Hello Guest, if you are reading this it means you have not registered yet. Please take a second, Click here to register, and in a few simple steps you will be able to enjoy our community and use our OpenViX support section.

View Entry Info: Enigma2 Timeshift Bug

Category:
Possible Bug
What ViX Image build number are you using?
Please provide your ViX Team image build number. Menu > Information > About > Build number > ENTER THIS NUMBER e.g. 4.2.028
5.2.034
Have you tried a flash WITHOUT settings restore?
Have you tried this? PLEASE SELECT YES OR NO.
No
Have you tried a flash WITH settings restore?
Have you tried this? PLEASE SELECT YES OR NO.
No
Attachments
Results 1 to 11 of 11

Thread: Enigma2 Timeshift Bug

Hybrid View

Previous Post Previous Post   Next Post Next Post
  1. #1

    Title
    Junior Member
    Join Date
    Apr 2014
    Posts
    6
    Thanks
    0
    Thanked 6 Times in 3 Posts

    Enigma2 Timeshift Bug

    Hi there,

    Am fairly confident I've uncovered a longstanding bug (regression) with time shift that should be a fairly easy fix. I've proved it by downloading the enigma-src and adding further debug to the code.

    I say regression because the part of the code in question suspiciously has a commented line with a different check commented out to be replaced by something that definitely isn't working.

    If you trawl the forums you'll often run up against people complaining about the error message that they have to format their drive with ext2/ext3 to support hard links and sure most of them probably screwed up but there are a minority like me that have done things correctly but have been affected by this bug which makes them feel they've not formatted their drive correctly when in reality it's a totally misleading error message. So...

    In:

    https://github.com/OpenViX/enigma2/b...s/Timeshift.py

    At about line 884:

    # if filename.startswith("timeshift") and not os.path.splitext(filename)[1]:
    if filename.startswith("timeshift") and not filename.endswith(".sc") and not filename.endswith(".del") and not filename.endswith(".copy"):

    We are effectively checking to make sure we don't enter this part of the code for certain file extensions but what is happening is we are not checking for ".ap" files and for some reason (of which I'm not quite sure), sometimes on rare occasion one of those exists and is being processed here. So that later when we try to do:

    os.link("%s%s.sc" % (config.usage.timeshift_path.value,filename), "%spts_livebuffer_%s.sc" % (config.usage.timeshift_path.value,self.pts_eventc ount))

    .. with an .ap file it is actually trying to hard link a "ap.sc" file which will never exist! (It's either .ap or .sc) causing the os.link to fail and the exception to think that it just can't make a hard link because of the drive format whereas in reality, it shouldn't be processing this .ap file.

    The fix seems simple (but maybe implemented better by someone more familiar with this code), to add 'and not filename.endswith(".ap")' to the end of the 'if' statement at the top and to add a delete for old '.ap' files with all those other deletes just below it.

    Hope this helps! This is working well for me and should help people out with the occasional error about hard links and ext3 formats. Happy to chat over.

    Cheers.
    Last edited by vampyrebat; 25-03-19 at 23:55.

  2. #2
    abu baniaz's Avatar
    Title
    Moderator
    Join Date
    Sep 2010
    Location
    East London
    Posts
    23,335
    Thanks
    6,421
    Thanked 9,146 Times in 6,224 Posts
    Can you attach your modified file please?

  3. #3

    Title
    Junior Member
    Join Date
    Apr 2014
    Posts
    6
    Thanks
    0
    Thanked 6 Times in 3 Posts
    No problem! Please find attached. Includes the debug lines proving the problem the output of which is here:

    < 207.559> hl1a /media/hdd/timeshift/timeshift.yiRdFW.ap
    < 207.559> hl1b /media/hdd/timeshift/pts_livebuffer_1
    < 207.559> hl2a /media/hdd/timeshift/timeshift.yiRdFW.ap.sc
    < 207.559> hl2b /media/hdd/timeshift/pts_livebuffer_1.sc
    < 207.564> [Skin] processing screen MessageBoxSimple:
    < 207.569> [GUISkin] warning, skin is missing element autoresize in <class 'Screens.MessageBox.MessageBox'>(Creating Hardlink to Timeshift file failed!
    The Filesystem on your Timeshift-Device does not support hardlinks.
    Make sure it is formatted in EXT2 or EXT3!

    [Errno 2] No such file or directory)< 207.569>

    Obviously that debug output is prior to adding the check for .ap files. I hope this helps!
    Attached Files Attached Files
    Last edited by vampyrebat; 26-03-19 at 00:19.

  4. The Following 3 Users Say Thank You to vampyrebat For This Useful Post:

    abu baniaz (26-03-19),Clabs (28-03-19),Valiant (26-03-19)

  5. #4

    Title
    Junior Member
    Join Date
    Apr 2014
    Posts
    6
    Thanks
    0
    Thanked 6 Times in 3 Posts
    To prevent enigma2 trying to hard link .ap files line 885:

    if filename.startswith("timeshift") and not filename.endswith(".sc") and not filename.endswith(".del") and not filename.endswith(".copy"):

    Should probably become:

    if filename.startswith("timeshift") and filename.count('.') == 1:

    All seems to be working well for me now but maybe someone with a better understanding of that code can verify it for me and hopefully fix it for a future release.

    Thank you!

  6. #5
    birdman's Avatar
    Title
    Moderator
    Join Date
    Sep 2014
    Location
    Hitchin, UK
    Posts
    7,769
    Thanks
    235
    Thanked 1,656 Times in 1,305 Posts
    Quote Originally Posted by vampyrebat View Post
    if filename.startswith("timeshift") and filename.count('.') == 1:
    Which would appear to be checking the number of "." chars in the filename.
    From your analysis it seem that you really just want to know whether the file actually exists, so better to check for that rather than some indirect concept of why it might not exist.

    So,
    Code:
            if filename.startswith("timeshift") and os.path.isfile(filename):
    MiracleBox Prem Twin HD - 2@DVB-T2 + Xtrend et8000 - 5(incl. 2 different USBs)@DVB-T2[terrestrial - UK Freeview HD, Sandy Heath] - LAN/USB-stick/HDD

  7. #6

    Title
    Junior Member
    Join Date
    Apr 2014
    Posts
    6
    Thanks
    0
    Thanked 6 Times in 3 Posts
    Hi Birdman!

    It's a little different from that. I'm checking that I get a file without two extensions.

    i.e that block of code needs to process:

    timeshift.yiRdFW

    but not:

    timeshift.yiRdFW.ap or timeshift.yiRdFW.sc

    If it processes the .ap it will throw an exception. The original bug. I was toying with the idea of checking the filename from the dvb stream processor but I don't want to get my hands too dirty yet before I understand more of the system, I'm just a noobie

    Hope that explanation helps!

  8. The Following 2 Users Say Thank You to vampyrebat For This Useful Post:

    birdman (28-03-19),Clabs (28-03-19)

  9. #7
    birdman's Avatar
    Title
    Moderator
    Join Date
    Sep 2014
    Location
    Hitchin, UK
    Posts
    7,769
    Thanks
    235
    Thanked 1,656 Times in 1,305 Posts
    A fix has been added (to add ".ap" to the check list):
    Code:
     https://github.com/OpenViX/enigma2/commit/7105f6d7b2107b2ef4367531a0a3533faaf021ad
    However, such a convoluted test strikes me as indicating there is something fundamentally wrong elsewhere.
    MiracleBox Prem Twin HD - 2@DVB-T2 + Xtrend et8000 - 5(incl. 2 different USBs)@DVB-T2[terrestrial - UK Freeview HD, Sandy Heath] - LAN/USB-stick/HDD

  10. #8

    Title
    Junior Member
    Join Date
    Apr 2014
    Posts
    6
    Thanks
    0
    Thanked 6 Times in 3 Posts
    Thanks for the reply. I made that commit last night. There's now an open PR. I opted for adding just the .ap check for clarity and I know that definitely fixes the original bug I was encountering. I completely agree with what you say. This bug might even just be the side effect of a different bug. I wish I understood the code a bit better so I could help clean it up a bit, there are almost certainly other solution opportunities here. I might try on the weekend.

  11. #9
    abu baniaz's Avatar
    Title
    Moderator
    Join Date
    Sep 2010
    Location
    East London
    Posts
    23,335
    Thanks
    6,421
    Thanked 9,146 Times in 6,224 Posts

    Enigma2 Timeshift Bug

    Can you please submit any further pull requests to the Dev branch?

  12. #10

    Title
    Junior Member
    Join Date
    Apr 2014
    Posts
    6
    Thanks
    0
    Thanked 6 Times in 3 Posts
    Quote Originally Posted by abu baniaz View Post
    Can you please submit any further pull requests to the Dev branch?
    Makes sense! Sure thing.

  13. The Following User Says Thank You to vampyrebat For This Useful Post:

    abu baniaz (30-03-19)

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
This website uses cookies
We use cookies to store session information to facilitate remembering your login information, to allow you to save website preferences, to personalise content and ads, to provide social media features and to analyse our traffic. We also share information about your use of our site with our social media, advertising and analytics partners.