martinwguy said
Hi

  I only have a few kb/second in upload here. It'll take weeks to upload them all. There's be a higher data rate sending you a CD in the post! Can you be more selective, ask me for specific files out of the filesystems or info that will help you build a working system?

  You *do* build your images from the source code, don't you, as it sounds like you might be assembling them from binary bits and pieces, which is fine for your own personal use, but breaks the GPL if you publish it, since you have to distribute or make available the source code that corresponds to the binaries of any GPL'd objects that you publish... all in the spirit of allowing others to modify them, fix bugs, make enhancements, see how they work, check there are no trojans etc.

 

    M

Specifically, I would require the following:

-rw-r--r-- 1 root root   16777216 Sep 23 19:43 boot.img
-rw-r--r-- 1 root root    8388608 Sep 23 19:43 kernel.img
-rw-r--r-- 1 root root    4194304 Sep 23 19:42 misc.img
-rw-r--r-- 1 root root   33554432 Sep 23 19:44 recovery.img
-rw-r--r-- 1 root root  536870912 Sep 23 20:02 system.img

It's probably between 200 and 300 MB once packed with a good archive software setup with max compression. The Linux kernel and Android source code (with their modifications) for devices such as the Cube U23GT aren't made available by the manufacturer, so we, as most sites dealing with custom firmwares work on binaries only.

As for trojan risks/security, our website is around and produce custom firmware modifications since end of 2011, we have a reasonably large trusted user base and we are doing our best to produce quality and transparent work. Anyone highly concerned about security shouldn't probably Chinese tablets anyway.

Cheers.

(my answers in bold)

martinwguy said

> Specifically, I would require the following:

> -rw-r--r-- 1 root root   16777216 Sep 23 19:43 boot.img

> -rw-r--r-- 1 root root    8388608 Sep 23 19:43 kernel.img

> -rw-r--r-- 1 root root    4194304 Sep 23 19:42 misc.img

> -rw-r--r-- 1 root root   33554432 Sep 23 19:44 recovery.img

> -rw-r--r-- 1 root root  536870912 Sep 23 20:02 system.img

>

> It's probably between 200 and 300 MB once packed with a good archive

> software setup with max compression.

OK, it's dribbling up into http://martinwguy.co.uk/martin/opinions-on-teclast-p90hdch/51cube/U23GTC4/

Good luck!

I got the files, if no one else need them, feel free to delete the files. Next days should be quite busy here, but I should have a rebuilt firmware ready by the week-end.

> The Linux kernel and Android source

> code (with their modifications) for devices such as the Cube U23GT aren't

> made available by the manufacturer

Then they're breaking the law.  The Linux kernel is under the GPL. If

they distribute binaries of the Linux kernel, they MUST distribute the

full modified source code.

You can contact them about this and request it. If they don't satisfy,

the next port of call is license-violation@gpl-violations.org

  If Android uses busybox, you can Cc: gpl@busybox.net

 These are the FSF lawyers basically, acting on Linus' behalf (don't

image that he personally has time for such rubbish!) - but one way for

young lawyers to get fame and experience is to do pro-bono work on

cases like this. If they win, so much the better for all of us.

See http://www.itproportal.com/2011/08/16/google-faces-gpl-breach-claims-over-android-code/

Thanks for links. I think Most Chinese vendors are actually violating the GPL. I've read on different forums, including ours that some people filed requests there and it looks like manufacturers did not really care. I don't know if the cases have been seriously followed and haven't yet found any detailed report about these kind of things. Interestingly, there are also cases of manufacturers releasing source code that doesn't work or compile, we have a story here for the ARCHOS GamePad Kernel source code.

> so we, as most sites dealing with custom

> firmwares work on binaries only.

I applaud your work in the face of manufacturer closedness, but the

manufacturers are breaking the law and making you incapable of fixing

the software you use.

I know, that's sad. Luckily, there are some talented developers like Crew RKTablets who manage to (re)build Kernels and/or Android system for a few devices, without the help of the manufacturer. There are probably some Rockchip or other vendor binary blobs here and there in the firmwares, but it's better than nothing.

> As for trojan risks/security, our website is around and produce custom

> firmware modifications since end of 2011, we have quite a few trusted users

> are doing our best to produce quality and transparent work.

Of course.

> Anyone highly

> concerned about security shouldn't probably Chinese tablets anyway.

That's why I never put a useful password into it, but just created

a pippopoppo@gmail.com to use with it.

However, if there is the source code, you can check each package

against the official release of that package and see changes, then

compile it and check if the resulting binary is the same as what the

manufacturer is distributing.  It's a different world, believe me, and

it's what the GPL is all about.

Can't agree more, unfortunately, we're still far from being there yet with tablets from Chinese manufacturers 

Cheers.

If Cube's U23GTC4-W stock update works with the device that had the WiFI issue, our custom firmware should work. 

We haven't yet rebuilt a custom firmware using Cube's update or martinwguy files, that should happen in a couple of days.

Cheers.

I've just posted our custom firmware for the Cube U23GTC4-W tablet, post #1 in current topic has been updated.

I've quickly checked martinwguy firmware dump, it seems to be designed for the U23GTC4, not the U23GTC4-W. This kernel is more recent than current official firmware offered by Cube for download on their website. I'll post a rebuilt firmware based on martinwguy dump if next official update from Cube takes too long to come.

Cheers.

Using that firmware should get you fixed.

Cheers.