Firmware 4.31.19086

[Katsunami]

Originally Posted by Sirtel

The Sage button bug is much older than 4.31. If anything, 4.31 improved it. And the Libra 2 has not had any bugs with buttons AFAIK.

The main issues with 4.31 were crashes and database corruptions when sideloading. The Libra 2 tended to freeze and crash when doing several things in quick successsion (like switching fonts, searching etc), but again this was older than 4.31.

Hm... thanks. It seems I avoided the bugs because:
- I had a bunch of books on the device so I didn't need to sideload yet
- The KA1 has no buttons

4.31 (somehow) seems to have massively improved both the refresh/page-flip speed of the reader and the touch accuracy. The latter is one thing which has annoyed me since I had the KA1; slower and less accurate than my 2012 Paperwhite. Therefore I wouldn't want to return to 4.20; even though I now have to reboot to activate the dark mode. (Fortunately I don't really need it that often.)

I'll look into 4.32.

Originally Posted by meeera

Oh yeah I'm already avoiding 4.31 after multiple database corruptions using it. I'm just not rushing headlong to go from .30 to .32, until there are a few more reports confirming that those issues have likely been fixed.

I wonder if Kobo will ever learn. It was like this 6 years ago. I would have assumed that their software would have been improved / rewritten to the point that adding a new functionality would not break other functions.

[MGlitch]

Originally Posted by Katsunami

I wonder if Kobo will ever learn. It was like this 6 years ago. I would have assumed that their software would have been improved / rewritten to the point that adding a new functionality would not break other functions.

You realize this isn’t how software development goes right? In fact the more features added the more likely it is that adding one will have potential affects on others, especially when those features interact in any way.

Kobo does have a beta testing group, but even with this they’re not going to catch every issue. There are too many variables outside of Kobos control. Namely the computer the kobo device is plugged into. And as has been mentioned a lot 4.31s bugs didn’t even affect everyone on this forum.

[NiLuJe]

Originally Posted by davidfor

For whatever reason, the design is that WiFi gets enabled whenever Bluetooth is enabled.

That's because the kernel only exposes a single way to power on the single chip that handles BT & Wi-Fi, so using BT *requires* powering-on the Wi-Fi chip (because it's also the BT chip).
Historically, killing Wi-Fi on NTX boards entailed powering it down, too (probably for power efficiency reasons), so powering the WI-Fi chip was somewhat fairly heavily tied with enabling Wi-Fi .

Which means that making the two *distinct* UI-wise is tricky: you'd have to uncouple toggling Wi-Fi power from toggling Wi-Fi connection (which probably entails mostly always leaving the chip powered on, which is... meh', especially on a board that already suffers from terrible power management. Ironically enough, that BT+Wi-Fi chip might be the most power efficient thing in there ).

[Quoth]

Originally Posted by NiLuJe

Which means that making the two *distinct* UI-wise is tricky: you'd have to uncouple toggling Wi-Fi power from toggling Wi-Fi connection (which probably entails mostly always leaving the chip powered on, which is... meh',

No. Because quite evidently you CAN use the GUI to turn off WiFi after turning on BT. Most gadgets have WiFi and BT in one chip for a few years. I've Android and Linux. I was doing SW & drivers on Linux and Windows to do with BT, WiFi, 3G, 4G etc from 2006.

I've ONLY EVER seen this issue on FW 4.3x.x of Kobos. It can quite easily be sorted. If it's a design decision, it's unique to Kobo.
It's absolutely not a kernel or chip issue that stops the Kobo Sage, Libre 2 and Elipsa from turning on or off BT or simply showing status and making sure the WiFi state is whatever it was last set to by the user.

It absolutely uses more power to have both on.

I've no objection to people leaving WiFi on, but EVERY maker other than Kobo (that I know) doing any ereader, phone, Laptop or Tablet with WiFi and BT, running Linux, Windows or Android lets the user control WiFi & BT independently.
I've 3 x Android phones, 4 x models Android tablets, Boyue Mars eink, played with Kindle PW4 and set up more models of Laptop than I can count with Linux. All control BT & WiFi separately from the user point of view.

Actually many of the WiFi + BT chips also have FM radio, but Apple disables that because they have no earphone socket; the earphone cable is used as the FM aerial, oddly even on my 10" tablet which is big enough to have an internal FM aerial.

No doubt Kobo has made all kinds of excuses for the bizarre 4.30 & 4.31, at least the 4.32 is usuable as you can look at BT setting without risk and you can turn off the WiFi.
The 4.28.x was certainly unfinished GUI on the Elipsa as there was no real indication if BT was on or off and it wasn't properly done at all.
No doubt in the next FW release KOBO will finally have an indicator on the icon bar beside WiFi and fix the bug, because no matter what they call it, it is a bug.

I'd also like the WiFi & BT icons on the tool bar like * (light) and Aa is when you are reading and tap the page. WiFi on/off used to be accessible there.

[Quoth]

Originally Posted by John F

i can't think of another device where enabling one enables the other.

Never seen or heard of it in about 20 years. The chips with BT & WiFi combined are pretty standard and in phones years before ereaders had BT. Most have FM radio too. There is even an Android app (or was) to enable FM radio on some phones & tablets where the maker didn't enable it, worked on very few models as often the aerial wasn't connected.

Actually a lot of BT speakers and other gadgets actually have the WiFi in the BT chip, but just disabled.

[NiLuJe]

Never have I said that this wasn't an NTX quirk, but the fact is that the kernel begs to differ:

static void _ntx_wifi_bt_power_ctrl (int power_status)
{
	int bus;
	int iOldStatus;

	iOldStatus = gi_wifi_power_status;

	//if (14 == gptHWCFG->m_val.bCPU)	// B300
	{
		printk("Wifi / BT power control %d\n", power_status);

		if (power_status & 0x3)	{
			printk("wifi/BT power on\n");
			sunxi_wlan_set_power(1);
		} 

		/* Bluetooth power control */
		if (0 != gptHWCFG->m_val.bBT) {
			if ((power_status & 0x2) == (iOldStatus & 0x2)) {  /* Bluetooth dis status */
				printk("Bluetooth already %s.\n",(power_status & 0x2)?"on":"off");
			} else if (power_status & 0x2) {
				printk("Bluetooth On\n");
				sunxi_bluetooth_set_power(1);
				gi_wifi_power_status |= 2;
			} else {
				printk("Bluetooth Off\n");
				gi_wifi_power_status &= ~2;
			}
		}

		/* Wifi power control */
		printk(KERN_ERR"[DEBUG] power_status:%d , iOldStatus:%d \n",power_status,iOldStatus);
		if ((power_status & 0x1) == (iOldStatus & 0x1)) {  /* check Wifi status */
			printk ("Wifi already %s.\n",(power_status & 0x1)?"on":"off");
		} else if (power_status & 0x1) {
			printk("Wifi On\n");
			//gpiod_direction_output(GPIO_WIFI_RST, 1);
			gi_wifi_power_status |= 1;
		} else {
			printk("Wifi Off\n");
			gi_wifi_power_status &= ~1;
		}

		//printk("Wifi/BT bus:%d \n",bus);
		bus = sunxi_wlan_get_bus_index() ;
		if( power_status&0x3 ) {
			if ( (!!power_status) != (!!iOldStatus) ) {
				sunxi_wlan_set_dis(1);	 /* Wifi dis pull high */
				//printk("sunxi_mmc_rescan_card !!\n");
				msleep(300);
				printk(KERN_DEBUG" sunxi_mmc_rescan_card bus%d!!\n",bus);
				sunxi_mmc_rescan_card(bus);
				msleep(300);
			}
		}
		else if ( 0==power_status ) {
			if ( power_status != iOldStatus )
			{
				sunxi_wlan_set_dis(0);	/* Wifi dis pull low */
				printk(KERN_DEBUG" remove card bus%d!!\n",bus);
				sunxi_mmc_set_card_removed(bus);
				msleep(300);
			}
		}

	
		if (0 != gptHWCFG->m_val.bBT) {
			if ( !(power_status&2) )	{
				sunxi_bluetooth_set_power(0);
				//printk("bt off\n");
			}
		}

		if(0==power_status) {
			printk("wifi/bt device power off\n");
			sunxi_wlan_set_power(0);
			sunxi_mmc_set_card_removed(bus);
			msleep(300);
		}

	}
	//printk("%s() : done \n",__FUNCTION__);
}

That's the ioctl handler.

Keep in mind that I want to make the distinction between powering (hw/sw) and enabling (ui) and connecting (hw/sw/ui) very clear, as the three could theoretically be distinct.

[davidfor]

Originally Posted by Laen

No Problems with 64gb (had ist half of a year in it) and 128gb intern or 64gb extern.
The Problems are now away after factory reset.
the Problems was only the 4 last read books on the home screen and die cover in stand by mode.
all other worked fine.
with database corrupt all would be not working.

there was other users on the web that had 128gb intern.

Are you sure those users where using a 128GB card in a Glo? I have seen that it works in a Clara HD and probably a Glo HD. But, the Glo was around a long time before these larger cards and might not have full hardware support for it. As I said, people had issues with some 32GB cards in the Glo. And as you are having problems, and you have made this change, it is something you need to check.

I can state that my Glo is working with all of the recent firmware versions. It only has a 2GB card in it. That is not the original as that failed some time ago. I only use it for testing, but, it is working correctly.

As to a corrupt database, the symptoms you describe are symptoms of a corrupt database. What happens when the database becomes corrupt depends on exactly how it is corrupt. The device can run for a long time with a corrupt database, it just can't save changes such as reading position. These will be lost either when the device is restarted or as more books are opened.

[Laen]

Originally Posted by davidfor

Are you sure those users where using a 128GB card in a Glo? I have seen that it works in a Clara HD and probably a Glo HD. But, the Glo was around a long time before these larger cards and might not have full hardware support for it. As I said, people had issues with some 32GB cards in the Glo. And as you are having problems, and you have made this change, it is something you need to check.



I can state that my Glo is working with all of the recent firmware versions. It only has a 2GB card in it. That is not the original as that failed some time ago. I only use it for testing, but, it is working correctly.



As to a corrupt database, the symptoms you describe are symptoms of a corrupt database. What happens when the database becomes corrupt depends on exactly how it is corrupt. The device can run for a long time with a corrupt database, it just can't save changes such as reading position. These will be lost either when the device is restarted or as more books are opened.

And I can say it works.

This Show how to use
https://www.mobileread.com/forums/sh...d.php?t=309576

And this too
https://www.reddit.com/r/kobo/commen...m_source=share

I have the 128gb works fine after factory reset.
Perhaps the error Was through the Partition extend.

The only 2 errors was the 4 last reading books in the home screen and the stand by picture of the reading book.

I had 64gb intern more then a year without probs and now the 128gb which works without problems after the reset


Gesendet von meinem SM-G991B mit Tapatalk

[abikyoukan]

I sent a DM to the Kobo Help team on twitter about this and is supposedly with the development team but decided to post here just in case someone sees this. Worth a try.

I use my device in Japanese and when using my kobo with PDF files, when using a dictionary in words, I cant select the whole word. Long pressing a word, only shows me the definition of one character and not the entire word.

This works with epubs books normally because I can select the word appropriately, however with pdf it doesn't give me an option to select the size of the word. Following is a quick video demonstrating this.
With epubs, regardless of the orientation, I can select the size of the word appropriately. As seen here

Would love to see this fixed since even though I used a PDF file with pictures and selectable text, it doesn't work with a pdf file with pure text.

Thanks in advance and sorry for my poor english.

[davidfor]

Originally Posted by abikyoukan

I sent a DM to the Kobo Help team on twitter about this and is supposedly with the development team but decided to post here just in case someone sees this. Worth a try.

I use my device in Japanese and when using my kobo with PDF files, when using a dictionary in words, I cant select the whole word. Long pressing a word, only shows me the definition of one character and not the entire word.

This works with epubs books normally because I can select the word appropriately, however with pdf it doesn't give me an option to select the size of the word. Following is a quick video demonstrating this.
With epubs, regardless of the orientation, I can select the size of the word appropriately. As seen here

Would love to see this fixed since even though I used a PDF file with pictures and selectable text, it doesn't work with a pdf file with pure text.

Thanks in advance and sorry for my poor english.

See my response to your other post.