Possible bug? - Page 3

DiapDealer · 09-08-2021, 07:42 PM

I should have some time coming up to test what you guys are cooking up on both virtual and physical Linux and Windows machines. Just let me know what's needed.

Do we need to consider a cross-platform utility class/function that we can easily call from anywhere, or is it something that will probably be OK with a bit of well-placed MainWindow ifdef-ing?

KevinH · 09-08-2021, 09:00 PM

It all depends. I would love to create a utility routine to create a table of keycode, shift used, and nativeVirtualKeyCode (basically a quick and dirty keyboard mapping of base and shifted keys).

This would work crossplatform. I just do not know how the nativeVirtualKeyCode is determined or how to create a map given those values are only provided inside a QKeyEvent and we can not force people to hit each key once with and without shift just to build a table each time Sigil is started.

Perhaps let's just focus on getting Windows fixed as macOS seems to function okay with the current code for the most part.

DiapDealer · 09-08-2021, 11:18 PM

Anything helpful or relevant here?

https://github.com/kovidgoyal/calibr...eyboard.py#L33

https://github.com/kovidgoyal/calibr...rtcuts.py#L109

KevinH · 09-09-2021, 10:28 AM

Yes!

Quote:

Originally Posted by DiapDealer

Anything helpful or relevant here?

https://github.com/kovidgoyal/calibr...eyboard.py#L33

Kovid is handling knowing if a shift is needed or not by checking if the key when uppercased == the key when lowercased. That should be true for all numbers and punctuation and all special symbols.

It is very similar to the code we use - ours uses unicode character classes - to accomplish something close.

Let's give Kovid's way a try to see how that works.

Excellent! I have my fingers crossed.

KevinH · 09-09-2021, 10:54 AM

Nope! That code is very broken on macOS. It removes the Shift from the sequence Ctrl+Shift+1 since all digits have no upper or lower and so meet the criteria for removing the SHIFT.

Perhaps it will help Windows. I will try that next.

KevinH · 09-09-2021, 11:09 AM

Okay, I have ifdef'd calibre's approach for use on Windows only. I pushed it to master with some additional debug output and used [deploy] in the commit message.

Would you give it a try and let me know if that works any better than what we have on Windows?

Please try a bunch of code sequences including Ctrl+Shift+1, Alt+Shift+1 and without the shift.

Sarmat89 · 09-09-2021, 11:15 AM

Some keyboards have lowercase letters in the Shift register.

KevinH · 09-09-2021, 11:29 AM

Quote:

Originally Posted by Sarmat89

Some keyboards have lowercase letters in the Shift register.

Understood but they should work fine in the Windows test since they differ between upper and lower case.

DiapDealer · 09-09-2021, 11:47 AM

Quote:

Originally Posted by KevinH

Okay, I have ifdef'd calibre's approach for use on Windows only. I pushed it to master with some additional debug output and used [deploy] in the commit message.

Would you give it a try and let me know if that works any better than what we have on Windows?

Please try a bunch of code sequences including Ctrl+Shift+1, Alt+Shift+1 and without the shift.

That seems to working with everything I've thrown it so far (on an English standard Windows keyboard).

Ctrl+Shift+1 results in a working Ctrl+! shortcut (displayed and saved).
Ctrl+Alt+1 results in a working Ctrl+Alt+1 shortcut (displayed and saved).
Alt+Shift+1 results in a working Alt+! shortcut (displayed and saved).
Alt+Shift+G results in Alt+Shift+G that works
Ctrl+Shift+9 results in Ctrl+(
Alt+Ctrl+9 results in Alt+Ctrl+9

All seem to work and are stored/restored from the .ini in working order.

I've tried lots more than that, but nothing seems to messing up that I know of. I'm not a much of a shortcut user, though, so I might have missed something.

KevinH · 09-09-2021, 12:08 PM

Wonderful!

@BeckyEbook would you please try current master on your International keyboard and see if it shows the correct mappings too.

@DiapDelaer - when you get a free moment can you test things are still fully working on Linux as well. Linux still uses our original code so let's hope it just works.

Thanks!

Quote:

Originally Posted by DiapDealer

That seems to working with everything I've thrown it so far (on an English standard Windows keyboard).

Ctrl+Shift+1 results in a working Ctrl+! shortcut (displayed and saved).
Ctrl+Alt+1 results in a working Ctrl+Alt+1 shortcut (displayed and saved).
Alt+Shift+1 results in a working Alt+! shortcut (displayed and saved).
Alt+Shift+G results in Alt+Shift+G that works
Ctrl+Shift+9 results in Ctrl+(
Alt+Ctrl+9 results in Alt+Ctrl+9

All seem to work and are stored/restored from the .ini in working order.

I've tried lots more than that, but nothing seems to messing up that I know of. I'm not a much of a shortcut user, though, so I might have missed something.

DiapDealer · 09-09-2021, 12:15 PM

I'll have to wait until I get home to try on Linux. It _should_ work, since it did before. My only worry is that it might only be on newer versions of Qt on Linux that it "just works". But if I get a chance, I'll try to check it out on some older Qts using some Linux VMs.

KevinH · 09-09-2021, 12:29 PM

Great!

BTW, I have been able to use a variant of calibre's approach and an ifdef to macOS to handle the macOS broken case of Meta+Shift+! (and @#$ etc) to properly make it Meta+! as it should be.

So hopefully we will have all platforms working better than they were before.

DiapDealer · 09-09-2021, 12:48 PM

Sounds like plan!

BeckyEbook · 09-09-2021, 01:58 PM

@KevinH:

I built a new Sigil and I have mixed feelings.

1. I don't like (no offense) the way Calibre does this. For me, Ctrl+Shift+1 is Ctrl+Shift+1, not Ctrl+!
2. In the current master version, Shift is omitted, so I can see Ctrl+! or after pressing Alt+Shift+2 I can see Alt+@
3. When in the field with the shortcut I put a Polish (and probably also a different one) diacritic, I see some unreadable double-byte sequence of characters.

I would suggest looking at another solution that shows the shortcuts in a more friendly way.

That's why I checked out two other GPL 3.0 projects on Github and maybe this code will help:

https://github.com/hluk/CopyQ/blob/6...tform.cpp#L263

https://github.com/QYQYMBA/ChiyaProj...indow.cpp#L148

In these applications, the Windows keyboard shortcuts display values Ctrl+Shift+1 (not Ctrl+!).

Becky

KevinH · 09-09-2021, 02:42 PM

The first link you sent is the simplest and it appears to use the Windows only approach you mentioned earlier.

But I am confused ... did calibre's code work correctly and you just did not "like" the Shortcut shown, or did it actually fail with some key sequences and that is why you did not "like" the Shortcut shown?

And what do you mean by "when in the field with the shortcut"?

Do you mean the "Ctrl+!" output shown? If so, that is not meant to be an input field to be edited. Entering other keys there will create issues. For example the del key or backspace will erase the key sequence completely. It is not meant for editing directly but to display the key sequence entered.

Or do you mean the table entry itself or when editing the ini?

Quote:

Originally Posted by BeckyEbook

@KevinH:

I built a new Sigil and I have mixed feelings.

1. I don't like (no offense) the way Calibre does this. For me, Ctrl+Shift+1 is Ctrl+Shift+1, not Ctrl+!
2. In the current master version, Shift is omitted, so I can see Ctrl+! or after pressing Alt+Shift+2 I can see Alt+@
3. When in the field with the shortcut I put a Polish (and probably also a different one) diacritic, I see some unreadable double-byte sequence of characters.

I would suggest looking at another solution that shows the shortcuts in a more friendly way.

That's why I checked out two other GPL 3.0 projects on Github and maybe this code will help:

https://github.com/hluk/CopyQ/blob/6...tform.cpp#L263

https://github.com/QYQYMBA/ChiyaProj...indow.cpp#L148

In these applications, the Windows keyboard shortcuts display values Ctrl+Shift+1 (not Ctrl+!).

Becky

09-08-2021, 11:18 PM	#33
DiapDealer Grand Sorcerer Posts: 27,549 Karma: 193191846 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	Anything helpful or relevant here? https://github.com/kovidgoyal/calibr...eyboard.py#L33 https://github.com/kovidgoyal/calibr...rtcuts.py#L109 Last edited by DiapDealer; 09-08-2021 at 11:34 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[BUG] - M96 out of memory - [BUG]	Alf77	Onyx Boox	5	02-05-2015 11:47 AM
Another bug that I wonder if others have seen	PeterT	Kobo Reader	16	06-08-2013 09:48 PM
DR800 Help, I've got a bug!! A bug on my screen!!	Franky	iRex	4	06-21-2011 11:45 AM
Embedded font bug or CSS bug in ADE	JSWolf	ePub	10	06-11-2011 02:34 PM
PRS-505 bug or eBookLib bug?	porkupan	Sony Reader	3	10-07-2007 10:44 PM

09-08-2021, 07:42 PM	#31
DiapDealer Grand Sorcerer Posts: 27,549 Karma: 193191846 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	I should have some time coming up to test what you guys are cooking up on both virtual and physical Linux and Windows machines. Just let me know what's needed. Do we need to consider a cross-platform utility class/function that we can easily call from anywhere, or is it something that will probably be OK with a bit of well-placed MainWindow ifdef-ing?

09-08-2021, 09:00 PM	#32
KevinH Sigil Developer Posts: 7,644 Karma: 5433388 Join Date: Nov 2009 Device: many	It all depends. I would love to create a utility routine to create a table of keycode, shift used, and nativeVirtualKeyCode (basically a quick and dirty keyboard mapping of base and shifted keys). This would work crossplatform. I just do not know how the nativeVirtualKeyCode is determined or how to create a map given those values are only provided inside a QKeyEvent and we can not force people to hit each key once with and without shift just to build a table each time Sigil is started. Perhaps let's just focus on getting Windows fixed as macOS seems to function okay with the current code for the most part.

09-09-2021, 10:54 AM	#35
KevinH Sigil Developer Posts: 7,644 Karma: 5433388 Join Date: Nov 2009 Device: many	Nope! That code is very broken on macOS. It removes the Shift from the sequence Ctrl+Shift+1 since all digits have no upper or lower and so meet the criteria for removing the SHIFT. Perhaps it will help Windows. I will try that next.

09-09-2021, 11:09 AM	#36
KevinH Sigil Developer Posts: 7,644 Karma: 5433388 Join Date: Nov 2009 Device: many	Okay, I have ifdef'd calibre's approach for use on Windows only. I pushed it to master with some additional debug output and used [deploy] in the commit message. Would you give it a try and let me know if that works any better than what we have on Windows? Please try a bunch of code sequences including Ctrl+Shift+1, Alt+Shift+1 and without the shift.

09-09-2021, 11:15 AM	#37
Sarmat89 Evangelist Posts: 482 Karma: 2267928 Join Date: Nov 2015 Device: none	Some keyboards have lowercase letters in the Shift register.

09-09-2021, 12:15 PM	#41
DiapDealer Grand Sorcerer Posts: 27,549 Karma: 193191846 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	I'll have to wait until I get home to try on Linux. It _should_ work, since it did before. My only worry is that it might only be on newer versions of Qt on Linux that it "just works". But if I get a chance, I'll try to check it out on some older Qts using some Linux VMs.

09-09-2021, 12:29 PM	#42
KevinH Sigil Developer Posts: 7,644 Karma: 5433388 Join Date: Nov 2009 Device: many	Great! BTW, I have been able to use a variant of calibre's approach and an ifdef to macOS to handle the macOS broken case of Meta+Shift+! (and @#$ etc) to properly make it Meta+! as it should be. So hopefully we will have all platforms working better than they were before.

09-09-2021, 12:48 PM	#43
DiapDealer Grand Sorcerer Posts: 27,549 Karma: 193191846 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	Sounds like plan!

09-09-2021, 01:58 PM	#44
BeckyEbook Guru Posts: 692 Karma: 2180740 Join Date: Jan 2017 Location: Poland Device: Misc	@KevinH: I built a new Sigil and I have mixed feelings. 1. I don't like (no offense) the way Calibre does this. For me, Ctrl+Shift+1 is Ctrl+Shift+1, not Ctrl+! 2. In the current master version, Shift is omitted, so I can see Ctrl+! or after pressing Alt+Shift+2 I can see Alt+@ 3. When in the field with the shortcut I put a Polish (and probably also a different one) diacritic, I see some unreadable double-byte sequence of characters. I would suggest looking at another solution that shows the shortcuts in a more friendly way. That's why I checked out two other GPL 3.0 projects on Github and maybe this code will help: https://github.com/hluk/CopyQ/blob/6...tform.cpp#L263 https://github.com/QYQYMBA/ChiyaProj...indow.cpp#L148 In these applications, the Windows keyboard shortcuts display values Ctrl+Shift+1 (not Ctrl+!). Becky

Advert

Advert