Menu ▾ ▴

#1487 Linux Task Coach crashes when iphone attempts to synchronize

closed
nobody
None
linux
2026-01-20
2014-01-20
Jim Daniels
No

OS is Linux Mint 14 x64 running on Lenovo X61 laptop on the same wifi network.
Once I figured out that I have to have Task Coach running on the linux box (may want to add that to the documentation with a step-by-step because it took so long to get that far), I found that attempts from an iphone 3 with an old iOS and the latest Task Coach caused the linux box to fail in unpredictable ways.

On one occasion, my linux box refused to accept mouse or keyboard input after I attemtped to sync. I reset the linux box by putting it to sleep with the power switch and re-awaking. The latest crash popped up an error window on the the linux box, then switched to a "Warning - kWin" dialog saying "the window "Error" is not responding. It bongs to the application python (Process ID =15873, hostname = localhost). Do you wish to terminate this application, etc., etc."

So it looks like even the error window in Task Coach has bugs.

On previous occasions, an error window appeared on the iphone. Both types of error are fatal for the linux side and not fatal for the iphone side.

The two cases are the following:
1) I start the iphone app, then start the linux application, then sync. This crashes the linux application with errors.
2) I start the iphone app, tell it to sync, it can't find anything so waits. I then start the linux application; the iphone now sees the linux box and I try to sync. This causes immediate linux crash without an error window.

On the bright side, the iphone app is rock solid. :(

Discussion

  • Jérome Laheurte

    The iPhone 3 shouldn't even be supported, I guess this is the new App Store "feature" where you can download older versions still compatible with your iOS... Could you find out the version of the iPhone app ?

     
  • Jim Daniels

    Jim Daniels - 2014-01-22

    Version 3.4.4.

    Why would you kill the iphone3 compatibility? All features on the standalone app that I've tried work perfectly on the 3. Also, it's the only reasonably priced unlocked international phone I could get. I'm waiting for the iphone4 unlocked market to drop in price before upgrading. I'm sure you know how the US has been treating phone unlocking.

     
  • Jérome Laheurte

    Nope I was wrong, that's the latest version. As a side note, keeping compatibility with older phones is a royal PITA. XCode updates don't make it easy, so each time I have to make a release I spend a day or two wondering why it won't build for older archs...

    Anyway, that means it uses the latest sync protocol, which makes me wonder what's happening. I sync with a Linux PC on a daily basis and never had those kind of problems. Yet another distro to install for testing purposes :)

     
  • Jérome Laheurte

    BTW are you using a particular flavor of Mint (XFCE, Gnome or KDE) ?

     
  • Jim Daniels

    Jim Daniels - 2014-01-25

    Here are the details on my install:
    Linux Mint KDE
    uname -a
    Linux Occulus 3.5.0-25-generic #39-Ubuntu SMP Mon Feb 25 18:26:58 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux

    Output of lsmod:
    Module Size Used by
    hid_generic 12541 0
    usbhid 46987 0
    hid 100411 2 hid_generic,usbhid
    nls_iso8859_1 12714 0
    usb_storage 48839 0
    snd_hrtimer 12745 1
    nfnetlink_log 17846 0
    nfnetlink 14328 1 nfnetlink_log
    autofs4 36534 1
    snd_hda_codec_analog 97588 1
    arc4 12530 2
    snd_hda_intel 33492 2
    snd_hda_codec 134213 2 snd_hda_codec_analog,snd_hda_intel
    snd_hwdep 17699 1 snd_hda_codec
    snd_pcm 96668 2 snd_hda_intel,snd_hda_codec
    thinkpad_acpi 81199 0
    snd_seq_midi 13325 0
    rt2800usb 22662 0
    rt2800lib 58731 1 rt2800usb
    crc_ccitt 12708 1 rt2800lib
    rt2x00usb 20665 1 rt2800usb
    rt2x00lib 54939 3 rt2800usb,rt2800lib,rt2x00usb
    snd_rawmidi 30513 1 snd_seq_midi
    coretemp 13401 0
    mac80211 540032 3 rt2800lib,rt2x00usb,rt2x00lib
    snd_seq_midi_event 14900 1 snd_seq_midi
    pcmcia 49008 0
    snd_seq 61555 3 snd_seq_midi,snd_seq_midi_event
    cfg80211 206797 2 rt2x00lib,mac80211
    kvm 414071 0
    microcode 22804 0
    snd_timer 29426 3 snd_hrtimer,snd_pcm,snd_seq
    snd_seq_device 14498 3 snd_seq_midi,snd_rawmidi,snd_seq
    yenta_socket 31818 0
    pcmcia_rsrc 18289 1 yenta_socket
    psmouse 95595 0
    pcmcia_core 22570 3 pcmcia,yenta_socket,pcmcia_rsrc
    snd 78921 15 snd_hda_codec_analog,snd_hda_intel,snd_hda_codec,snd_hwdep,snd_pcm,thinkpad_acpi,snd_rawmidi,snd_seq,snd_timer,snd_seq_device
    lpc_ich 17062 0
    serio_raw 13216 0
    snd_page_alloc 18485 2 snd_hda_intel,snd_pcm
    soundcore 15048 1 snd
    nvram 14363 1 thinkpad_acpi
    tpm_tis 18681 0
    bnep 18141 2
    rfcomm 46620 4
    bluetooth 209249 10 bnep,rfcomm
    mac_hid 13206 0
    parport_pc 32689 0
    ppdev 17074 0
    binfmt_misc 17501 1
    nfsd 255927 2
    nfs 301262 0
    lp 17760 0
    parport 46346 3 parport_pc,ppdev,lp
    lockd 76909 2 nfsd,nfs
    fscache 61094 1 nfs
    auth_rpcgss 40956 2 nfsd,nfs
    nfs_acl 12838 2 nfsd,nfs
    sunrpc 229747 6 nfsd,nfs,lockd,auth_rpcgss,nfs_acl
    firewire_ohci 40402 0
    sdhci_pci 18617 0
    sdhci 32646 1 sdhci_pci
    firewire_core 64369 1 firewire_ohci
    crc_itu_t 12708 1 firewire_core
    i915 520615 3
    drm_kms_helper 49113 1 i915
    drm 288721 4 i915,drm_kms_helper
    i2c_algo_bit 13414 1 i915
    video 19391 1 i915
    e1000e 199273 0

     
  • Jim Daniels

    Jim Daniels - 2014-01-25

    Or you can just post a debug version with diagnositc output and have me send you the output. It might be something simple.

     
  • Jérome Laheurte

    We'll come to that eventually, but I already have a bunch of VMs for this kind of things :)

     
  • Jérome Laheurte

    Okay, works for me. Do you see anything if you reproduce this after launching Task Coach from the command line ?

     
  • Jim Daniels

    Jim Daniels - 2014-01-25

    Sure do!
    /usr/bin/taskcoach.py
    QDBusConnection: session D-Bus connection created before QCoreApplication. Application may misbehave.

    Now I try to sync
    python: Fatal IO error 11 (Resource temporarily unavailable) on X server :0.
    [xcb] Unknown sequence number while processing queue
    [xcb] Most likely this is a multi-threaded client and XInitThreads has not been called
    [xcb] Aborting, sorry about that.
    python: ../../src/xcb_io.c:274: poll_for_event: Assertion `!xcb_xlib_threads_sequence_lost' failed.
    Aborted (core dumped)

     
  • Jérome Laheurte

    Okay, let's try something else. Give it a try with the attached taskcoach.py, after installing faulthandler:

    apt-get install python-faulthandler

     
  • Jim Daniels

    Jim Daniels - 2014-01-26

    The linux box no longer crashes. However, there is no debug info either. The Linux box pops up an "Error" dialog with no content (nothing is written in the box). The Linux TaskCoach now becomes unresponsive. It will not respond to screen redraw on resize nor menu items.

    The iphone app says "Protocol negotion..." and is likewise unresponsive.

    I noticed another problem with the iphone app. When I started the sync attempt, my iphone was on the 3G network. However, when I switched over to the wifi network so I could find my Linux box, the iphone couldn't find it. I had to kill the app and restart with the iphone on the local wifi network in order to find the Linux box. This may be an issue with BonJour or possibly that your app doesn't notice the change in network.

     
    • Jérome Laheurte

      OK never mind, I found out. You didn't fill in a password in the iPhone preferences. Task Coach is supposed to show a message box in this case but it seems that under some environments MessageBox is not thread-safe (works on Ubuntu)... I'll fix that ASAP.

      As for the second problem, it's Bonjour. The discovery code is actually taken from an Apple sample and doesn't do anything related to the network type...

       
  • Jérome Laheurte

    A fix was made and checked into the source code repository of Task Coach. The fix will be part of the next release. You will get another notification when that release is available with the request to install the new release and confirm that your issue has indeed been fixed.

    If you like, you can download a recent build from http://www.fraca7.net/TaskCoach-packages/latest_bugfixes.py to test the fix.

    Because a fix has been made for this bug report, the priority of this report has been lowered to 1 and its resolution has been set to 'Fixed'.
    Thanks, Task Coach development team

     
  • Jérome Laheurte

    This bug should be fixed in the latest release of Task Coach. Can you please install the latest release of Task Coach and confirm that this bug has indeed been fixed?

    Thanks, Task Coach development team

     
  • Jim Daniels

    Jim Daniels - 2014-02-16

    I'm using the Mint distribution, which hasn't been updated. Also, I see no updates on iphone. I'm assuming the update is on the Linux side. Should I uninstall the package and install directly from the distribution?

     
  • Jérome Laheurte

    You'll only get automatic updates if you configured the PPA (https://launchpad.net/~taskcoach-developers/+archive/ppa); if not you'll have to upgrade by hand. Or wait for the Debian packager to update to this version.

    Yes, there is no update on the iPhone side.

     
  • Jim Daniels

    Jim Daniels - 2014-02-16

    The software claims to be up to date. My About Task Coach window is Version 1.3.99.6161, February 16, 2014

    Assuming this is the latest, it's no good. It crashed the KDE window manager hard. I still had mouse & keyboard, but nothing else. CPU time was maxed for a single CPU. I restarted using your debugging taskcoach.py. Here is the output:
    QDBusConnection: session D-Bus connection created before QCoreApplication. Application may misbehave.
    python: Fatal IO error 11 (Resource temporarily unavailable) on X server :0.

    Is this the right version? How the hell does KDE get fubar'ed so hard by a user-level app?

     
    • Aaron Wolf

      Aaron Wolf - 2014-02-16

      1.3.99 is NOT the stable build, that's the testing build for future versions that are not complete yet. http://www.taskcoach.org/download_daily_build.html has "latest bugfixes" which is the stable version but pre-release and "latest features" is not stable. So if you meant to test the stable bugfix release, use the first one instead.

       
  • Jim Daniels

    Jim Daniels - 2014-02-16

    Congratulations! The new version works fine. However, I would change the wording of the help message that pops up when there is no password. You ask to set a password in the iphone section. Since I'm not familiar with the software, I tried the iphone app first.

     
  • Jérome Laheurte

    OK I'll make the message clearer. Thanks for reporting.

     
  • Jérome Laheurte

    • status: fixed-and-released --> closed
     

Log in to post a comment.