You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Ubuntu Rendering Issue When There is a Long View or Trigger #1001
Ubuntu Rendering Issue When There is a Long View or Trigger
There is a rendering issue on the 'Database Structure' tab when using DB Browser for SQLite under Ubuntu when there is a long view or trigger. I suspect there is some problem with limitless horizontal scrolling.
It happens before I've loaded any data. The trigger is something I'm
writing and related to Canadian taxes. You could probably reproduce it with
multiple kilobytes of random 'ipsum lorem' comments above a simple select
type view. I can probably make a demonstration database that would exhibit
the error if you're interested.
I've just prepared a test database containing two long lorem ipsum views. xx is one long line, yy are many lines. Not sure if they are long enough though. For both views it's struggling (xx prints one long empty line as a tooltip, yy fills the entire screen with its tooltip) but I never get the weird output from your screenshot.
@justinclift Do you still have an Ubuntu VM somewhere to test this? I've only tryed this on KDE. My test db is here.
@yep, I can test this. My Ubuntu VM is really Kubuntu, but it's pretty easy to download standard Ubuntu 16.04.2 and whip up a new one. I'll add this to my task list for this week.
Aha, I knew this sounded familiar. Pretty sure this bug is another occurance of #920, and the solution is to compile it against a newer Qt library.
That being said, I'm trying to compile DB4S on Ubuntu 16.04.2 right now, and it's failing due to a different (known) problem. Looking into solutions for that next... 😄
@MKleusberg As a data point, just tried your test database on an Ubuntu 16.04.2 VM here, compiling using Qt 5.6 (from a 3rd party PPA).
No rendering issue like the above showed up, but the "Database Structure" tab took ages (~9.5 seconds) to show up. Switching from "Database Structure" to any of the other tabs was instant. From them back to "Database Structure"... wait 9.5 seconds again. Every time.
So, something not quite right there. Possibly just something weird with the VM, VMware Fusion, or whatever else too. I'll muck around with a few settings next and see if that fixes the problem.
Ahh, this is interesting. I just installed Qt 5.8 from the official Qt download site (qt.io). When DB4S is compiled using that, the slowness issue isn't present.
Looks like whatever the problem is, it's been fixed in Qt sometime after Qt 5.6.1.
It really looks like both Qt 5.5 and 5.6 on Ubuntu aren't fantastic for DB4S. For anyone having weird slowness or screen corruption issues, it'll need to be compiling using Qt 5.7 or above instead.
Should we add a warning to the build instructions then that we discourage using Qt < 5.7 even though it compiles just fine? Doesn't seem like we can do much else about it.
@christopherthompson81 Did you ever get around to trying a build created with a newer version of Qt? eg something that shouldn't have this rendering problem
Activity
justinclift commentedon May 1, 2017
Oh, that's interesting. Is the data set which you're using to trigger that, a public one?
Just thinking that might be helpful for debugging. 😄
christopherthompson81 commentedon May 1, 2017
justinclift commentedon May 1, 2017
Yep, definitely interested, that would really help. 😄
MKleusberg commentedon May 8, 2017
I've just prepared a test database containing two long lorem ipsum views.
xxis one long line,yyare many lines. Not sure if they are long enough though. For both views it's struggling (xxprints one long empty line as a tooltip,yyfills the entire screen with its tooltip) but I never get the weird output from your screenshot.@justinclift Do you still have an Ubuntu VM somewhere to test this? I've only tryed this on KDE. My test db is here.
justinclift commentedon May 9, 2017
@yep, I can test this. My Ubuntu VM is really Kubuntu, but it's pretty easy to download standard Ubuntu 16.04.2 and whip up a new one. I'll add this to my task list for this week.
justinclift commentedon May 9, 2017
Aha, I knew this sounded familiar. Pretty sure this bug is another occurance of #920, and the solution is to compile it against a newer Qt library.
That being said, I'm trying to compile DB4S on Ubuntu 16.04.2 right now, and it's failing due to a different (known) problem. Looking into solutions for that next... 😄
justinclift commentedon May 9, 2017
@MKleusberg As a data point, just tried your test database on an Ubuntu 16.04.2 VM here, compiling using Qt 5.6 (from a 3rd party PPA).
No rendering issue like the above showed up, but the "Database Structure" tab took ages (~9.5 seconds) to show up. Switching from "Database Structure" to any of the other tabs was instant. From them back to "Database Structure"... wait 9.5 seconds again. Every time.
So, something not quite right there. Possibly just something weird with the VM, VMware Fusion, or whatever else too. I'll muck around with a few settings next and see if that fixes the problem.
justinclift commentedon May 9, 2017
Well, it's not caused by VMware Fusion's implementation of "Graphics Acceleration" for Linux. Same problem both ways.
justinclift commentedon May 9, 2017
Tried it with a fresh build with Qt 5.5 (using the initial compatiblity patch) instead. Same behaviour with that database.
Using a different database (eg this), DB4S has no delay. The "Database Structure" tab shows immediately when switching at all times.
justinclift commentedon May 9, 2017
Ahh, this is interesting. I just installed Qt 5.8 from the official Qt download site (qt.io). When DB4S is compiled using that, the slowness issue isn't present.
Looks like whatever the problem is, it's been fixed in Qt sometime after Qt 5.6.1.
justinclift commentedon May 9, 2017
Qt 5.7 doesn't have the problem either.
It really looks like both Qt 5.5 and 5.6 on Ubuntu aren't fantastic for DB4S. For anyone having weird slowness or screen corruption issues, it'll need to be compiling using Qt 5.7 or above instead.
MKleusberg commentedon May 12, 2017
Should we add a warning to the build instructions then that we discourage using Qt < 5.7 even though it compiles just fine? Doesn't seem like we can do much else about it.
justinclift commentedon May 12, 2017
Yep. If you want to do it, then go for it. Otherwise I'll get it done later on today. 😄
MKleusberg commentedon May 12, 2017
@justinclift Done 😃
justinclift commentedon Feb 16, 2018
@christopherthompson81 Did you ever get around to trying a build created with a newer version of Qt? eg something that shouldn't have this rendering problem
karim commentedon Aug 2, 2018
Closing this due to lack of response.