Broken edge swipes in landscape in Äijänpäivänjärvi

Question

24

Broken edge swipes in landscape in Äijänpäivänjärvi

bug
1.1.4.29

asked 2015-04-28 11:38:49 +0200

pichlo

4062 ●64 ●92 ●92

updated 2015-04-29 11:00:03 +0200

jaacoppi
591 ●5 ●16 ●16

http://fromasailfishosnoo...

After the upgrade to Äijänpäivänjärvi, edge swipes in landscape behave inconsistently. They work fine for some apps but they act as if the device were in protrait for others. So, for example, a left edge swipe is interpreted as a top edge swipe and closes the app instead of switching to the home screen. Likewise for other edge swipes.

Steps to reproduce:

Turn your phone to landscape: speaker and USB port on the left, mic and LED on the right, buttons on the top.
Start one of the affected apps, e.g. File Browser.
Swipe from the left edge.
Tada!

It does not behave like that for all the apps, only for some. A quick test found: File Browser, Webcat, Modrana, Web Pirate, LLsvPlayer. It worked fine before the OS update.

EDIT Addition to the steps to reproduce. Once an affected application is started in portrait, the issue does not show until it is started again in landscape TWICE. Curiouser and curiouser.

edit retag flag offensive close delete

Comments

2

First need to heck these applications, as for example my applications are not affected, and i didn't added any hacks for orientation in it.

coderus ( 2015-04-28 13:09:23 +0200 )edit

1

Thanks for the suggestion, @coderus. And to @szopin on TMO. It turns out the problem is caused (or at least triggered) by the "Eventview controls attached page" patch. See also this TMO post for details.

(Caveat: It is possible that other patches trigger it too. I stopped experimenting on the first hint of success.)

pichlo ( 2015-04-28 15:43:31 +0200 )edit

unfortunately, it's not true and this patch is not related to your problem.

coderus ( 2015-04-28 15:55:57 +0200 )edit

denial n.

(Psychology) a psychological process by which painful truths are not admitted into an individual's consciousness.

:)

It is unfair to blame the patch and I was not trying to do that. Only trying to explain why some users were able to reproduce it whilst other were not. I suspect the patch somehow exposes a code path - a switch from the launcher to the app's own screen perhaps - that is otherwise not exercised and thus exposes the bug.

We can continue arguing and/or pretending that the issue does not exist or we can work together on finding the true cause and a solution. Your call.

pichlo ( 2015-04-28 17:20:52 +0200 )edit

3

omg, i'm developer of this patch, and i know it's not related. you can reproduce same on clean system with no patches enabled.

coderus ( 2015-04-28 17:33:15 +0200 )edit

First need to heck these applications, as for example my applications are not affected, and i didn't added any hacks for orientation in it.
coderus ( 2015-04-28 13:09:23 +0200 )edit
Thanks for the suggestion, @coderus. And to @szopin on TMO. It turns out the problem is caused (or at least triggered) by the "Eventview controls attached page" patch. See also this TMO post for details.
(Caveat: It is possible that other patches trigger it too. I stopped experimenting on the first hint of success.)
pichlo ( 2015-04-28 15:43:31 +0200 )edit
unfortunately, it's not true and this patch is not related to your problem.
coderus ( 2015-04-28 15:55:57 +0200 )edit
denial n.
(Psychology) a psychological process by which painful truths are not admitted into an individual's consciousness.
:)
It is unfair to blame the patch and I was not trying to do that. Only trying to explain why some users were able to reproduce it whilst other were not. I suspect the patch somehow exposes a code path - a switch from the launcher to the app's own screen perhaps - that is otherwise not exercised and thus exposes the bug.
We can continue arguing and/or pretending that the issue does not exist or we can work together on finding the true cause and a solution. Your call.
pichlo ( 2015-04-28 17:20:52 +0200 )edit
omg, i'm developer of this patch, and i know it's not related. you can reproduce same on clean system with no patches enabled.
coderus ( 2015-04-28 17:33:15 +0200 )edit

Answer 1

14

answered 2015-05-10 04:18:17 +0200

drcouzelis

5165 ●27 ●36 ●35

updated 2015-05-10 04:18:39 +0200

Ok, I think I've figured this out.

As you can see on the changelog[1] there are many references to "Bind allowedOrientations to defaultAllowedOrienatations". In other words, something official changed in regards to default orientations. I then looked at many of the official applications (for example, "jolla-notes") and found this code:

allowedOrientations: defaultAllowedOrientations
_defaultPageOrientations: Orientation.All

I then applied that to "/usr/share/harbour-file-browser/qml/pages/DirectoryPage.qml" by changing this:

allowedOrientations: Orientation.All

To this:

allowedOrientations: defaultAllowedOrientations

Edge swipes now work correctly in File Browser.

I think two things need to be done now for this issue to be truly resolved:

Find official documentation that explains this change to default orientations and how it works.
Update all applications, both those in the Jolla Store and in the Warehouse.

[1]: https://together.jolla.com/question/89803/changelog-114-aijanpaivanjarvi/

edit flag offensive delete publish link

Comments

2

I can confirm the same works with filetug. So obviously there has been some yet undocumented change in the API...

casanunda ( 2015-05-11 18:59:53 +0200 )edit

Broken edge swipes in landscape in Äijänpäivänjärvi

Comments

1 Answer

Comments

Question tools

Stats

Related questions

Broken edge swipes in landscape in Äijänpäivänjärvi

Comments

1 Answer

Comments

Question tools

Public thread

Stats

Related questions