one time only, 3.) Ignore all cell phone tower time and date information after that.
On the L5 you can disable on-line time updates. Whether that means ignoring the cell towers, I will be able to explore once I get the replacement modem (the one I had, had the wrong firmware). I’ve been promised a replacement shipped to me, no cost.