Constant Crashing Using Webif

[Bangord30]

Possibly fixed by this (going solely by the description text):

https://github.com/oe-alliance/e2ope...63fb4b025d8657

Guys, how in plain English when there has been a pull request(?) or a file amended on github, do you amend this file on your own box? I can never see a simple download file button on github?! Maybe I'm missing but always looks confusing to me

[ccs]

You can "view" the .py file and then there is a "raw" button, which you can use to see the whole file without numbering, and you can right click raw and "save target as", or select all and copy and paste.

I'm sure this isn't the way, but it does work.

[birdman]

I'm one of the devs, of course I have it already

So post the crash log.

[Ev0]

So post the crash log.

Here is mine (different box and image, but i'm sure the crash will be the same)

[birdman]

Here is mine (different box and image, but i'm sure the crash will be the same)

That log just contains garbage characters once it comes to the log part:

xb9\x10\r\x01U<\x01\xc1\xf5\xadc\xdci&l\xfc\x17\xb bd\xd1/\xf4\xb9\x0e\x05\x9d\xfbm\xcf\xbeMup\xabo\x00\x7f\ xcbG$\xfbW\x0b\xf0.....

so nothing useful.
It could be a memory overwrite within enigma2, or a damaged file-system on your debug log storage device.

[Ev0]

That log just contains garbage characters once it comes to the log part:

so nothing useful.
It could be a memory overwrite within enigma2, or a damaged file-system on your debug log storage device.

It's the same garbage that filled my screen when it crashed

[Bangord30]

I sent mines to Vix through the box if helps? It gave me a ref no to quote. If I can find it again?! 😉

Sent from my VF695 using Tapatalk

[SpaceRat]

It's the screenshot

Python is well known for useless error messages, here you can see that something about the hex string (single quoted) containing the screenshot failed.

Nothing OWIF can do about it.

You just need Betacentauri's "Allow NUL characters in Console" and MiLo's "buffer and setbuffer" changes in enigma2.
Plus "Fix compilation of OpenPLi4/mipsel ..." .

OpenViX has all these changes and so do OpenATV, OpenHDF and OpenBH.

Gesendet von meinem Siemens C25 mit Tapatalk

[SpaceRat]

That log just contains garbage characters once it comes to the log part:

Not really garbage, when you look at the end of the line ...

\xb9\lots_of_more_hex_numbers\xd5|\x89\x1e\x8d\x0b',)) failed)

It might look like garbage in the first place, but using "\x" is just many programming language's way of indicating that the following characters represent a hex byte inside a string rather than their literal character, just like \n represents a line feed ("new line") and \r represents a (carriage) "return".
\x is used to process, log or show (assumed or truely) non-printable characters.

In fact, \x0a (ASCII char 10 dec or 0a hex) could also be used to represent a line feed and \x0d (ASCII char 13 dec or 0d hex) could also be used to represent a carriage return, \n and \r are just convenience shortcuts to it.

Well, as I sense that's interesting but doesn't help, back to the log:
It doesn't help either.
Python logs help shit, they never go back far enough to log the real error and this is true for this log too.
We only see that "blabla',))" failed, so we can take a wild guess and learn that some call - which was not logged, it would be somewhere before those hex nums - containing the picture data (Those hex nums) as single-quoted (We can see the ' ) failed.
Thanks Python, that helps.

All I can say is, that OpenBH does contain the necessary changes for the new screenshot routine in OWIF:

[SpaceRat]

I sent my log to WOS few nights ago as my duo 2 has been crashing for a good few weeks now(every update done as and when) and I've been reflashing but upon reading the log it's always something to do with openwebif. Closed it down on my pc and not used it for a few nights and box been fine.

Sending them to /dev/null would have had the same effect

No OpenViX dev is actively participating in OWIF development.
If you experience any OWIF problems, create a ticket on the OWIF development site https://github.com/E2OpenPlugins/e2o...enWebif/issues or at least post publically, so that I eventually become aware of it.
A ticket (issue) on github is the preferred - as in "most reliable and fastest" - way though, as I'm by far neither the most active nor the best OWIF developer but just the only one reading here.

I can fix small problems, try to solve PEBKAC problems, but in many cases I will just open an issue myself

[birdman]

Nothing OWIF can do about it.

Not about the text, but there hopefully is something it can do about why the error occurred. That something is what we are looking for....

You just need Betacentauri's "Allow NUL characters in Console" and MiLo's "buffer and setbuffer" changes in enigma2.
Plus "Fix compilation of OpenPLi4/mipsel ..." .

Converting the string into binary isn't any great problem (a simple Perl one-liner probably), but still the result won't be of much use given that the crash log only has a 16kB buffer.

OpenViX has all these changes and so do OpenATV, OpenHDF and OpenBH.

So Vix already has the fixes you reckon need to be added?

[SpaceRat]

So Vix already has the fixes you reckon need to be added?

Yes, they are contained within their enigma2 git, that's why I do not understand that there are any problems.

If you compile what's visible in the ViX or OpenBH gits, you definitely should get a working combination of OWIF and enigma2 ...
... unless of course teams freeze enigma2 at a certain SRCREV on the one hand and continue to build "latest" for everything else, which would be plain stupid.

[Rob van der Does]

Yes, they are contained within their enigma2 git, that's why I do not understand that there are any problems.

The youngest release image (4.2.026) is from before that commit date.
And OP is using an even (much) older version (4.2.012).

[SpaceRat]

But then they shouldn't have an OWIF with piping for the screenshots either ...

[Rob van der Does]

Yep, correct.