Frustration with development pace
#21
(10-07-2020, 04:17 PM)natasha Wrote: I am happy to hear that one or two users are brave enough to use PP as a daily phone.

However, I am with the majority of comments. I bought PP in July, and for the first months, I have been updating the phone with Mobian every 2-4 days. Now it is on my desk, resting there...I do not think I am going to update until a few weeks or months... If I saw any improvement, I would keep updating, playing, etc. but I do not see any movement. We have Daily releases: Some apps work, next day they broke, next they are up again, ... but nothing really changes.  At this pace, probably it will never make it. People will get tired before it can make it.

I am with the idea of crowdfunding to hire developers. That is a great idea and might be the only way. I will also contribute as much I can (mainly if I see progress)
I am using the PinePhone as a daily driver. It's a pain and I have to apologize for its screwups often, but I figured it was good to eat my own dog food for the parts I do work on and for bug reports for the parts I don't.

Battery life is still terrible, but thankfully I'm a hermit so it doesn't much matter. It is a serious problem when I do go out for extended times though.

I don't think a dev for a specific OS is a good idea right now. I think what we need right now is infrastructure, e.g. more kernel development which is used by almost all PinePhone OSes, and a distro-agnostic calls and messages app. I'm very, very slowly working on one in GTK3 and Rust. I don't have a lot of free time.

We also need more work on ModemManager, because it seems to dislike the quirks of the Quectel modem, and it seems to be the cause of calls becoming uncontrollable but still active (can't hang up, window disappears). I get the feeling ModemManager isn't widely used for anything but mobile broadband stuff, so the code paths for phone and SMS aren't well tested.

Bless his soul, megi has done a lot in the kernel, more than anyone else, (we'd really be boned without him), but it's clear he's only one guy working in his free time, and it's not enough. He's also not using it as a daily driver, which of course would shield him from the regressions that seem to sprout up like weeds with every commit.

We really need a full-time kernel developer. Not someone for Sailfish or Mobian or UBports, but someone to work on the linux kernel, along with bits and bobs in other troublesome system components.

(10-07-2020, 05:57 AM)jed Wrote: As I understand it, all of the PinePhone specific development is done by volunteers. Instead of rubbishing how quickly those volunteers are progressing the software in their free time, you could contribute.
Thank you for parroting the same unhelpful garbage I've seen written in other threads I lurked on that dared to question the speed of development. I'm not criticizing the contributors, and if you read my post you'd know I have already touched on the issue of why it might be taking time. As far as contributing, I already have and I am. This is not a normal or acceptable pace of progress for a project of PinePhone's stature.
  Reply
#22
It occurs to me, we could donate to megi to try and get him able to afford to work on the PinePhone near full time. He has a contribute page on his site:
https://xnux.eu/contribute.html
  Reply
#23
(10-07-2020, 04:35 AM)fsflover Wrote: Thank you for starting this topic, it's an important thing to discuss. I hope someone from the development community describes the current situation and problems and maybe explains what has to be done and how long it might take.
I second this.  Hope to hear from the Devs.

They have the most realistic idea of the challenges, possible timelines and very importantly the resources needed to meet the challenges in perfecting Pinephone's basic
functionality.

Personally, my $350 plus investment in my pinephones (BH and PMOS conv) would be considered well spent if industry-standard quality basic calls and texting (SMS & MMS) as well as more reasonable battery life were realized and were rock solid.

How much funding would be required from the community and Pine64?  The Devs might have some idea.

I myself am not overly frustrated at the present state of affairs but I agree that with more resources applied, we should be able to do better and fear that frustration will only grow if this is not addressed and that is not good for this wonderful Pine64 experiment.
  Reply
#24
(10-07-2020, 02:53 PM)lot378 Wrote: Please provide a breakdown of the issues encountered with EG25-G for the benefit of the community because they/we are all interested.

Reading through https://xnux.eu/devices/feature/modem-pp-reveng.html already provides a good explanation: there are at least some developers at Quectel who are very poorly skilled and don't really know what they're doing (based on the unreasonable amount of system() in the firmware code).

On top of that, the latest 2 firmware releases both state (pages 8+11) that they "Solved the problem of USB disconnection" (spoiler: nope, is still there).
Because yes, the modem issues (missed call, no more mobile data...) we have when resuming from sleep are entirely due to the modem resetting the USB connection on its side and its QMI interface vanishing (AT commands still work as they use another interface, but only for calls and SMS).

The problem isn't that we don't understand how the modem works, but that (for now) it just doesn't work as it should. They're obviously aware of the problem and seem to be working on it, so there's still hope that modem issues will be solved within a few months (years?).
  Reply
#25
(10-07-2020, 02:23 AM)firefox-58 Wrote: ATM I am using with joy my other Linux based smartphone: A SONY Xperia X running stable and flawlessly on Sailfish OS. Calls and SMS work, Caldav/Carddav works, ActiveSync works, Camera works, Browser and internet are working and the battery lasts for 2-3 days.

Hello fellow sailor Smile

While I think we can all agree that progress on call functionality is frustrating I think it's worth commenting that the comparison to Sailfish OS has a number of issues:

1. Sailfish is backed by a commercial company founded by ex-Nokia developers, has paying personal customers and corporate clients and launched it's first phone in 2013 (7 years ago).
2. It's UI (which is awesome) is closed source.
3. Saying the browser works is a stretch, it's only just got an updated engine in the 3.4 early access release and still has stability issues (ref: https://forum.sailfishos.org/t/browser-p...ose/2590/8). Prior to this update it didn't work with many sites due to the outdated engine, it's still not great.
4. You are fortunate to have an Xperia X, mine died (hardware), I replaced with XA2 on which *phone calls* are plagued by a bug (no audio in either direction) which has been known about across *4 releases* (ref: https://forum.sailfishos.org/t/3-4-0-22-...io/2446/18). I have used Sailfish as a daily driver for over two years and unless this is fixed ASAP I'm going to have to ditch :/

I love Sailfish too but it's really not a good comparison to the PinePhone project which is 100% open source community driven and only 6 months ish of development in vs the many years that Sailfish has been in development (not to mention it's precursor projects in Nokia itself).

I just felt it was worth pointing out Smile I fully agree that it would be great to:

- Get an update from devs on the issues surrounding the modem functionalities
- Get told where to direct donations to accelerate the work on this area (like others I'm more than happy to funnel some cash to this cause!)
  Reply
#26
(10-08-2020, 03:01 AM)a-wai Wrote:
(10-07-2020, 02:53 PM)lot378 Wrote: Please provide a breakdown of the issues encountered with EG25-G for the benefit of the community because they/we are all interested.

Reading through https://xnux.eu/devices/feature/modem-pp-reveng.html already provides a good explanation: there are at least some developers at Quectel who are very poorly skilled and don't really know what they're doing (based on the unreasonable amount of system() in the firmware code).

On top of that, the latest 2 firmware releases both state (pages 8+11) that they "Solved the problem of USB disconnection" (spoiler: nope, is still there).
Because yes, the modem issues (missed call, no more mobile data...) we have when resuming from sleep are entirely due to the modem resetting the USB connection on its side and its QMI interface vanishing (AT commands still work as they use another interface, but only for calls and SMS).

The problem isn't that we don't understand how the modem works, but that (for now) it just doesn't work as it should. They're obviously aware of the problem and seem to be working on it, so there's still hope that modem issues will be solved within a few months (years?).

Thank you. I'd agree that's some pretty lazy programming to use system() calls to echo values in to files (besides the security risks that are obviously inherent). It seems to me Pine64 need to appoint a project manager to get the Quectel issues recognised resolved tested deployed to updated firmware, otherwise expect them to just roll around making the same mistakes like its groundhog day.
  Reply
#27
(10-08-2020, 03:01 AM)a-wai Wrote: Reading through https://xnux.eu/devices/feature/modem-pp-reveng.html already provides a good explanation: there are at least some developers at Quectel who are very poorly skilled and don't really know what they're doing (based on the unreasonable amount of system() in the firmware code).
I read some of the source for their QFirehose updater utility prior to running it, and that seriously made me reconsider updating my modem, as I wasn't sure I could trust it to do the job and not brick the modem. Miraculously, it worked as intended. But yes, Quectel seems to have some people who definitely do NOT belong in systems programming working on their modem firmware, which is horrifying. If they'd open source their firmware (when hell freezes over), I'm sure the community could quickly fix at least some of these issues.
  Reply
#28
(10-07-2020, 03:33 PM)bcnaz Wrote: occasionally it will quit and go back to the log-in screen. That can be frustrating, yes
  but I just sign back in and continue texting.

The audio/talk works on some Distro's now,  good enough to use,   though quality is still improving.

For data,  I prefer using hot spot to my laptop over the web browsers for sure, but some of the on phone browsers are getting better.
 YES, the data works, at least for me it does.  Using with both AT&T and T-Mobile carriers. (GSM)
  The phone/modem did not see any of the 4 different Verizon sim cards that I have tried.

Not sure how Mobian can work on your Braveheart, but does not work on my UBIports CE phone. Mobian often logs out and I cannot log in back. When I enter the password, it asks for the password again, forever. Only reboot helps.

My SIM card is recognized without issues and data works, but after some time it suddently stops working until I reboot. After reboot, I have to wait for maybe 5 minutes before it finds the data connection again. It also goes to "airplane mode" all the time by itself and when I switch it off, it switches on Bluetooth for no reason.

Firefox suddenly stopped fitting the screen when in vertical mode (horizontal mode works fine, but keyboard takes 2/3 of space). Telegram only works fine in vertical mode, so I have to rotate the screen all the time. The horizontal mode rotates differently every time, so I have to adjust myself.

Calls work, but can the phone reliably wake up from suspend when it gets a call? When charging, the phone does not switch off the screen automatically and even sometimes switches it on by itself.

It's all frustrating. Many of these issues look quite simple and important for everyday use. I am not asking to fix them ASAP for free for me. I am asking what are the actual problems, what is to be done and how long it could take at all.
  Reply
#29
(10-08-2020, 05:23 AM)fsflover Wrote:
(10-07-2020, 03:33 PM)bcnaz Wrote: occasionally it will quit and go back to the log-in screen. That can be frustrating, yes
  but I just sign back in and continue texting.

The audio/talk works on some Distro's now,  good enough to use,   though quality is still improving.

For data,  I prefer using hot spot to my laptop over the web browsers for sure, but some of the on phone browsers are getting better.
 YES, the data works, at least for me it does.  Using with both AT&T and T-Mobile carriers. (GSM)
  The phone/modem did not see any of the 4 different Verizon sim cards that I have tried.

Not sure how Mobian can work on your Braveheart, but does not work on my UBIports CE phone. Mobian often logs out and I cannot log in back. When I enter the password, it asks for the password again, forever. Only reboot helps.

My SIM card is recognized without issues and data works, but after some time it suddently stops working until I reboot. After reboot, I have to wait for maybe 5 minutes before it finds the data connection again. It also goes to "airplane mode" all the time by itself and when I switch it off, it switches on Bluetooth for no reason.

Firefox suddenly stopped fitting the screen when in vertical mode (horizontal mode works fine, but keyboard takes 2/3 of space). Telegram only work fine in vertical mode, so I have to rotate the screen all the time. The horizontal mode rotates differently every time, so I have to adjust myself.

Calls work, but can the phone reliably wake up from suspend when it gets a call? When charging, the phone does not switch off the screen automatically and even sometimes switches it on by itself.

It's all frustrating. Many of these issues look quite simple and important for everyday use. I am not asking to fix them ASAP for free for me. I am asking what are the actual problems, what is to be done and how long it could take at all.

The problems seem to be GRAVE...... no implementation of PHOSH runs properly. Only in plasma-mobile incoming calls are ringing also when the phone sleeps. But you cannot accept the call.. The lockscreen brings up nothing that looks as an phone interface. Sou You cannot receive/take the call. This bug is old but nobody can or will fix it.

Maybe the networkmanager of KDE is some more sophisticated. But plasma-mobile itself is far far from being usable.

So NONE of the available operatings systems on the PinePhone work usable. And this after a lot of months. I fear that the modem will never be functional. And it is NO excuse that the devs for the modem driver need backward disassembling. I think Pine has bought the modem chips and not stolen it. So I cannot imagine that there will be no documentation available for it. If there are firmware bugs they must also appear in the products of other customers of the modem producer. They also would need an firmware update if so.

Sometimes I think that most devs are proud about what they habe newly implemented in various builds. But have forgotten that this is a at first  a PHONE
  Reply
#30
(10-08-2020, 04:19 AM)lot378 Wrote:
(10-08-2020, 03:01 AM)a-wai Wrote:
(10-07-2020, 02:53 PM)lot378 Wrote: Please provide a breakdown of the issues encountered with EG25-G for the benefit of the community because they/we are all interested.

Reading through https://xnux.eu/devices/feature/modem-pp-reveng.html already provides a good explanation: there are at least some developers at Quectel who are very poorly skilled and don't really know what they're doing (based on the unreasonable amount of system() in the firmware code).

On top of that, the latest 2 firmware releases both state (pages 8+11) that they "Solved the problem of USB disconnection" (spoiler: nope, is still there).
Because yes, the modem issues (missed call, no more mobile data...) we have when resuming from sleep are entirely due to the modem resetting the USB connection on its side and its QMI interface vanishing (AT commands still work as they use another interface, but only for calls and SMS).

The problem isn't that we don't understand how the modem works, but that (for now) it just doesn't work as it should. They're obviously aware of the problem and seem to be working on it, so there's still hope that modem issues will be solved within a few months (years?).

Thank you. I'd agree that's some pretty lazy programming to use system() calls to echo values in to files (besides the security risks that are obviously inherent). It seems to me Pine64 need to appoint a project manager to get the Quectel issues recognised resolved tested deployed to updated firmware, otherwise expect them to just roll around making the same mistakes like its groundhog day.
Piggybacking on this:

If as believed by the Devs, the core firmware of the EG25-G is fundamentally flawed, then it would appear that it's time for Pine64 to depart from its volunteer community based software development at least as it pertains to this issue.

The powers-that-be at Pine should attempt to address the modem's problems with Quectel and obtain some type of solid NDA agreement of cooperation and information sharing, and then commission a qualified developer to begin fixing the EG25 problems.  Seems like its worth a try anyway.
  Reply


Possibly Related Threads…
Thread Author Replies Views Last Post
  PinePhone app development WhiteHexagon 14 3,548 05-17-2023, 10:41 PM
Last Post: Kevin Kofler
  App Development rkveale 4 4,085 05-16-2021, 09:43 AM
Last Post: axel
Question Where are the development discussions at? mnjm97 2 3,701 08-28-2020, 01:19 AM
Last Post: rocket2nfinity

Forum Jump:


Users browsing this thread: 2 Guest(s)