- User Since
- Aug 20 2016, 6:10 PM (197 w, 6 d)
Apr 29 2018
Mar 21 2017
Mar 3 2017
I confirmed that the mouse cursor can still be moved by using the touchscreen, and the keyboard still works, so this is definitely a touchpad issue. Plugging in an external mouse works fine too.
Mar 1 2017
I see. In that case, is there any chance we'll ever get an option to allow the use of one password for encryption as well as login, where you could still change your user password, causing the login prompt to appear as usual? I don't care if this isn't default behavior. Merely having an option would be a major productivity bonus, especially in an IT office context like mine, where we're rebooting various machines all day long.
Feb 28 2017
2-3 seconds would be a major improvement! But... could we use UUIDs later as they start to come online? In other words, could we still do the same task, but just in the background after login, or maybe while the password is being typed in (there's a few seconds right there)?
Nice work, macifom. It seems to me that we need a distinction between "I'm closing the lid because I'm going to be away for a while" and "I'm closing the lid because I have external HDMIs or Thunderbolts or whatever other displays connected that I want to use instead". In the former case, we need to lock ASAP. In the latter, I would actually say that the lock countdown should be reset, because a lid closure is evidence of user activity, not unlike a keystroke or mouse movement.
Feb 19 2017
Feb 12 2017
Considering that you got the login background but could not see the prompt, which is centered, that sounds like a resolution mismatch between your monitor and your laptop screen, which is not being handled properly when you're logged out. This is a delicate issue because it begs the question of what settings the driver should be able to see when you're logged out, vs. what it shouldn't. (I suppose this is one reason why Solus quite sensibly separates logged in from logged out background images.)
Feb 9 2017
Thanks for the clarification, and I'm pleased to see that you're actively monitoring OpenSSL security alerts. You might want to mention this issue in the docs, especially if it applies to other libs as well. Seriously no one would think that libxyz.4.3 means version 5.1. I guess we need to disregard the name and use "eopkg info xyz" instead.
Wait a second... then why do we have libssl.so.1.0.0 instead of libssl.so.1.0.2?
Feb 7 2017
Update: I booted with the mouse plugged in, and never removed it. Just before doing so, I tested the touchpad, and it was dead. Maybe this is because I happened to boot up with the lid closed, as I was on external HDMI? I think it's safe to say at this point that this is easily reproduced within a workday. Locking and unlocking the screen a few times may be necessary. Perhaps there's a way to always make it happen, but I haven't had time to test it in a directed manner.
Feb 4 2017
Feb 2 2017
I just tested this a bit further. I booted the laptop connected to a monitor. Then I disconnected it while still in mirror mode. After that, I let it fall into lock mode with a black screen. As expected, I could not see the password I was typing in until I connected an external monitor. I typed the password, but did not yet hit Enter because I wanted to see what would happen if I unplugged and plugged the HDMI several times in succession, with plenty of resync time in between. I ended up hanging the machine with the password dialog showing after maybe 5 cycles. Maybe this wouldn't have occurred, had I allowed more resync time between plug events. Then again, robust software would not need extra time in order to prevent hanging the machine. Unfortunately, I have no idea where in the stack between me and the hardware the problem actually resides.
Jan 25 2017
I think I see what's going on now. When I delete the "Video" folder (which is easily done), the bookmark to it on the left panel of Files still remains. If I right click on that bookmark, "Remove" is grayed out, meaning that I'm not able to remove the bookmark. However, sometime later (after a reboot?), something figures out that the target folder was deleted, so the bookmark removes itself automatically. (I would have expected to be able to do this manually via the right click menu). It's not the worst problem, and I could live with it, but it's weird how "Remove" was disallowed.
Jan 23 2017
Click on the image below to zoom. This is what it looks like on Intel 615 graphics, right here on Phabricator in Firefox. As you'll notice, it's a purely horizontal phenomenon, so the dividing lines between bug reports look normal while the text does not. All other windows look normal.
Jan 22 2017
Thanks DataDrake. I vote to increase the priority of this task.
Same here, although in FF it tears regardless of whether fullscreen, and even with this website here. Ironically, as I was entering this comment, the whole screen got uniform tearing all over the browser window. I took a screenshot to attach, but the act of doing so fixed all the tearing. I suspect that this is because taking a screenshoot refreshes things, so the problem goes away. The problem happens maybe once an hour, seemingly at random. I will see it for a split second, for example, during scrolling, when presumably ever single window update has some tiny chance of tearing everything, but otherwise does so correctly. I don't know where in the stack between me and the hardware this problem lies.
Jan 15 2017
Nov 24 2016
It happened again just recently, but unfortunately it's not consistent so I can't be sure of the answer to your question. If you like we can leave this open for a while and I'll report back if it recurs.
Nov 23 2016
Nov 22 2016
I repeated the experiment on the newer machine. This time, I got the low battery warning with 12 minutes left. But then, a few minutes later, it lost power without ever hitting the critical warning. It's as though the alerts are triggered based on estimated time remaining, instead of percentage remaining. I doubt that's the case, as of course it's pretty much impossible to predict time remaining. In any event, something is causing this variable behavior from run to run and machine to machine.
Nov 18 2016
I figured I'd help out by trying this on a completely different laptop. The one described above has continued to shut down without warning when the battery gets down to less than half, more or less. Out of about 10 times, it only gave a battery warning once. Most of the other times, it could actually be restarted successfully while still on battery. The machine check errors have never reappeared, so I don't know what to make of this. It could well be a hardware or firmware problem.
Well sure maybe it's something else, although it would be useful to test what happens if you artificially kill a few packets during an attempted upgrade. Anyway I'm glad you're looking into this.
Perhaps this is due to problems retrying network transactions that fail due to lost packets. It seems as though the algorithm doesn't have a robust ability to adjust to network conditions which vary over time, especially if certain control packets are lost. At best, it seems to hang there for half an hour before timing out or succeeding. To be fair, I doubt whether this is a Solus problem, as Software Center has always been unreliable in this manner in every distro I ever tried.
Nov 17 2016
Nov 15 2016
This happened again, but in a different way. I did actually get a "battery critical low" message this time, but at the time, the battery still had either 2 bars, or had just gotten down to 1 bar. Seems a little premature for that message, which in most cases occurs with 10 minutes or less remaining. Maybe this is inevitable because my battery interface provides only this large granularity information. (I probably had 30 minutes left.) In any case, it would be worth sitting around and letting a battery die to see how it all behaves, and in particular to identify what's going on with lost file blocks that should have been written long before poweroff. (I plugged in this time, so lost nothing.)
Nov 7 2016
Nov 6 2016
Nov 1 2016
Oct 23 2016
Oct 20 2016
It's good that you can actually install guest tools, but who knows how Solus-aware they actually are, and therefore how well they're actually optimizing. The point is that, ideally, you would just choose "Solus" from the dropdown during VM creation. Recompiling the hypervisor should not be required. This should be easy. It's not Solus' fault that it's not, but someone would need to reach out to the various VM players in order to address this.
To clarify, this can only be fixed by contacting the particular virtualization project in question.
Oct 11 2016
Google just released Project Noto, which is over 100,000 free international character glyphs for open source use:
Aug 26 2016
Aug 20 2016
I just noticed that this was first reported here: