Sigil-0.9.4 Released

[DiapDealer]

Originally Posted by sdjskr

And I am expecting some refreshed manual to compile the source code from Sigil in Windows environment. Someone promised it. I also want to find some clue to resolve this issue.

I am that someone (though "promised it" is strong wording). Please don't get your hopes up that it's going to be very soon. My schedule is quite full, and documentation is the first thing to get pushed to the back burner. I will get to it, but it's not going to be immediate.

[KevinH]

sdjskr

Sorry, your video will not work for me. It just shows an error message. If switching to CodeView and back makes the cursor reappear, then it is definitely a lost focus issue. If you lose focus when hitting spacebar or return, but not when adding composed korean characters, that means something else is stealing focus from the tab widget. That something else may not be Sigil at all.

So please try one more thing. In BookView make the problem happen, but instead of switching to CodeView and back, simply use your mouse to click anyplace inside the BookView tab or on the tab marker at the top of the screen. Both of those should change focus. What happens then?

KevinH

[sdjskr]

Originally Posted by KevinH

Sorry,
Your video will not show for me. If switching to codeview and back makes the cursor reappear, then it is definitely a lost focus issue. If you lose focus when hitting spacebar or return, but not when adding composed korean characters, that means something else is stealing focus from the tab widget. That something else may not be Sigil at all.

So please try one more thing. In BookView make the problem happen, but instead of switching to codeview and back, simply use your mouse to click anyplace inside the BookView tab or on the tab marker at the top of the screen. Both of those should change focus. What happens then?

KevinH

Once the cursor disappears, it does not show even if I click anyplace.
I tried this on Windows 7, nothing else was installed but Sigil. The problem still exists.

As for the video, I'll give you another link. Please, pay attention to the location of the cursor when I switched to CodeView.

https://www.youtube.com/watch?v=JUHSUP4MF48

I have found something. If I minimize Sigil window, the cursor reappears when I get it back.

And, I'm not sure if this is helpful but, East Asian languages like Korean take about 3 bytes in UTF-8(surrogate pairs), 2 bytes in UTF-16(BMP).

[KevinH]

The location of the cursor in CodeView is working as dsigned. CodeView and BookView are two separate views but are not simultaneous. When you go from BookView to Codeview, the complete contents of the BookView is extracted and used to load into CodeView. And visa-versa.

The problem is entering a carriage return/linfeed after inputing Korean results in focus being lost to the BookView widget. So when you do that something is stealing focus away from the widget. By Going to CodeView andback, focus is regained. By minimizing and back focus is regained. But by clcking inside Bookview focus is not regained.

Just in case these is some key accelerator assigned for the enterkey by mistake, please try after renaming the Sigil.ini file (to save it) and restarting Sigil to get back to default.

If that does not help, we will have to figure out what is intercpting the focus and taking it away from BookView. That should not and does not happen on other systems.

KevinH

[sdjskr]

Originally Posted by KevinH

The location of the cursor in CodeView is working as dsigned. CodeView and BookView are two separate views but are not simultaneous. When you go from BookView to Codeview, the complete contents of the BookView is extracted and used to load into CodeView. And visa-versa.

The problem is entering a carriage return/linfeed after inputing Korean results in focus being lost to the BookView widget. So when you do that something is stealing focus away from the widget. By Going to CodeView andback, focus is regained. By minimizing and back focus is regained. But by clcking inside Bookview focus is not regained.

Just in case these is some key accelerator assigned for the enterkey by mistake, please try after renaming the Sigil.ini file (to save it) and restarting Sigil to get back to default.

If that does not help, we will have to figure out what is intercpting the focus and taking it away from BookView. That should not and does not happen on other systems.

KevinH

What do you think?
Would that be the reason why WebView is eliminated in Qt 5.6?
Not enough for some non ascii characters?

https://www.youtube.com/watch?v=jh46E5gd5dE

Webview losing cursor

[KevinH]

QWebView was deprecated in Qt5.6 because Qt could not support two different webkit implementations. QtWebkit and Chromes QtWebEngine. So they chose Chromes WebEngine over Apples webkit. The problem is Chromes WebEngine does not have all of the capabilities of Qt webkit and has as many or more bugs. Sigil has no plans to move to Qt 5.6 as Sigil needs Qt webkit. Any move would require a large rewrite of Sigil and loss of functionality in Bookview. My tests show QWebEngine is missing basic audio/video support and is not nearly as stable as Qtwebkit is right now.

KevinH

[KevinH]

No, all that shows is something is stealing focus from the QWebpage on your system. We already deduced that. The question is what and why is focus being taken. It must have something to do with some graphics driver or system level utility stealing focus, as this problem does not exist on other platforms or even reports in other languages.

A google search for losing focus on Windows 7 results in a long list of problem reports and possible causes. Something else is grabbing the focus away but only when switching from Korean back to hit the enter key.

KevinH

[KevinH]

In your mini Qwebview QtCreator program, could you try adding a slot and registering it to the microfocusChange signal so that we can confirm focus is actually being lost and not that the cursor is simply being made invisible.

[KevinH]

Alternatively, add your own QinputMethod event handler and log all of the inputMethodEvents being generated to see if any are making the cursor invisible.

[sdjskr]

Originally Posted by KevinH

No, all that shows is something is stealing focus from the QWebpage on your system. We already deduced that. The question is what and why is focus being taken. It must have something to do with some graphics driver or system level utility stealing focus, as this problem does not exist on other platforms or even reports in other languages.

I have tried on different machine. Now, it's a Japanese version of Windows 7.

https://www.youtube.com/watch?v=-2MM4zbkaTE

This time, Return works, even though I have to hit twice, and the cursor does not turn into the under bar like when typing Korean.

However, if I press Space bar, it starts losing focus again!.

Do you still think this is limited to my machine or Korean character?

Wow, now I can post without the permission? Thank you.

[KevinH]

After doing lots of web searches and some code reading, I think the problem is a bug in QinputMethod in Qt that is triggered by your Windows 7 IME for Korean. I think it is related to automatic closing of partially composed characters either done by a space or by enter. This bug may cause the loss of focus or it may inadvertently cause the cursor to become invisible.

If we can actually figure out what is happening (lost focus vs invisible cursor) we may be able to add workarounds to the code. I would try adding your own QinputMethodEvent handler to your simple code example and use QDebug to log all of the QInputMethod events so that we can see exactly what Qt is receiving from the Windows IME under Korean on your system. So you need to subclass QWebPage/QWebView and overload the QinputMethodEvent handler and log what comes in event wise from your IME.

KevinH